The revenue model of an online South 

African stockbroking platform 

 

 

 

Wesley Ryan Bester  

Student number: 2607202 

Email: wesleybester.wb@gmail.com 

Supervisor: 

Euphemia Godspower-Akpomiemie, Ph.D. 

 

 

A research report submitted to the Faculty of Commerce, Law and 

Management, University of the Witwatersrand, in partial fulfilment of the 

requirements for the degree of Master of Management in the field of Digital 

Business. 

 

Johannesburg, 2024 

 

 



ii 
 

Abstract 

As technology advances at an unprecedented pace, many businesses and industries 

must adapt to the increasingly digital world. The online stockbroking industry is no 

exception and requires significant attention and change to keep up with the times. 

Business and revenue models in the stockbroking industry in South Africa have 

remained essentially unchanged over the past few decades. The variable-rate 

brokerage fee charged on transactions executed remains the primary source of 

income. This revenue model has rapidly become unsustainable with the decrease in 

these fees over the past few years. The study's main objective is to investigate 

revenue models that are more suitable for the digital trading environment. The study 

examines the background and appropriateness of alternative revenue models and 

platform models, along with the use of prospect theory to guide customer preferences. 

This quantitative case study utilises secondary data from a South African bank's online 

stockbroking division, analysing over 334,000 trades over 10 years. The entire dataset 

is analysed by looking at its descriptive, and inferential statistics, as well as time-series 

analysis. The study investigates the relationship between frequency, transaction 

amount, and their effect on brokerage over time, along with their association using the 

Chi-square model. Secondly, a model is built to predict a fixed monthly subscription 

fee for clients to replace the outdated variable-rate brokerage model. Clients will then 

choose a model of best value to answer the second research question. The study 

addresses two hypotheses, it also finds brokerage fees highly correlated to transaction 

values and inversely related to trade frequency. Based on the results, the model 

developed can effectively predict future fixed monthly subscription fees for online 

stockbroking platforms.  

Keywords 

Online trading, stockbroking, business model, revenue model, profit formula, platform-

based models. 

 

 



iii 
 

 

DECLARATION 

 

I, Wesley Bester, declare that this research report is my work except as indicated in 

the references and acknowledgements. It is submitted in partial fulfilment of the 

requirements for the degree of Master of Management in the field of Digital Business 

at the University of the Witwatersrand, Johannesburg. It has not been submitted before 

for any degree or examination in this or any other university. 

 

Name: Wesley      Signature:

 

 

Signed at: Sandton.  

 

On the 26th day of February 2024 

 



iv 
 

DEDICATION 

 

I dedicate this research to my supportive parents, siblings, friends, and colleagues 

who have fostered my growth and thirst for knowledge. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



v 
 

ACKNOWLEDGEMENTS 

 

I acknowledge with appreciation and thanks the following people: 

• My parents, siblings, friends, and family have provided a foundation for my 

growth and development while encouraging my quest for knowledge. 

• Dr Euphemia Godspower-Akpomiemie, my supervisor for her understanding 

and support throughout this project.  

• Dr John Nasila and the other WBS professors for their assistance and 

guidance. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



vi 
 

 

Table of Contents 

DEDICATION ................................................................................. iv 

ACKNOWLEDGEMENTS ............................................................... v 

LIST OF TABLES ........................................................................... ix 

LIST OF FIGURES ......................................................................... xi 

CHAPTER 1. INTRODUCTION ....................................................... 1 

1.1 Purpose of the Study .............................................................................................................. 1 

1.1.1 Statement of Purpose.................................................................................................. 1 

1.1.2 Framework of a Traditional Brokerage Model....................................................... 1 

1.2 Context of the Study ............................................................................................................... 2 

1.3 Research Problem ................................................................................................................... 8 

1.4 Research Questions ............................................................................................................... 9 

1.5 Significance of the Study ...................................................................................................... 9 

1.6 Delimitations of the Study ................................................................................................... 10 

1.7 Definitions of Terms .............................................................................................................. 10 

1.8 Assumptions ........................................................................................................................... 11 

1.9 Chapter Outline ...................................................................................................................... 11 

CHAPTER 2: LITERATURE REVIEW ........................................... 13 

2.1 Introduction............................................................................................................................. 13 

2.2 Background of the Study .................................................................................................... 13 

2.2.1 Introduction to Business Models ............................................................................... 15 

2.2.2 Introduction to Revenue Models ................................................................................ 17 

2.2.3 Introduction to Global Trends in Online Stockbroking ........................................ 18 

2.3 Alternative Revenue Models in Stockbroking ................................................................ 19 

2.3.1 The Influence of Pricing ............................................................................................... 19 

2.3.2 Digital Transformation and its Impact on the Financial Industry ...................... 21 

2.3.3 International Stockbroking Platform Robinhood is Setting Global Trends .... 22 

2.4 Platform Business Models are Suitable for Stockbroking ......................................... 24 

2.4.1 Critical Success Factors of Platforms’ Business Models ................................... 25 



vii 
 

2.4.2 Network Effects .............................................................................................................. 26 

2.4.3 Movement from Traditional Business Models to Platform Business Models 27 

2.5 Theoretical Framework ........................................................................................................ 29 

2.5.1 Application of Prospect Theory (PT) ......................................................................... 29 

2.6 Conclusion of Literature Review ....................................................................................... 30 

2.6.1 Summary of Findings on Alternative Revenue Models in Stockbroking ........ 30 

2.6.2 Summary of Findings on Platform Business Models ........................................... 30 

Chapter 3: Research Methodology ............................................. 33 

3.1 Research Approach .............................................................................................................. 33 

3.2 Research Design ................................................................................................................... 33 

3.3 Data Collection Methods ..................................................................................................... 34 

3.4 Population and Sampling .................................................................................................... 34 

3.4.1 Population ........................................................................................................................ 34 

3.4.2 Sample and Sampling Method .................................................................................... 34 

3.5 Descriptions of Variables .................................................................................................... 35 

3.6 Procedure for Data Collection............................................................................................ 37 

3.7 Data Analysis Strategies and Interpretation .................................................................. 37 

3.8 Possible Limitations and Challenges of the Study ...................................................... 39 

3.9 Quality Assurance ................................................................................................................. 40 

3.9.1 External Validity .............................................................................................................. 40 

3.9.2 Internal Validity ............................................................................................................... 40 

3.9.3 Reliability .......................................................................................................................... 40 

3.10 Ethical Consideration ......................................................................................................... 41 

CHAPTER 4: PRESENTATION, DISCUSSION AND INTERPRETATION 
OF RESULTS ................................................................................ 42 

4.1 Introduction............................................................................................................................. 42 

4.2 Data Cleaning ......................................................................................................................... 42 

4.3 Descriptive Analysis and Interpretation .......................................................................... 43 

4.3.1 Time Periods in Years – First Year 2013 ................................................................... 43 

4.3.2 Time Periods in Years = Third Year 2015 ................................................................. 45 

4.3.3 Time Periods in Years = Seventh Year 2019 ............................................................ 48 

4.3.4 Time Periods in Years = Eighth Year 2020 ............................................................... 50 

4.3.5 Time Periods in Years = Tenth Year 2022 ................................................................. 53 

4.3.6 Summary of Descriptive Analysis.............................................................................. 56 



viii 
 

4.4 Inferential Analysis and Interpretation ............................................................................ 57 

4.4.1 Time Periods in Years = First Year 2013 ................................................................... 58 

4.4.2 Time Periods in Years = Third Year 2015 ................................................................. 62 

4.4.3 Time Periods in Years = Seventh Year 2019 ............................................................ 66 

4.4.4 Time Periods in Years = Eighth Year 2020 ............................................................... 70 

4.4.5 Time Periods in Years = Tenth Year 2022 ................................................................. 74 

4.4.6 Transaction Value vs Brokerage 2013–2022 (Full Yearly Periods) .................... 77 

4.5 Brokerage Regression Modelling with Time-Series Data ........................................... 83 

4.5.1 Introduction ..................................................................................................................... 83 

4.5.2 Time-Series Analysis ..................................................................................................... 84 

4.5.3 Time-Series Prediction ................................................................................................. 90 

4.5.4 Summary of Time-Series Analysis ............................................................................ 96 

4.6 Modelling Customer Decision-Making between a Fixed or a Variable Brokerage 

Charge ............................................................................................................................................. 97 

4.6.1 Summary of Customer Decision-Making ............................................................... 101 

 CHAPTER 5. CONCLUSIONS AND RECOMMENDATION……..100 

5.1 Introduction ........................................................................................................... 102 

5.2 Conclusion ............................................................................................................................ 103 

5.2.1 Time periods, trading frequency and transaction values on brokerage rates . 103 

5.2.2 Differences between a subscription-based payment model and the existing 

transaction-based model on brokerage rates .................................................................... 104 

5.3 Recommendations .............................................................................................................. 105 

5.4 Suggestions for Further Research ................................................................................. 105 

References…………………………………………………………….106 

Appendices……………………………………………………………115 

 

 

 

 

 

 

 



ix 
 

LIST OF TABLES 

 

Table 1: Consistency Matrix, objectives, data detail and analysis method………... 40 

Table 2.1: Brokerage amount in 2013 …………………………………………….…...44 

Table 2.2: Brokerage amount 2015…………………………………………………….47 

Table 2.3: Transaction amount 2019…………………………………………………...48 

Table 2.4: Brokerage amount in 2020……………………………………………....….52 

Table 2.5: Transaction amount in 2022………………………………………...…..….53 

Table 3.1: Guidance on Chi-square tables: presentation and interpretation……....56 

