Multisided Platforms: Classification and Analysis

Round 1

Reviewer 1 Report

Submitted manuscript presents an interesting approach to analyze and understand network effects on multi-sided platforms. Developend and presented theory is inspired by the Bass model to formally measure and describe complex interactions and network effects on multisided platforms.   The authors have identified the following research gap: none of the previous scientific contributions and papers address temporal evolution of platform services using analytic or system dynamic methods.

The problem description is not fully convincing and remains rather vague. What is the tacked problem here, to apply the Bass model is interesting, but why does it show potential to be applied to the field of multi-sided platforms. 

There is no discussion of alternative approaches existing. Furthermore, simulations and game theory is no referred to nor is it argued why the chosen approach is chosen or adequate for the addressed problem. Is there further empirical work, if yes, it should be mentioned or reviewed.

The research is from my perspective relevant and constitutes an interesting way to shed light into complex interactions and behaviours in relation to platform-based business models. However, the theoretical foundation of multi-sided platform remains rather vague and is not sufficiently well reflected and reviewed. The authors refer to Parker et al. however they do not detail the platform model of the authors, for example participants, value unit, filter results in core interaction. The underlying concepts remain rather weak and are not defined in reference to state-of-the-art literature or review. Role models such as complementors or boundary resources are not discussed or mentioned why they can be excluded from the analysis.

The Bass model and formula is introduced but it is not convincingly shown where and how the model is improved (partly only parameters or variables in the functions are renamed). The original model uses density function and cumulative probability, thus it would be useful to display the density function and cumulative probability to ease comprehension and to highlight where the paper delivers new insights in relation to this approach. The imitiation and innovation coefficients are not further detailed in relation to internal and external influence as in the original model. For me the empirical data used refering to examples such as Facebook are not transparent and it is difficult to judge its reliability and if the data to calculate the coefficients are appropriate.  From my perspective, the papers tries to tackle to many different examples, but does not one in adequate detail, thus I would recommend to have less examples but more detail (especially in relation to applied data). On page 12, figures are shown, visualizing the data. What is the software used here? What is the data used for the visualizations visible and how can it be reviewed? How is it possible to reproduce the shown figures and results? An interesting aspect is the link to design parameters of multi-sided platforms. The target group of the research is not clear, because the data is only of value, if it can be used to make decisions and improve designs and influencing factors of platform solutions.

Overall the authors introduce various formulas and equations, but do not argue or show proof that the used equations are mathematical correct. Often solutions of the equations are presented, without indicating or proofing why this is claim is true. Here additional sources are required, or mathematical proof has to be presented, e.g. as appendix.

The concept service remains as well rather vague from a theoretical point of view. Overall the model appears too simplified and is rather oriented towards a good-oriented logic and not a service-dominant logic, where experience or relationship, resource integration and linkages play a crucial role. Respective literature is not reviewed and adequately discussed.

The results are not discussed and the limitations of the research is not presented or reflected. Finally, the conclusion and summary is rather short. It is recommended to discuss outcomes from a practioner viewpoint. What can platform owners learn from the paper.

Author Response

Dear Reviewer.

We thank you for your comments and suggestions. The comments below follow the order in which they appear in the review.