Table 3.2: Summary of time periods vs brokerage results for 2013……………..…58 

Table 3.2.1: Chi-square test in 2013 for time periods vs brokerage…………….….58 

Table 3.3: Summary of frequency vs brokerage results for 2013………….............59 

Table 3.3.1: Chi-square test 2013 for frequency vs brokerage……………………..60 

Table 3.3.2: Crosstabulation in 2013 for time periods vs brokerage………….……60 

Table 3.4: Summary of results in 2015 of time periods vs brokerage ……………..62 

Table 3.4.1: Chi-square test in 2015 for time periods vs brokerage………………..62 

Table 3.4.2: Summary of results in 2015 for frequency vs brokerage……………...63 

Table 3.4.3: Chi-Square test 2015 for frequency vs brokerage ………..…………..64 

Table 3.5: Summary of results in 2019 for time periods vs brokerage………….....66 

Table 3.5.1: Crosstabulation in 2019 for time periods vs brokerage…………….…66 

Table 3.5.2: Chi-square test for time periods and brokerage in 2019………….…..67 

Table 3.5.3: Summary of results for 2019 of frequency vs brokerage…………..…67 

Table 3.5.4: Chi-square test in 2019 of frequency vs brokerage………………...…68 

Table 3.5.5: Crosstabulation in 2019 of frequency vs brokerage……………..….…69 

Table 3.6: Summary of results for 2020 of time periods vs brokerage …………….69 



x 
 

Table 3.6.1: Chi-square in 2020 for time periods vs brokerage…………..………...70 

Table 3.6.2: Crosstabulation in 2020 of time periods vs brokerage…………….…..70 

Table 3.6.3: Summary of results for 2020 of frequency vs brokerage ………….….71 

Table 3.6.4: Chi-square in 2020 of frequency vs brokerage ……………………..….72 

Table 3.7: Summary of results in 2022 of time period vs brokerage…………....…..73 

Table 3.7.1: Chi-square in 2022 of time periods vs brokerage …………………..….74 

Table 3.7.2: Summary of results in 2022 for frequency vs brokerage…………...….75 

Table 3.7.3: Chi-square in 2022 of frequency vs brokerage ………………………...75 

Table 3.7.4: Crosstabulation in 2022 of frequency vs brokerage………………...….76 

Table 3.8: Summary of results 2013 to 2022 for transaction value vs brokerage....77 

Table 3.8.1: Chi-square of transaction value vs brokerage over the entire period...78 

Table 3.8.2: Crosstabulation of transaction value vs brokerage………………….….78 

Table 4.1: Reliability tests for 2013 to 2022………………………………………...….80 

Table 5.1: Descriptive statistics………………………………………………………….83 

Table 5.2: KPSS Test………………………………………………………………….….85 

Table 5.3: Model selection criteria…………………………………………………...….89 

Table 5.4: Causal estimates for the linear model…………………………………...…90 

Table 5.5: Goodness of fit statistics……………………………………………………..91 

Table 5.6: Random sample of data to show predictions of the model……………….94 

Table 6.1: Historic brokerage fees on the same sample client set…………………...97 

Table 6.2: Monthly fixed predicted brokerage vs historic monthly variable fees……98 

 

 

 



xi 
 

LIST OF FIGURES 

 

Figure 1.1: A brief view of the traditional brokerage model ....................................... 2 

Figure 1.2: JSE equities market, order flow and clearing. ......................................... 4 

Figure 2.1: Capgemini Consulting Business Model Framework………………..........16 

Figure 2.2: Five critical dimensions of configuring a price model…………………….19 

Figure 2.3: Critical success factors on platform-based business models…………...25 

Figure 2.4: Revenue mix and technology investments in banks adopting platform 

models………………………………………………………………………………………28 

Figure 2.5: Conceptual Framework……………………………………………………..31 

Figure 2.6: Analytical Framework………………………………………………………..32 

Figure 3.1: Diagrammatic outline of the modelling………………………………….....38 

Figure 4.1: Absolute Tx amount for 2013……………………………………………....43 

Figure 4.2: Absolute Transaction amount in 2015……………………………………..45 

Figure 4.3: Brokerage data for 2019…………………………………………………….49 

Figure 4.4: Transaction amounts for 2020……………………………………………...51 

Figure 4.5: Brokerage data in 2022……………………………………………………..55 

Figure 4.6.1: Frequency of deals in 2013………………………………………………61 

Figure 4.7.1: Frequency of deals in 2015………………………………………………65 

Figure 4.8.1: Chart in 2020 of frequency vs brokerage……………………………….72 

Figure 4.9.1: Chart of time periods in 2022………………………………………….....74 

Figure 4.10.1: Brokerage <= 7000 over the period vs frequency…………………….79 

Figure 4.11.1: Brokerage of 33601+ over the period and frequency……………...…79 

Figure 5.1: Variable relationships and outliers with averages……………………..…84 



xii 
 

Figure 5.2: Original time series compared to differenced………………………….…86 

Figure 5.3: Autocorrelation function, original vs differenced………………………….87 

Figure 5.4: Yearly and monthly seasonality check…………………………………….87  

Figure 5.5: Variable relationships and outliers…………………………………………89  

Figure 5.6: Linear regression and ARIMA predictions…………………………………92 

Figure 5.7: PLOT ACF for residuals of the ARIMA model, ensuring no more 

information is left for the linear model…………………………………………………....93 

Figure 5.8: Daily actual brokerage vs predicted brokerage……………………………95 

Figure 6.1: Monthly prediction vs historic fees…………………………………………98 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



 

1 
 

CHAPTER 1. INTRODUCTION 

 
 

1.1 Purpose of the Study 

1.1.1 Statement of Purpose 

This case study examined the practicality of applying a new revenue model to the 

South African online stockbroking business model. Developing a revenue model 

aligned with a higher frequency of online trading can improve the way in which clients 

are charged. This case study showcases online investing platforms with alternative 

revenue models other than a purely traditional per-deal model, which only partially 

captures the new digital nature of trading. This case study employed a quantitative 

approach to assess the viability of an alternative revenue model. 

1.1.2 Framework of a Traditional Brokerage Model   

The framework of a traditional brokerage model in Figure 1.1 examines the challenges 

encountered in the stockbroking industry from a business standpoint. It aims to visually 

lead the reader through the issues and unique ideas presented in this research paper, 

and lays out the traditional model for brokerage charges, displaying the two main 

research problems, namely: competing through lower brokerage and not being geared 

towards digital channels. It proposes an analysis of historic deals to visualise trends 

in online brokerage rates, showing various influences while assessing whether a need 

exists for a new revenue model to address existing problems.  



 

2 
 

 

 

Figure 1.1: A brief view of the traditional brokerage model (Source: Author’s 

compilation)  

1.2 Context of the Study  

As the world becomes increasingly digitally driven, customers prioritise providers who 

meet their changing needs, expectations, behaviours and accessibility (Fu & Mishra, 

2022). This digital transformation and adoption trend has been accelerated by the 

COVID-19 pandemic, which affected various industries, including the financial sector 

(Kutnjak, 2021). The online stockbroking industry, in particular, experienced a surge in 

client growth, as investors preferred online platforms that catered to their needs 

(Ingrassia, 2021; Tan, 2021). According to Fu and Mishra (2022), the pandemic led to 

a remarkable surge in the daily downloads of finance-related apps, surpassing the 

typical growth rate by 21–25%. 

 



 

3 
 

South Africa is not excluded from this digital transformation in trading. South Africa’s 

Johannesburg Stock Exchange (JSE) moved from an open outcry market, where retail 

client orders were communicated to their stockbrokers and executed on the exchange 

floor, to electronic trading. The JSE commenced its digital transition in 1996 when it 

switched to an electronic trading system (Dicle & Levendis, 2013). Willing parties 

exchanged the opposite side of the trade in a traditional floor-based trading system 

executed through verbal communication. The JSE electronic trading system was 

additionally supported in 1997 by a real-time news service. Reporting price information 

was followed by the adoption of the London Stock Exchange’s digital order book, the 

Stock Exchange Electronic Trading Service (SETS) in 2002 (Dicle & Levendis, 2013). 

The Johannesburg Equities Trading system required orders to flow through the central 

order book unless the trade met specific off-book trading criteria1. The study by Dicle 

and Levendis (2013) regarding the JSE implementation of the SETS trading system 

showed evidence that trading activity almost doubled, and trading was more cost-

effective. These are the results of digitalisation, with enhanced functionality for both 

retail and institutional investors. 

The JSE was established on the 8th of November, 1887 and is Africa's oldest, largest 

and most liquid stock exchange (Lukasiewicz, 2019). The JSE follows an order-driven 

trading model and a limit-order trading system. The JSE is also included as one of the 

top 20 exchanges in the world using Market Capitalization2. The JSE’s stability is 

essential for offshore investors, who comprise most of South Africa’s daily trading 

volumes3. South Africa is highly competitive from an emerging market’s perspective 

and earns 49% of its equity market revenues from foreign sales, ranking fourth behind 

Taiwan, Chile, and Korea, respectively4. Taking these matters into account, a South 

African online stockbroking platform will be used as a case study, which is sufficiently 

comprehensive from a global, emerging market and African perspective. Global trends 

will apply to South Africa, and the established exchange provides sufficient historical 

data for analysis. 

 
1 Johannesburg Stock Exchange. (2023). Equity Rules, Rule 116.(page 44)(https://www.jse.co.za)  
2 South Africa Banking & Financial Services Report: (Q3- 2023). www.fitchsolutions.com/bmi 
3 Annual Results Presentation [Audited results, 9 March 2023]. www.jse.co.za 
4 Goldman Sachs. (2023). EM in Focus_ South African Equities - Weighing Domestic Risk Premium 
Against Better Fundamentals for Exporters  



 

4 
 

The JSE focuses on providing direct market access to clients through its members 

who follow the JSE regulations5. Client orders pass directly onto the exchange, 

matching buy and sell orders through the exchange central limit order book. Figure 1.2 

provides a visual cue of the order flows on the JSE. There is a clear illustration of the 

relationship between buyers and sellers and how the JSE trading system matches 

these orders. The JSE accounting system records all transactions, before the JSE 

clearing system sends the details to STRATE, the clearing and settlement authority. 

 

 

 

Figure 1.2: JSE equities market, order flow and clearing. (Source: Equities Operations 

| Johannesburg Stock Exchange (jse.co.za))6 

 
5 Johannesburg Stock Exchange. (2023). Equity Rules, Rule 116. (https://www.jse.co.za) 
6 Equities Operations | Johannesburg Stock Exchange (https://www.jse.co.za) 



 

5 
 

The South African online investing environment provides its clients direct market 

access to the exchange through the JSE’s central limit order book. Orders are matched 

against those with the best price on the screen7. In this case study, the South African 

parameters of trading through the central order book will remain and the study will not 

follow international off-book trading flows. The focus is to modify or improve how 

clients pay fees to better suit all stakeholders involved. Clients are charged additional 

fees over and above brokerage fees, including minimum transaction fees and monthly 

platform admin fees. These fees vary between platform providers but only make up a 

small portion of the total revenue earned. As equity brokerage rates are the primary 

source of revenue for this researcher, only those rates are accounted for in this case 

study. 

Transaction fees need to be considered by investors and traders alike and are likely 

to influence their choice of stockbroking provider. Woodside-Oriakhi et al. (2013) 

reviewed the various transaction costs applicable to executing financial assets and 

summarised them as fixed, variable, or a mixture of the two. Fixed costs are paid 

regardless of the transaction size, while variable costs are based on the transaction 

value (Woodside-Oriakhi et al., 2013). 

Globally, there has been a shift, particularly in the United States (US), to attempt to 

democratise trading by applying a zero brokerage or commission fee to transactions 

executed through online investing platforms (Tan, 2021). These platforms charge 

higher interest on deposits and sell their order flows to wholesalers in the market in 

return for a share in the spread earned. There are additional service subscriptions for 

premium services or tools. Robinhood, one of the pioneers of zero brokerage trading, 

has recently gained popularity, albeit initially launched in 2013. Robinhood was the 

first to market with this idea and successfully implemented its solution on a large scale, 

while numerous other platform providers have only recently followed in their footsteps 

(Berkow, 2021). A deeper dive into Robinhood’s successes and failures can illuminate 

where their model can be improved, mainly as pertaining to which aspects will and will 

not work in the South African environment. Important points need to be raised with 

regard to the various other fees that are charged as well as the end benefits passed 

 
7 Johannesburg Stock Exchange. (2023). Equity Rules, Rule 116.(page 44)(https://www.jse.co.za) 



 

6 
 

onto the clients of brokerage saved versus the increased spread which the client ends 

up paying for (Berkow, 2021; Chlistalla & Lutat, 2011).  

Platforms such as Robinhood in the US market have primarily earned most of their 

revenue by receiving a share in the spread earned for their client’s order flow (Dowdy, 

2023). These client orders are routed to a wholesaler who executes the orders off-

exchange and gives the clients a price slightly better than seen on screen. The 

revenue generated from sharing in the spread earned by the wholesalers can be much 

greater than the brokerage charged, as clients cross the spread between the bids and 

the offers (Ingrassia, 2021).  

Business models outline how a business can create and deliver value to its clients 

(Teece, 2010). Looking deeper into the business architecture, Teece (2010) found that 

revenues and costs leading to profits are essential for enterprises to deliver value to 

their clients. The central theme is creating and delivering value to customers while 

making a profit. At the same time, Johnson et al. (2008) focus more on the customer 

value proposition that competing offerings do not address. The four key elements 

which should be integrated to form a comprehensive business model are customer 

value proposition, profit formula, essential resources, and critical processes (Johnson 

et al., 2008). Later in this research, the terms profit formula and revenue model are 

used interchangeably. 

Revenue models are the blueprint for defining enterprise success with viable 

architectures for revenues and costs (Teece, 2010). Sources of revenue do vary over 

time and across industries, while technology has been aiding in reducing costs 

associated with information (Teece, 2010). The multiple revenue stream approach 

(Teece (2010) has seen online businesses following an extension to what the cinemas 

used to employ, with soundtracks being sold and additional memorabilia. Johnson et 

al. (2008) see the revenue model as part of the overall profit formula. The overall profit 

formula has four components: revenue model, cost structure, margin model, and 

resource velocity (Johnson et al., 2008). According to Johnson et al. (2008), the 

revenue model components can be viewed as the product of price and volume in its 

most basic form. 



 

7 
 

The digital era has changed various aspects of accessibility to the traditional 

stockbroking environment, making it more accessible, convenient, and affordable for 

retail investors to invest in the stock market. Online investing platform developments 

allow individuals to invest through personal computers, mobile devices, and tablets8. 

The increase in online investing has changed platforms’ offerings by including mobile 

trading apps and linking them to investment communities, and social media where 

investors can learn from each other. 

The rise of digital environments and the introduction of new models impact older 

business models, and change is required to survive this disruption. In an interview on 

disruption, Clayton Christensen touched on three innovations that affect a business’s 

growth and, indirectly, its profit margin (Christensen, 2020). The efficiency innovation 

mentioned by Christensen, namely, ‘when companies try to do more with less’, is 

interpreted as reducing costs by improving practical efficiencies until a point. 

Efficiencies do not create new growth and will not outlast a decreasing profit margin. 

Significant disruptions have been brought on in this digital era in all sectors. 

Businesses that did not keep up with these changes, such as Eastman, Kodak and 

Sears, were leaders in their sector until such time as they failed to keep up with the 

times (Christensen, 2020). Businesses can be created by repositioning, developing 

new business models, or repackaging existing revenue models. This has been 

exemplified by successful companies such as Uber and Netflix, which have effectively 

addressed their customers' needs (Christensen, 2020). 

Traditional stockbroking revenue models were primarily based on generating revenue 

from brokerage or fees, namely, a transaction-based revenue model and did not 

include other revenue sources. Additional services include investment and trading 

services such as advisory, management, commissions, custodian, and earning 

additional interest on various products9. These revenue models were traditionally 

based on execution on behalf of the clients, with lower volumes, added execution risk, 

and less frequency, and all managed through human interaction.  

 
8 FSCA Financial Sector Outlook Study 2022 
9 Equities Rules, 116 (2023) & Market Regulation | Johannesburg Stock Exchange 
(https://www.jse.co.za) 



 

8 
 

The rise of digital technology, and especially online trading platforms, has created a 

need for a new revenue model. Therefore, a change is needed, as all industries are 

undergoing significant changes caused by digitisation10. This study has tested and 

developed an alternative model to collect revenue from retail investing clients who use 

online platforms to execute their own trades. This case study examined South African 

online stockbroking platforms while drawing insights from international platforms. The 

author analysed secondary data and developed a new platform revenue model by 

forecasting a fixed monthly subscription amount and comparing it against the old 

variable costs for retail investing clients. The aim was to replace the lost profit margins 

experienced from diminishing brokerage charges. 

1.3 Research Problem 

The online stockbroking industry in South Africa faces a significant challenge in 

refining and changing its ways to charge clients for online platform execution 

services11. Traditional stockbroking revenue models were not designed for the 

increased competition of today’s digital era, with online platform providers lowering 

brokerage charges to remain competitive12. Due to the high JSE exchange costs, 

South African online stockbrokers cannot profit from low-value transactions unless 

accompanied by significantly increased volumes13. Online stockbroker revenue 

models need to be updated to align with other digital platform revenue models, or 

potentially be left uncompetitive and outdated.  

Profit margins earned by these online trading platforms are continually decreasing due 

to firms needing to reduce their brokerage to remain competitive. The reduced 

profitability further impacts the industry and the organisation by failing to attract top 

talent, who typically follow industries where profitability and earning potential are the 

highest (He, 2018). Based on this identified problem, two sub-problems emerged:  

 
10 FSCA Financial Sector Outlook Study 2022 
11 South Africa's Security Dealing Sector 2020: Market Challenges Poised by COVID-19 Pandemic 
(yahoo.com) 
12 www.fitchsolutions.com/bmi (South Africa Banking & Financial Services Report Q3-2023, pg. 51) 
13 www.fitchsolutions.com/bmi (South Africa Banking & Financial Services Report Q3-2023, pg. 51) 



 

9 
 

The current problem online stockbrokers face is they primarily compete against each 

other by lowering their transaction-based brokerage rates. The question that remained 

unanswered is: How do online stockbroking platforms replace this loss in fees? 

Secondly, traditional brokerage models, predating digital trading platforms, operated 

on lower volumes with a personalised touch, relying solely on phone or email for trade 

execution by a licensed broker. 

Retail clients in South Africa use online stockbroking platforms to trade for themselves 

without assistance. Previously, a stockbroker was needed to execute these needs with 

no alternative. Since the digitalisation of the JSE in 1996, the way clients are charged 

through a transaction-based fee has remained the same, while online trading 

capabilities have changed considerably. 

1.4 Research Questions 

Based on the explained problems, the study’s primary aim was to explore alternative 

revenue and business models for online stockbrokers better suited to a digital trading 

environment. The following research questions were developed accordingly: 

Research question 1: What are the effects of trading frequency and transaction values 

on brokerage rates over time? 

Research question 2: What are the differences between the subscription-based 

payment model and the existing transaction-based model on brokerage rates? 

1.5 Significance of the Study  

The findings of this research case study will assist existing online stockbroking 

platforms in South Africa in refining their revenue models and staying competitive while 

keeping their profit margins intact. This will also instil new entrants into the market with 

confidence, as they will be able to employ the updated platform revenue model. The 

research findings may also enable the regulators to fill in any gaps in legislation that 

might not have been considered for policy creation. Global literature covers early 

adopters of new revenue models in stockbroking, such as those by Robinhood, 

charging zero brokerage fees (Tan, 2021). 



 

10 
 

Moreover, the findings of this study will contribute to the academic literature on online 

stockbroking platforms and the various revenue models applicable in South Africa, 

which are currently limited. This study gauges the validity of platform models being 

applied to the online stockbroking industry in South Africa. Ultimately, attempting to 

implement the best-fitting model against historical revenue data to validate and confirm 

a change in the revenue model is necessary and possible.  

1.6 Delimitations of the Study  

This study is delimited as follows: 

I. South Africa is geared towards on-exchange trading through the JSE’s 

central order book; this premise was followed, and no off-exchange 

solutions were considered.  

II. The scope of this study consisted of a new revenue model to predict fixed 

subscription fees that can be applied to online investing platforms in South 

Africa.  

III. Regulatory changes during the study were not considered.  

IV. Tax concerns are beyond the scope of this study and were not considered. 

1.7 Definitions of Terms 

The following are the definitions of terms and key concepts of this research study. 

I. JSE – Johannesburg Stock Exchange, the primary exchange in South 

Africa.   

II. Business model – This describes the architecture of how businesses 

create and deliver value to their customers and how they extract their share 

of the value (Teece, 2018).   

III. Revenue model – This financial architecture defines how profits are made 

and extracts the value capture portion of the business model. Revenue 

models refer to how firms generate revenue within a business model (Penier 

et al., 2020). 



 

11 
 

IV. A digital platform business – This is described by Armstrong and Lee 

(2021) as using digital technologies to connect buyers and sellers, creating 

network effects, reducing frictions, allowing smoother transactions and 

externalising physical assets (pp 445-460). Digital platform businesses can 

either be in the form of industry, technology, or multi-sided platforms. 

V. Platform ecosystem participants – The ecosystem has five main 

components.  Firstly, the platform owner defines the purpose, the business 

model and how the platform is governed. Second, the complementors 

provide complementary offerings. Third and fourth are the buyers and the 

consumers. Last is the vehicle through which the platform is accessible. 

Platforms must ensure constant engagement, a continuous exchange of 

value and data, and constant feedback between all participants (Armstrong 

& Lee, 2021, pp. 472–478). 

VI. Online stockbroking platform – This digital platform provides clients 

access to execute buy and sell orders through a secure platform. 

1.8 Assumptions  

I. This study assumes that competition in South African online investing 

platforms is based on brokerage costs charged against deal size and the 

volume of trading execution. 

II. An assumption was made that the new revenue model can be statistically 

and financially tested by comparing it to historical data and modelling it to 

predict future trends.  

III. The updated revenue model will benefit clients and platform providers. 

Clients will select the model with the best value. 

1.9 Chapter Outline  

Chapter 1 starts with the purpose of this study, which explores the idea of an 

alternative revenue model for South African online investing platforms. This chapter 

outlines the basics of a business model, described by Teece (2018) as an architecture 

of how businesses create value for clients and extract their profits. The background 

introduces the core themes contained in the study, including stockbroking, pricing, 

platform models, and international trends in stockbroking. 



 

12 
 

Chapter 2 consists of a systematic literature review, followed by Chapter 3, which 

describes the research methodology deployed. Chapter 4 presents the research 

findings and discusses these findings. 

Chapter 5 presents the author's conclusions, various recommendations, and areas for 

future research. 

 

  



 

13 
 

CHAPTER 2: LITERATURE REVIEW 
 

2.1 Introduction 

This chapter follows a systematic approach described by Kitchenham and Charters 

(2007) as identifying, evaluating and synthesising the available research. This 

literature review examines available research in relation to the aim of this study.  

Relevant theories are explored to delve deeper into the primary aim of exploring 

alternative revenue models in stockbroking in South Africa. The background 

introduces relevant topics and themes to guide and address the research questions. 

Revenue models are discussed, with business models being crucial in choosing 

pricing mechanisms. International literature is reviewed for variety and reach, as well 

as South African literature, to guide the best-in-class approach when aligning the 

South African perspective. The review also focuses on digital platforms, the required 

ecosystems, and their relevance to the stockbroking industry. A theoretical overview 

of prospect theory (PT) is carried out, as well as its applicability to address the 

research question on customer decision-making behaviour. This chapter concludes 

with research question summaries, the literature’s conceptual framework, and an 

analytical overview of connecting variables.  

2.2 Background of the Study  

The primary motivation of this study is to change how online stockbroking businesses 

charge clients in South Africa. The first sign of digitisation of the exchange in South 

Africa occurred in 1996 when the JSE transitioned towards being fully electronic for 

trading functionality (Dicle & Levendis, 2013). There is a need for South African online 

stockbroking platform providers to further develop, match and implement the relevant 

international best practices to remain relevant and competitive. 

The JSE’s automated trading system, now known as TradeElect, licensed by the 

London Stock Exchange, processes this order-driven market (Wyk et al., 2015). The 

JSE central order book accepts orders from investors through a registered JSE broker 

through TradeElect, where it matches automatically with an opposite order (Wyk et al., 

2015). Globally, equity markets are considered vital in the financial system as a whole 

and are a critical source of finance in the real economy (Wyk et al., 2015). Equity 



 

14 
 

markets are essential for the financial system, and the role of stockbrokers in 

facilitating a viable platform for clients’ needs to be understood and geared towards 

long-term sustainability in the new digitally-driven environment. 

South African Policy and Regulation 

The Financial Services Conduct Authority (FSCA) oversees the regulation and 

supervision of the securities exchange of South Africa. Their aim is to ensure capital 

markets are efficient, fair, and transparent by providing a regulatory framework for 

stock exchanges to conduct business. This ensures investors can buy and sell shares 

in a conducive environment14. The FSCA sets the guiding framework of 

the Financial Markets Act 19 of 2012, which is the overarching requirement for 

stockbrokers to abide by, and which the JSE equity rules expand on15. 

 

The Financial Markets Act requires a stockbroker to be an authorised user and sets 

the standard on pricing and fees. This includes two rules: 1. Stockbrokers ‘must 

disclose to their clients the fees for their services, which disclosure must give the 

specific monetary amount for each service rendered; or if such amount is not pre-

determinable, the basis of the calculation’. 2. Stockbrokers ‘may charge a fee for 

different categories of transactions. The JSE expands further on how stockbrokers’ 

price their fees in terms of the JSE equity rules: 

• A mutually agreed fee may be charged to the client in advance of such a 

transaction. 

• Profits are to come only from agreed commissions or fees. 

• In respect of a transaction other than the agreed commission or fee, full 

disclosure and accurate information about the fees and charges is required. 

• Fees and charges are to be reflected in specific monetary terms. 

International retail trading apps such as Robinhood have become immensely popular 

in the US, attracting both first-time and experienced investors. These new applications 

offer enhanced convenience and reduced trading expenses, which disrupt 

conventional brokerage business models by providing zero commissions (Tan, 2021). 

 
14 Financial Markets Act 19 of 2012  
15 Johannesburg Stock Exchange. (2023). Equity Rules, Rule 116.(https://www.jse.co.za) 



 

15 
 

Stock exchanges and brokers address the fundamental issues investors face by 

reducing information asymmetries by publishing prices, providing infrastructure, and 

matching buyer and seller transactions (Feyen et al., 2021). To stay competitive 

against big technology companies, financial institutions have had to embrace new 

technologies and break down the production of financial services to improve their 

efficiency (Feyen et al., 2021).  

 

As equity markets become entirely digitally driven, updating business and revenue 

models will become vital to unlocking value and exploring the latest trends, such as 

platform business models. Platforms can be seen as intermediaries that allow and 

provide for exchange between participants who deploy the use of platforms (Goldfarb 

& Tucker, 2019). Platforms can monopolise supply- and demand-side economies of 

scale while pursuing a winner-takes-all position (Croxson et al., 2021). This study 

determines if the current traditional transaction-based pricing model for stockbroking 

platforms would be better replaced by a subscription-based model.  

2.2.1 Introduction to Business Models 

Historically, business strategy and traditional measures such as resourced-based 

views dominated the literature. During the past two decades, the literature has shifted 

to the business model as a unique concept spurred on by the growth of the internet 

and e-commerce (Teece, 2010). In the past, strategy theories that preceded business 

models, such as the resource-based view of the firm or its positioning, posited that 

value creation was only a result of producers and not customers. This approach 

focused on supply-side factors and a single competitive advantage (Barney & Arikan, 

2005; Porter, 1996; Porter & Kramer, 2011).  

A company's business model refers to the framework it uses to manage the critical 

aspects of providing value to customers, receiving payment for that value, and 

converting it into profit. This definition, coined by Teece (2010), encompasses a 

business's overall structure and strategy. According to various scholars, the value 

proposition of a company shapes its business model, including the manner in which it 

creates and captures value for its clientele (Massa, Tucci & Afuah, 2017). Teece (2018) 

further narrows down his reasoning on business models to focus on identifying the 

target customers and the methods of revenue generation. Strong dynamic capabilities, 



 

16 
 

organisational design and business models work together and are seen as 

interdependent functions of each other for success (Teece, 2018). Teece (2018) 

considers the interactions between these as the central theme, where solid dynamic 

capabilities enable the effective implementation of business models. 

Saebi et al. (2017) bring an adapted dimension into the business model literature, as 

the management adapts their business model to the changing environment. According 

to research, businesses tend to change their business models more readily in 

response to threatening conditions rather than perceived opportunities (Saebi et al., 

2017). Depending on the perceived threat, these adaptions to the business model can 

be large or small. 

Internal and external forces influence a business model’s environment. Ramdani, 

Binsaif, Boukrami and Guermat (2020) summarise past literature elaborating on 

internal and external challenges. According to Ramdani et al. (2020), there are two 

internal challenges that affect the business model: top management and 

organisational culture. A business can face seven external obstacles that may impact 

its operations. These include crises, client demands, regulatory changes, 

advancements in technology, competitive pressures, and industry/service provider 

influences that affect the overall environment (Ramdani et al., 2020). 

To provide clarity of underlying components within the business model, drawing from 

Schön (2012), a modular approach assists in graphically showing the interactions. 

 

Figure 2.1: Capgemini Consulting Business Model Framework. (Source: Schön, 

2012) 



 

17 
 

The linkages between the three central dimensions: value proposition, revenue model 

and cost model, are shown in Figure 2.1. In each dimension, relevant influencing 

factors, such as pricing logic and the specific channel to market within the revenue 

dimension, are linked to the overall synergy they produce (Schön, 2012).  

The literature builds a foundation of how businesses and business models must be 

adapted from a broader perspective then just a single approach. David Teece 

pioneered the early thinking of business models in the digital era and separated these 

models from being termed strategy (Saebi et al., 2017). There can be minor or 

significant adaptations made to a business model depending on one’s circumstance, 

as Saebi et al. (2017) posited, while Linnenluecke (2017) shows that sticking to one’s 

core business model through all occasions can work. With the advent of digitisation, 

the stockbroking industry has undergone significant changes. A more drastic approach 

to the business model is necessary to remain relevant and meet customers' evolving 

needs. A redesign of the revenue model can address these needs. 

2.2.2 Introduction to Revenue Models 

 

According to Clauss (2017), businesses can capture financial value by exploring 

alternative revenue streams, adjusting their pricing strategies, and evaluating their 

profitability and sustainability. Clauss (2017) takes this one step further, where 

attention is drawn to changing business models by introducing new cost structures, 

particularly new revenue models. In a business model, the value capture aspect is 

determined by the revenue model applied to determine its financial success and by 

adjusting revenue streams and cost structures to be optimal (Vaska et al., 2020). 

Ramdani et al. (2020) compiled an interesting study across numerous business 

divisions within the financial sector, and of particular interest is the asset management 

division within which stockbroking is a crucial activity. The value proposition is the 

essential dimension with notable important sub-dimensions, including revenue 

streams, where it was found fee-based models are the most dominant (Ramdani et 

al., 2020). The stock brokerage division responded to four challenges: new technology, 

crises, client demands and competitive pressure. It addressed them by deploying 

multi-brokerage models, charging clients trading commissions, and lending revenues 

(Ramdani et al., 2020). The revenue streams in the investment banking division were 



 

18 
 

not consistent and were determined on a per-deal basis, including fixed fees, 

transaction-based fees, and success fees (Ramdani et al., 2020). 

Vaska et al. (2020) relay a final finding on value capture, asserting that transaction-

based revenue models are not viable for long-term application. If no further action is 

taken and transaction-based revenue is examined in isolation, Vaska and colleague’s 

findings will be considered accurate according to this research study. The reality is that 

there will be slight adjustments made throughout the business’s life cycle, as affirmed 

by Clauss (2017), who found that adjustments in some way or form should be made 

to impact either cost or revenue structures.  

2.2.3 Introduction to Global Trends in Online Stockbroking  

Recently, there has been a rise in retail trading activity due to the increasing popularity 

of financial technology (FinTech) apps used for investing, including Robinhood, a well-

known US-based retail trading app (Tan, 2021). These new retail trading clients use 

apps such as Robinhood due to lowered trading costs and increased convenience, 

disrupting traditional brokerage models with zero-commission trades (Tan, 2021). This 

speaks to the US households reflecting their financial culture, which embraces more 

aggressive risk-taking and actively managing their investments (Fligstein & Goldstein, 

2015). First-time investors rushed to join zero-cost platforms such as Robinhood to 

play the ‘game’ of speculation, with the platform’s user-friendly interfaces and focus 

on user engagement, which lures them into this gamified trading space (Tan, 2021).  

According to Lazaro and Verges' (2022) timeline, the impact of digitalisation on 

brokerage charges in the US is evident. The timeline reveals that with the emergence 

of day traders and a shift towards online trading, commission charges have now been 

reduced to zero, with 25% of trades now being conducted online. South Africa is a 

laggard regarding these global trends; the country is experiencing diminishing 

brokerage charges to remain competitive as online trading increases. However, South 

Africa has not reached zero-brokerage trading. Section 2.3 presents various themes 

that have the most significant impact on alternative revenue models in stockbroking. 



 

19 
 

2.3 Alternative Revenue Models in Stockbroking  

The literature provides an overview of the challenges and applications that are 

particularly relevant to the revenue models in stockbroking. Several themes are 

highlighted, the most pertinent of which are expanded. 

2.3.1 The Influence of Pricing  

All businesses aim to create a price model that effectively entices customers to pay 

for their products or services. The price model should be configured to ensure a steady 

influx of revenue while delivering long-term value to the customers (Petri, 2014). In a 

study by Petri (2014), a slight change in the price model resulted in a radical shift in 

the business model itself. The evidence showed that changes could affect cash flow 

by shifting dimensions within the price model (Petri, 2014). According to Penier et al. 

(2020), a revenue model refers to how a business generates income. According to 

Penier et al. (2020), the earned revenue model involves providing a service and 

charging a fee to generate revenue for the business. Critical dimensions for a price 

model are clearly described by Petri (2014) in Figure 2.2, along with a description of 

the five key attributes. 

 

Figure 2.2: Five critical dimensions of configuring a price model. (Source: Petri, 2014, 

p. 5) 



 

20 
 

In Figure 2.2, the scope of the offering refers to either a single attribute being paid for, 

or a package of bundled items being bought together. The next dimension, the pricing 

base, was influenced by cost, and the customer or competitor information was used 

to price in the second dimension. The third dimension, influence, relates to buyer or 

seller negotiation over the price, with exogenous pricing beyond their control. The price 

formula connects volume and price and their specific combinations to determine a final 

price. The fifth and final dimension, temporal right, relates to the customer’s time 

frame, with perpetual being the longest and pay-per-use the shortest. According to 

Petri (2014), the pricing in Figure 2.2 is a reliable and effective method for creating 

sustainable and profitable revenue streams, which can significantly improve the overall 

business model.  

In transaction markets, platforms can generate revenue from people joining the 

platform and using it by sharing in the monetary value (Filistrucchi et al., 2014). Fees 

can be divided into two categories: transaction-based fees, which are simply 

transaction fees, and service-based fees, consisting of service, connection and 

membership fees (Kemppainen et al., 2018). 

Modifying the price formula must be considered to address the research objective of 

whether an alternative revenue model is better than the traditional stockbroking 

brokerage model. A net monthly fixed subscription could be implemented as opposed 

to the current fixed fee plus a fee per transaction. This small change could have a 

significant impact on the industry. Petri (2014) states that to adjust the price model to 

have the desired effect, it should consider the overall business model and attempt to 

amplify and enhance its core features. Business models in resilient firms deal with 

challenges quicker, recover faster, and develop different or non-conventional ways of 

working while under pressure (Linnenluecke, 2017).  

South African stockbroking businesses still in existence can be described as resilient 

firms, as defined by Linnenluecke (2017). Stockbroking fees in South Africa are 

primarily transaction-based fees, described by  Kemppainen et al. (2018) as one of 

the primary categories of fees. Looking at Petri (2014) and the success achieved by 

changing the price formula, this could have the desired effect, namely, increasing the 

sustainability of online stockbroking platforms. The next theme presents digitalisation 

as a catalyst for global change, linking the traditional stockbroking business model 



 

21 
 

poised for transformation and highlighting the importance of exploring alternative 

revenue models. 

2.3.2 Digital Transformation and its Impact on the Financial Industry 

The role of technological innovation in driving economic growth and industrial 

transformation is widely acknowledged. Rapid speed and constant evolution bring 

transformative changes (Gomber et al., 2018). Digital transformation harnesses new 

technology, allowing firms to capture value through platforms and improve customer 

relationships (Gomber et al., 2018). Vaska et al. (2020) identified four value creation 

dynamics brought on by digital transformation; the most relevant to stockbroking would 

be the revision and extension of existing services online and the offering of new value 

propositions precisely how clients want them. In the financial sector, disruptive 

technologies are employed against traditional business models to facilitate 

sustainability and harness the sharing economy (Vaska et al., 2020).  

The cost of digital technology can impact economic actions, while digital economics 

aims to study how cost changes can alter economic models (Goldfarb & Tucker, 2019). 

Digital environments offer numerous benefits, such as reduced search costs, almost 

costless digital replication, individual behaviour tracking, and digital verification. These 

factors combine to create significant cost savings and improve overall economies 

(Goldfarb & Tucker, 2019).  

Liu et al. (2011) described digital transformation or ‘digitalisation’ best by stating it as 

integrating digital technologies into business processes. This aligns with the problems 

outlined in Chapter One with stockbroking revenue models not accounting for the 

digital era. In today's digital world, businesses must utilise digital technologies and 

platforms to gather and integrate data in order to stay competitive in platform 

economies (Petrakaki et al., 2018). The goal of digital transformation is to enhance an 

organisation by initiating changes through communication, information, computing, 

and connective technologies (Vial, 2019). Daugherty et al. (2016) have stated that 

platform-based business models are the most significant change in the world since the 

18th century Industrial Revolution. 



 

22 
 

South Africa, an emerging market economy such as Taiwan, allowed its first electronic 

trading through the exchange in 1996 (Dicle & Levendis, 2013). These two markets 

are good for comparative reasons across this study while looking to the US and Europe 

for best-in-class examples in the stockbroking industry. Lin et al. (2021) analysed the 

online stockbrokers in Taiwan, who first allowed an online electronic transaction on 17 

October 1997, where the order rate only accounted for 0.02% of stock exchange 

transactions. Taiwanese’s online stockbrokers comprised 53 firms in 2020, and their 

electronic orders accounted for 65% of overall market turnover, with commission-

based transactions accounting for 60% of transactions and brokerage fees accounting 

for 70% of the overall revenue (Lin et al., 2021).  

The study by Lin et al. (2021) discovered that many online stockbrokers offer 

brokerage discounts to entice customers, diminishing their overall efficiency value. 

FinTech has positively impacted the stockbroking industry, as brokers can now provide 

electronic trading and other valuable services. According to Lin et al. (2021), there has 

been notable growth in this area. The next theme explores how Robinhood, the global 

leader in online stockbroking platforms, became a market leader and disrupted 

traditional stockbroking models in the US. 

2.3.3 International Stockbroking Platform Robinhood is Setting Global Trends  

New technologies have made online trading more accessible for younger investors, 

democratising and revolutionising private retail investing and changing its ecosystem 

(Ingrassia, 2021). According to Tripathi and Rengifo (2023), the Robinhood Effect, 

which involves increased trading activity and stock holdings facilitated by the popular 

trading platform Robinhood and its fractional trading capabilities, has resulted in a 

significant influx of $53 billion in new investments into the stock market. In Chapter 

One, it was mentioned that Robinhood's zero-brokerage model is a critical factor in 

attracting investors. Robinhood sells its clients' order flow and receives payment to 

compensate for the revenue lost from not charging brokerage fees. A closer look at 

the payment for order flow is required for completeness. 

Payment for order flow (PFOF) is the revenue earned by a broker-dealer (i.e., 

Robinhood Financial LLC) for routing information to a market maker (i.e., Citadel LLC) 

from their online stockbroking platform on their client orders (Ingrassia, 2021). 



 

23 
 

Platforms such as Robinhood receive trading revenue via PFOF, allowing them to 

charge zero brokerage on their client’s trades. This trend is also occurring with other 

major online platform providers, who are altering their traditional brokerage models to 

zero-brokerage models and finding alternate means to generate returns (Ingrassia, 

2021). Robinhood declares its revenue through these streams: rebates from market 

makers and trading venues (PFOF), subscriptions, margin interest, stock loans, 

income generated from cash, and cash management16.  Many trades made by retail 

investors occur off-exchange. In these cases, the broker's inventory is used to match 

orders, or they are sold to a wholesaler (Eaton et al., 2022). Retail trading that takes 

place outside of exchanges is known as dark pool or dark market trading. Such trades 

must be reported to the Financial Industry Regulatory Authority within a span of 10 

seconds (Eaton et al., 2022).  

A study by Pagano, Sedunov and Velthuis (2021) on the realised spread between the 

buying and selling prices based on Robinhood’s user activity showed that the greater 

the retail client participation, the more market quality and lower costs were achieved. 

The impact on share price is also higher with increased user activity in regular periods, 

showing their users can affect price, which indicates lower market quality (Pagano et 

al., 2021). According to a recent study by Pagano et al. (2021), the participation of 

retail investors in financial markets can significantly affect trading quality, especially 

during times of market stress. Hence, institutional investors and regulators must take 

note of this finding. 

Other ways of addressing the expense of trading in the equity market include trading 

through additional or alternative exchanges that charge lower exchange fees, such as 

the Chi-X exchange in Europe (Chlistalla & Lutat, 2011). South Africa has also 

attempted to address these high exchange cost inefficiencies by starting three 

alternate exchanges: ZARX, A2X and the CPT exchange17. There has been no attempt 

to sell order flow in South Africa to market makers such as in the developed countries 

with higher transaction volumes and values. South African retail investors trade on-

 
16 https://robinhood.com/us/en/support/articles/how-robinhood-makes-money/ 
17 www.fitchsolutions.com/bmi (South Africa Banking & Financial Services Report Q3-2023, pg. 51) 



 

24 
 

screen through the JSE central order book and do not trade off-exchange through their 

brokers, who abide by the JSE member exchange rules18.  

Hypothesis 1: 

There is no relationship between time periods, trading frequency and market 

conditions and their effect on brokerage rates. 

To address the research question of whether an alternative revenue model will be 

better suited than the traditional transaction-based revenue model, hypothesis 1 is 

tested on the relationship between the time periods, trading frequency and market 

conditions. To test this hypothesis, the author analysed 10-year historical secondary 

data on time periods, trading frequency, transaction amounts, and brokerage rates. 

2.4 Platform Business Models are Suitable for Stockbroking 

The literature provides an overview of challenges and applications most relevant to 

different platform-based business models. Although numerous challenges are 

mentioned, the most significant are described below. 

The author expands on the brief introduction to platform models provided in Chapter 

One’s introduction and the background provided earlier in Chapter Two. In their book, 

Armstrong and Lee (2021) mentioned three major platform types to pay attention to: 

internal product platforms, industry (or technology) platforms and multi-sided market 

platforms (pp 445-460). Each platform type has a value focus area, internal or external, 

with upstream or downstream participation and various key success factors 

(Armstrong & Lee, 2021, pp. 472-478). The critical success factors highlighted here 

are ecosystem value opportunities, network effects, innovation and asset 

externalisation and frictionless transactions (Armstrong & Lee, 2021, pp. 445-478). 

The next platform model theme with relevance to stockbroking is the universal success 

factors which apply to all industries. 

 
18 Johannesburg Stock Exchange. (2023). Equity Rules, Rule 116. (https://www.jse.co.za) 



 

25 
 

2.4.1 Critical Success Factors of Platforms’ Business Models 

According to Rohn, Bican, Brem, Kraus and Clauss (2021), the network of participants 

on a platform and their information exchange interactions are the most important 

intangible assets for facilitating transactions. Creating value within a platform context 

involves leveraging the platform's technology and network externalities to facilitate 

connections between participants, whether on a business-to-business or consumer 

basis (Rohn et al., 2021). When determining the price structure of platforms, three 

factors come into play, namely: the relative size of cross-group externalities, the type 

of model used (subscription or payment-per-transaction), and whether or not the 

customer uses multiple platforms to complete the desired activity (Rohn et al., 2021). 

Figure 2.3 extends the three traditional business model factors, including start-up 

culture, platform architecture and the advocation of digital transformation to achieve 

success in a platform business model (Rohn et al., 2021).  

 

Figure 2.3: Critical success factors on platform-based business models. (Source:  

Rohn et al., 2021) 



 

26 
 

2.4.2 Network Effects 

The network effect refers to the influence that a particular number of platform users 

have on the value generated by every new user, which can be either positive or 

negative (Parker et al., 2016). Stated differently using Metcalf’s law, the value of the 

network grows non-linearly as the number of users of a network increases as users 

can now connect (Parker et al., 2016). Network effects are enhanced and exaggerated 

through frictionless entry and the ability to scale rapidly, maximising its value-building 

impact (Parker et al., 2016).  

Network effects occur when at least two or more participants are connected to a 

platform to interact with one another (Rohn et al., 2021). Croxson et al. (2021) summed 

it up best, stating that greater market power and the ability to leverage this power come 

with the increased market size. Tipping effects can be used to build market momentum 

and, in so doing, achieve platform leadership (Armstrong & Lee, 2021). Looking at the 

stockbroking industry in South Africa, a community network effect has not yet been 

effectively taken advantage of by existing stockbroking platform providers. 

Technology progression has threatened many industries’ business continuity and 

rendered established business models obsolete, seeing new tech giants such as 

Amazon, Uber and Alibaba reorganising and renewing the market (Niemimaa et al., 

2019). The world's most prominent players and giants such as Walmart and Target 

were forced to innovate their business models or fade into non-existence (Niemimaa 

et al., 2019). To achieve sustainability and success in platforms, certain factors must 

be considered. These include reaching a critical mass, facing competition, thriving in 

winner-takes-all markets, and fostering collaboration within business ecosystems. 

Ruutu et al. (2017) summarise that end users, developers, and service providers all 

form part of the requirements for gaining critical mass for self-sustaining growth and 

scaling a platform successfully. 

Internet-based platforms have crucial roles regarding the applicability and use of data 

and the corresponding network effects of end users where data accumulation can be 

substantial (Ruutu et al., 2017). Specific South African stockbroking platforms can be 

seen as having already achieved a critical mass for survival. These internet-based 

platforms need to extract relevant data and focus on exploiting network effects from 



 

27 
 

existing users to scale and grow exponentially. From this description, stockbroking 

platforms can be viewed as internet-based platforms.  

2.4.3 Movement from Traditional Business Models to Platform Business Models 

The dynamics of two-sided platform markets are shaped by various factors, including 

the platform's structure and the frequency and value of user transactions on the 

platform, and not solely determined by the fees charged (Rochet & Tirole, 2006). The 

swift expansion of FinTech and big tech companies into finance has led banks and 

financial institutions to adopt platform models to keep pace, utilising client data and 

automation to provide various third-party services tailored to multiple client groups and 

markets (Croxson et al., 2021). In platform markets, when more users are attracted, 

the average cost per user decreases, and the average return increases. The 

willingness of users to join a more comprehensive platform increases revenue 

(Croxson et al., 2021). Interoperability or multi-homing plays an important role, 

allowing users to interact with others on different platforms; even if just a particular 

group of users within a platform (Croxson et al., 2021). In financial services, this could 

entail using interoperable payment systems, which allows for excellent efficiency. The 

stockbroking industry could use different stock exchanges or various financial products 

on a platform. 

The end user’s willingness to transact on a platform is influenced by the platform’s 

membership, usage, and variable and fixed charges, ultimately determining the end-

user presence on the platform (Rochet & Tirole, 2006). Fixed fees such as 

membership or platform fees could be the most efficient way to capture end-user 

surplus; these fixed fees are in place to cover the platform’s costs (Rochet & Tirole, 

2006). Platform-based business models find that their primary income source is 

subscription fees (Croxson et al., 2021). This fee income closely fits the match-making 

business model, requiring little funding, balance sheet, regulation or supervision, with 

the core value coming from their network and data (Croxson et al., 2021). Online 

platforms can utilise their customers' data and advanced algorithms to determine 

everyone’s reservation price and offer personalised pricing slightly below that 

threshold. This practice, known as ‘value-extracting innovation’, may be considered 

monopolistic conduct, according to Croxson et al. (2021). 



 

28 
 

Many financial institutions are transitioning from traditional business models to those 

used by FinTech and big tech companies. These financial institutions generate 

revenue through fees instead of relying on net interest income. However, this shift 

requires significant investment in digital technologies to collect and analyse big data 

to personalise offers, such as what big tech platforms do (Croxson et al., 2021). 

According to Croxson et al. (2021), platform-based banks now generate almost 40% 

of their revenue from fees and non-interest income, where previously their revenue 

was dominated by interest income as the sole source of banking revenue. In contrast, 

their peers are only extracting 33% from this avenue. This shift is attributed to these 

banks investing 50% more than their peers in technology, data processing, and 

communication as a percentage of their overall expenditure. Figure 2.4 displays banks 

with platform business models in red against competing banks with more total assets 

above 50 billion USD in blue. 

 

Figure 2.4: Revenue mix and technology investments in banks adopting platform 

models (Source: Croxson et al., 2021) 

Looking at this global trend of adopting platform models, South Africa cannot be far 

behind and should consider adopting platform business models. As international 

financial institutions' structures change when they adopt platform models, this change 

is needed in the South African stockbroking industry to remain competitive. The author 

uses PT to determine if clients prefer a fixed monthly subscription-based pricing 

structure or the stockbroking industry's traditional fixed fee per transaction model. 

 



 

29 
 

2.5 Theoretical Framework 
 

2.5.1 Application of Prospect Theory (PT) 

In 1979, Kahneman and Tversky's research paper on decision-making under risk 

showed experimental proof of disregarding expected utility; their research is known as 

prospect theory (PT) (Barberis, 2013). This original work was updated in 1992, known 

as the ‘cumulative prospect theory’, and is still considered the best description of how 

people evaluate risk in experimental settings (Barberis, 2013). Kahneman and 

Tversky’s work was designed as a substitute for the expected utility theory, which did 

not consider how individuals make decisions in risk situations (Edwards, 1996). 

Barberis (2013) summarises the four critical elements of PT: 

I. Reference dependence – used from gains and losses measured relative to a 

reference point. 

II. Loss aversion – value function, where individuals are more sensitive to losses.  

III. Diminishing sensitivity – value function towards a gain is concave and convex 

for losses.   

IV. Probability weighting – overweight low probabilities and underweights higher 

probabilities.  

Chang and Nichols (1987) use PT in their tax audit risk paper, demonstrating the 

applicability to payments and not merely income, where a tax payment can be 

perceived as a gain or a loss depending on the individual. In their research, Kőszegi 

and Rabin (2009) emphasise the significance of reference points and how they relate 

to loss aversion in PT. Specifically, they explore the difference between anticipated 

and realised consumption affecting reference points. Drake and Freedman (1993) 

found that extra time and effort are applied when a considerable percentage discount 

or savings are likely to be gained. Their finding presented an alternative perspective 

to PT, which asserts that an individual's decision is influenced solely by the percentage 

of the discount. A shortcoming identified by Edwards (1996) was that the predictive 

ability of PT found limited support in earlier work when applying the theory. 



 

30 
 

The author used PT to analyse user behaviour on a trading platform, comparing fixed 

monthly subscription and transaction-based pricing models to determine which one 

user’s thought was of greater value. Alike to Chang and Nichols (1987), the author has 

shown that paying a fee structure that aligns with personal preferences can be 

advantageous for individuals and seen as a gain.  

2.6 Conclusion of Literature Review  

2.6.1 Summary of Findings on Alternative Revenue Models in Stockbroking 

Digital technology must be embraced and adapted to the stockbroking business’s 

unique environment.  Adjusting the stockbroking business revenue collection method 

is essential to compete in this digital environment and to remain relevant. Robinhood 

has set the trend in the US, where zero-brokerage costs have become a way to 

compete with all major traditional brokerage houses following in their footsteps to 

remain competitive (Pagano et al., 2021). Petri (2014) shows that a slight change in 

pricing can have significant industry impacts.  

2.6.2 Summary of Findings on Platform Business Models 

While platform-based business models are new, there has been a definite trend 

towards their adoption from traditional business models across various industries. A 

common theme for successful platform businesses is the harnessing of network 

effects, externalising assets, exploiting big data, and a drive towards a winner-takes-

all competition (Croxson et al., 2021; Rohn et al., 2021; Ruutu et al., 2017). The South 

African stockbroking industry is dominated by traditional banks, offering stockbroking 

as an additional service. Banks and financial institutions have globally adopted a 

platform-based business model (Croxson et al., 2021). The South African stockbroking 

industry should adopt a platform-based business model to remain competitive. The 

conceptual framework in Figure 2.5 outlines the key themes and concepts that have 

surfaced in the literature review. At the same time, the analytical framework (Figure 

2.6) expands on the PT and variables that will influence and lead to a customer 

choosing a monthly subscription fee or not. 



 

31 
 

 

Figure 2.5: Conceptual Framework (Source: Author’s compilation) 

 

Figure 2.6: Analytical Framework (Source: Author’s compilation)  

 

 

 

 

 



 

32 
 

  



 

33 
 

Chapter 3: Research Methodology 
 

3.1 Research Approach 

This research adopted a case study approach from a quantitative perspective to 

address the primary research objective, namely: to assess the viability of alternative 

revenue models for stockbroking in South Africa. According to Creswell and Creswell 

(2018), quantitative research involves using numerical data to explore relationships 

between variables and develop and test theories (pp.38-47). The author used 

secondary data from a South African bank to analyse historic brokerage data and then 

mathematically model over a newly proposed subscription or fixed monthly pricing 

model.  

The aim was to closely match the revenue generated from the traditional variable 

transaction-based model with a new fixed monthly subscription-based model. In this 

case study, the author adopted a positivist approach. This approach enabled a 

scientific and objective data analysis, essential for replication and generalisability, but 

which disregards subjectivity. This research has relied on systematic data collection, 

hypothesis testing through experimentation, and verification of results in alignment 

with the positivism paradigm (Grant & Giddings, 2002). 

3.2 Research Design  

This case study analysed the relationship between brokerage fees and independent 

variables such as time periods, trading frequency, and transaction value. The author 

used a cross-sectional design when performing descriptive and inferential analysis on 

data collected over 10 years from 2013 to 2023 in a South African bank. This case 

study followed a time-series design based on secondary data with no influence. Using 

correlation-based analysis to draw causal inferences with the help of logic assisted 

with comparing the actual observations and drawing conclusions (Edmondson & 

Mcmanus, 2007). This effectively addressed the primary aim of exploring alternative 

revenue and business models better suited to a digital trading environment. This 

research has encompassed experimental and descriptive research designs while 

concluding with a time-series analysis for completeness. Essential concepts focused 



 

34 
 

on included validity and whether the research set out to measure what it intended to 

accomplish (Field, 2018, pp.15-16). 

3.3 Data Collection Methods 

A South African bank granted the author access to historical client brokerage and 

transaction data for stockbroking clients and brokerage fees over 10 years for the 

purpose of analysis. The dataset included actual brokerage data from client 

transactions, including trade execution time and frequency. All available brokerage 

data was initially extracted from a database running an SQL query, extracting available 

data in an Excel file for upload and evaluation in the IBM Statistical Programme for 

Social Sciences (SPSS) program. Brokerage fees per transaction were the primary 

variable extracted. The data was extracted and analysed in secured environments with 

no distribution or access to anyone other than the author. 

3.4 Population and Sampling 

 

3.4.1 Population  

This case study was conducted on a South African bank focusing on its stockbroking 

division. Client characteristics included registered stockbroking accounts where clients 

executed shares themselves through the online platform (directly on the JSE while 

excluding ring-fenced clients belonging to third parties). This client database consisted 

of 334,000 online transactions starting from January 2013 and ending on 8 September 

2023. The 10-year historical period aligns with when the business started growing its 

online stockbroking functionality and the financial industry became more digitally-

driven and dependent. 

3.4.2 Sample and Sampling Method  

The author used the entire population as the sample, as the dataset was readily 

available for analysis upon extraction. The author implemented a census sampling 

design for the population of 334,000 transactions over the period to ensure precision, 

eliminate sample error, and explore relationships within subsets. Potential concerns 

with the dataset included issues of quality and population biases. The nature, 

quantum, and volatility of the daily trade data collected from 2013 to 2022 influenced 



 

35 
 

the decision to group the data into yearly intervals and apply a cross-sectional design 

when analysing descriptive and inferential statistics. 

3.5 Descriptions of Variables  

The Chi-square test was used to measure the association between the dependent and 

independent variables and the difference between their expected and observed 

frequencies.  

Chi-squared: 

  𝜒2 = ∑
(𝑂−𝐸)2

𝐸

𝑛
𝑖=1 ……………………………………………………………….Equation 2  

Where: 

𝜒2 is the test statistic, 

O is the observed value, and 

E is the expected value. 

Foundations of Prospect Theory: Understanding Client Choices 

Exploring the Equation of Value Determination: Where V represents the ultimate value 

selected by clients, each decision is influenced by the probability (p) and the weighted 

decision factor π(p). Under the condition that p + q equals 1, the equation incorporates 

the relationship x > y > 0 or x < Y < 0.  

V(x,p:y,q) = v(y) + π(p)[v(x) – v(y)]…………………………………………Equation 1 

Hypothesis 2: 

There is no difference for South African stockbroking clients when choosing a payment 

option. 

Using Equation 1 and adding a riskless component (remaining on the same brokerage 

rate) and a risky component (changing to a new subscription model) will predict a 

determinable value in the client’s view. A new subscription-based model was based on 

the historical secondary data collected and analysed. Based on the client's current 



 

36 
 

brokerage rate, PT values each hypothesis, and the client chooses the outcome with 

the highest value (V). 

H2a: Stockbroking clients will prefer to pay a fixed monthly subscription. 

H2b: Stockbroking clients prefer to pay a transaction-based fee per transaction.  

 

Using the prospect theory (PT) discussed in Chapter Two, clients could choose 

between a new subscription fee or the old transaction-based model.  

Based on PT, clients will choose the highest value (V), and each choice is the 

probability (p) and weighted decision 𝜋(𝑝) if p + q = 1, then x >y>0, or x<Y<0 (see 

Equation1). 

Prospect theory is still considered one of the best measures when applying risk-based 

decision-making to determine an outcome (Barberis, 2013). Decisions for 

stockbroking clients were based on wealth or a cost reference point and were either 

accepted or not, depending on the new subscription-based fees (Edwards, 1996).  

Prospect theory has limitations, as there is no precise answer as to what comprises a 

gain or a loss; the exact weightings are not defined (Barberis, 2013). Despite this 

limitation, a decision can be made once it is clearly defined where the greatest value 

lies for each outcome. 

 

This research tested the following hypotheses:  

The first hypothesis is tested using inferential analysis by way of conducting a Chi-

square test using yearly data groups with a 5% significance level against the observed 

p-value. 

Null Hypothesis 1: There is no relationship between time periods, trading frequency, 

transaction value, and their effect on brokerage rates. 



 

37 
 

The second hypothesis is tested by building a model that predicts fixed future 

brokerage fees at an individual client level and then compares those against historical 

actual fees.  

Null Hypothesis 2: There is no difference to South African online stockbroking clients 

when choosing a payment option. 

H2a: Stockbroking clients will prefer to pay a fixed monthly subscription. 

H2b: Stockbroking clients prefer to pay a transaction-based fee per transaction.  

3.6 Procedure for Data Collection 

  

Stockbroking client data was originally saved on an underlying database immediately 

after each transaction. This research extracted stockbroking data by providing specific 

search criteria and downloading an Excel file through a front-end management portal. 

Permission to access and use the data was granted by the South African bank, with 

specific exclusions followed. This historically saved brokerage data was both cost-

effective and easy to obtain. This data analysis used a licensed version of IBM SPSS 

statistics, version 28.0.1.0 (142). 

3.7 Data Analysis Strategies and Interpretation  

Descriptive statistics was applied as a base, as it is helpful in summarising data and 

describing the sample. However, it does not follow causal analysis (Fisher & Marshall, 

2009). The distribution was determined according to different categories of time-series 

analysis, specifically linear regression and the auto-regression integrated moving 

average (ARIMA) model, which could present past observations, time points, and 

errors as a linear combination, with non-linearity being a weakness (Lee & Tong, 2011). 

A case study by Eisenhardt (1989) focused on a single setting, with the research 

steered towards that setting while understanding the various dynamics present. The 

following steps were taken in this order. 

Firstly, the data was cleaned, and the dependent and independent variables were 

extracted. The historic brokerage data was analysed and compared using descriptive 

statistics to determine central tendencies and variances over time. After identifying 



 

38 
 

consistent patterns in historical data through time-series analysis methods such as 

linear regression and the ARIMA model, which specifically consider seasonality and 

transaction size during high trading volume months, a distinct model using the most 

suitable approach was developed. The hypothesis was tested by proving the null 

hypothesis that there is no relationship between time periods, trading frequency and 

transaction value and their effect on brokerage rates. A Chi-square test of association 

was used to understand any association between the selected two variables, namely, 

dependent and independent variables. For the Chi-square model formula for 

descriptive and inferential analysis, see Equation 2. 

Secondly, exploratory data analysis was performed on the already cleaned and 

analysed data to statistically model a new fixed monthly subscription model based on 

historic fee data. Additional variables, such as periods, trade activity (frequency), and 

market (economic conditions), were added to enhance the predictiveness of the 

model. ARIMA was applied to the model. Data was split into historical data and training 

sets for validity and generality tests, with periodicity of one-year intervals. The model 

was adjusted to accommodate the historic transaction-based fees. Once the model 

functioned as needed, it generated predictions for subscription fees. Model validation 

was performed by testing against a random selection of executed trades, ensuring 

accuracy and reliability. Time-series analysis was used to predict and forecast future 

values from the collected data. 

Finally, using the PT discussed in Chapter Two, the client had a perceived loss or gain 

when choosing between traditional transaction-based fees or the newly modelled 

subscription-based fees. Figure 3.1 gives a graphical depiction of the outline. 



 

39 
 

 

Figure 3.1: Diagrammatic outline of the modelling (Source: Author’s compilation) 

3.8 Possible Limitations and Challenges of the Study 

  

I. Measurement on longitudinal case studies – varying time points can affect 

comparability.  

II. Generalisability – even though it is a census case study sample (i.e., entire 

dataset obtainable used), these findings will be applied to the broader 

environment and industry. 

III. ARIMA requires data to be stationary, requires parameter estimates, and the 

data must be sufficient and of high quality to drive the right results. 

IV. If the initial hypothesis of a relationship between dependent and independent 

variables was false, indicating no association between them, would digitisation 

be a more significant motivator for online stockbrokers to adopt a platform 

model? 



 

40 
 

3.9 Quality Assurance  

Analysis in the real world has brought about numerous complexities previously tested 

in perfectly controlled environments (Lincoln & Guba, 1986). The following quality 

assurances should provide reassurance that this research has considered its quality 

and is aware of the rigour required with its implementation and analysis.  

3.9.1 External Validity  

The research attempted to be generalisable and applicable to various scenarios while 

attempting to provide precise results on historical data with accurate future trends. The 

linear regression model and ARIMA model were used in order that comparisons could 

be made. The concern with the ARIMA method’s inability to model non-linear data may 

pose a problem; however, linear regression proved to be more accurate and was used 

to complete the new model. 

3.9.2 Internal Validity  

  

Using a positivism approach, the author detected causal relationships in the data. 

Establishing a causal relationship with the historical dataset and the methods 

employed might be of concern; however, it was addressed by collecting data at varying 

time points, which solved the potential issue.  

3.9.3 Reliability  

The author ensured the findings could be replicated and dependable, providing 

consistent results19. This study used multi-item scales to ensure reliability. Cronbach’s 

alpha addressed reliability with its function testing high commonalities, hence weak 

uniqueness. The function ranges between zero and one, with 0.7 and above being a 

good measure, indicating only a 30% potential error20. The reliability function of scales 

and Cronbach’s alpha requires a minimum of three items per scale. 

 
19 Sony, M. (2023). Research Theory and Design – Quantitive approach. Class slides and lecture 

notes 
20 Sony, M. (2023). Research Theory and Design – Quantitive approach. Class slides and lecture 
notes 



 

41 
 

3.10 Ethical Consideration  

Data integrity, protection and anonymity were of utmost importance. The sensitive 

nature of the analysed and observed data necessitated assurance that neither the 

bank's name nor the client names were displayed. 

A consistency matrix combines the research questions, its objectives and aligns the 

relevant data with the applicable analysis method in a concise table for easy 

comprehension. In the fourth chapter, the author shares discoveries and insights that 

correspond to the approach used in the consistency matrix methodology. These 

findings are further explored through in-depth discussions and interpretation. 

Table 1: Consistency Matrix, objectives, data detail and analysis method 

Main objective: Alternative revenue models in stockbroking in South Africa  

Objective 
Research 
Questions 

Data Detail 
Data 
Analysis 
Method 

Alternative 
Revenue 
Models 

Independent 
variable 
relationships with 
brokerage rates 

A case study 
using historic 
brokerage data 

Descriptive 
statistics 

Alternative 
Revenue 
Models 

Independent and 
dependent variable 
relationships with 
brokerage rates 

A case study 
using historic 
brokerage data 

Inferential 
analysis 

Platform 
Business 
Models 

Subscription-based 
fees 

Case study data 
Time-series 
analysis 

Client 
Acceptance 

Yes or no – risk-
adjusted choice 

Historic fees vs 
subscription-
based fees 

Prospect 
theory 

 

 



 

42 
 

CHAPTER 4: PRESENTATION, DISCUSSION AND INTERPRETATION 

OF RESULTS 
 

 

4.1 Introduction 

This chapter presents the results of the study in four parts to align with the consistency 

matrix presented in Chapter Three. Each of the four parts is presented with the 

observed results, interpretation, discussion, and summaries of the findings. The first 

section contains a descriptive analysis of the variables. The second part uses 

inferential analysis to determine whether the first hypothesis statements were met. 

The third section uses time-series regression analysis to determine a fixed 

subscription price. The last section presents ideas on whether clients would accept 

the new proposed fixed monthly subscription fee versus the existing variable-rate 

brokerage fees to answer the second hypothesis. 

4.2 Data Cleaning 

This secondary data, obtained for all client trades between 1 January 2013 and 8 

September 2023, contained numerous fields and a total data point count of just over 

11 million line items. Only the most important columns of data were kept, namely: client 

account number, transaction amount, brokerage amount, date, branch and deal 

sequence. All other columns were removed from the dataset. Transaction amounts for 

sales were originally presented as negative amounts, which were then converted to 

positive-only numbers using Excel's absolute function. Transaction amounts were also 

shown in cents and divided by 100 to convert into Rand amounts, giving a final 

transaction amount in Rands with only positive numbers labelled Absolute Tx amounts.  

The website client data was the only data used after filtering out other branches. The 

deal sequence line was used to show the number of individual trades, which 

accounted for the total execution amount per deal per client. The brokerage amount, 

i.e., the fee charged by the online stockbroker, was captured in cents form and kept 

as such. All fields were limited to two decimal places. This data cleaning left 334,000 

online stockbroking client deals from 1 January 2013 to 8 September 2023. 



 

43 
 

4.3 Descriptive Analysis and Interpretation  

This descriptive analysis groups the extensive daily data into yearly intervals starting 

in January 2013 and ending in December 2022. A yearly cross-sectional view 

separately analyses the variables for annual independent observations. 

To remain thorough and not be repetitive, the author has not interpreted and discussed 

all 10 periods, but instead focused on at least two tear intervals starting with periods 

for 2013, 2015, 2019, 2020 and 2022 to ascertain the impacts of COVID-19. Both 

absolute transaction amounts (Absolute Tx amounts) and brokerage amounts have 

been analysed. Transaction amounts were broken down into seven intervals, while 

brokerage amounts were broken down into five intervals. The volatile nature of the 

daily trade data collected required the data to be grouped into the most popular 

intervals. Observations were produced for each nominated year for transaction 

amounts and brokerage data, with results presented in either a table or figure. 

4.3.1 Time Pe