1. Based on the feedback from the reviewer, the introduction has been extended to include the list of aspects studied in the paper:
   • proposal to extend the Bass equation to incorporate cross-side feedback term – an approach not found elsewhere in the literature. The Bass equation with this extension can then be used to analyse MSPs. The rational for the extension is elaborated in Section 3.1. The extension is natural and is equivalent to numerous examples in physics where theories are extended in the same way (one of the authors is a physicist by training).
   • to study the initialization problem and why it exists, in particular the chicken and egg problem.
   • to show that strong cross-side and same-side network effects may lead long latency periods and is, hence, a major strategic problem.
   • to show that, for the same reason, the market growth follows an S-curve.
2. Alternative approaches are now more explicitly mentioned in the conclusion of the paper: System dynamics and agent-based models. We have not found any examples in the literature where these methods have been used to analyze the dynamics of MSPs. The stochastic Polya urn method used by Arthur et al. to show that the evolution of even a simple technology may be subject to path dependence and thus end up in one out of several final states, is mentioned in Section 1. This method is mathematically too complex to be applied on MSPs and is therefore not mentioned further as an alternative approach.
3. The rationale behind the extension of the Bass equation is mentioned in point 1 above. The alternative to the approach by using probability density function and cumulative probability is to set up the equations in the “physicist’s way” as is done in this paper. There is nothing new to this. The empirical data are used to suggest parameter values that can be used in the example calculations – they have no other significance and are not used for other purposes. The purpose of the paper is not to analyse a few examples in depth but to show simple methods to analyse all seven possible two-sided platform configurations. Our hope is that other authors will use these methods for detailed analysis of each possible configuration.
4. More details concerning the solutions of the differential equations is not deemed to be necessary for two reasons: (i) the details not included in the derivation of the solution are simple and is found in any undergraduate textbook on calculus and (ii) the purpose of the paper is not to show detailed derivations but to focus on the results.
5. As mentioned in the introduction, the models are kept simple and does not include elements mentioned in the review. Including such elements would have complicated the mathematical models beyond simple analysis.
6. The conclusions of the paper have been extended to contain more precise summary of the paper.
7. The reference list is rather short because there is so little literature (we found none) on the temporal evolution of MSPs.

Reviewer 2 Report

The reviewed article is devoted to the system analysis of the IT platform market. The authors describe the essence of interaction between IT market participants using MSP-platforms.

The presentation of the material is clear and clear. All necessary links are provided. The algorithms and methods used to construct the classification are very well described. It is very appropriate to use the apparatus of differential equations to construct mathematical models and explain the real processes of interaction between market agents.

Remark 1. In my opinion, it was possible in conclusion to indicate that the approach can be applied to most marketplaces at the regional level (China, Russia, etc.)

Author Response

Dear Reviewer

We thank you for your comments and suggestions. Here follows some comments from our side, and how we have changed the paper according to the reviewer's suggestions

1. The approach can be used in any market. There is no need to indicate this explicitly in the conclusions.
2. The introduction has been extended to include the list of aspects studied in the paper:
   • proposal to extend the Bass equation to incorporate cross-side feedback term – an approach not found elsewhere in the literature. The Bass equation with this extension can then be used to analyze MSPs. The rational for the extension is elaborated in Section 3.1. The extension is natural and is equivalent to numerous examples in physics where theories are extended in the same way (one of the authors is a physicist by training).
   • to study the initialization problem and why it exists, in particular the chicken and egg problem.
   • to show that strong cross-side and same-side network effects may lead long latency periods and is, hence, a major strategic problem.
   • to show that, for the same reason, the market growth follows an S-curve.

The modelling principles are made clearer.

The conclusions have been expanded to summarize the results better.

3. The reference list is rather short because there is so little literature (we found none) on the temporal evolution of MSPs. However, we have now added six more references to the paper which are discussed in the introduction and related works part of the paper. 

Reviewer 3 Report

The reviewed article is focused on the overview of the different types of multisided platforms distinguished by network effects within and between groups. The article could be considered as an original, well structured, however the moderate improvements should be applied.

The initial part, dealing with the literature review should be improved, with wider study of related literature sources. The limited initial literature review is visible also in other parts, when only overall 16 references for whole article were quoted. Similarly, research design should be improved with appropriate formulation of research question/hypotheses. After that, current methodological part could be applied. For absence of the research question/hypotheses the following article´s section is a little bit complicated for the reader to follow main research idea. Subsequently, after the modifications of individual parts, it will be appropriate to supplement the part devoted to the findings and formulated conclusions.

Author Response

Dear Reviewer.

We thank you for your comments and suggestions.

We have included an additional 6 references, which are discussed as a part of the introduction or as a part of the related works. 

The introduction has been extended to include the list of aspects studied in the paper:

• proposal to extend the Bass equation to incorporate cross-side feedback term – an approach not found elsewhere in the literature. The Bass equation with this extension can then be used to analyze MSPs. The rational for the extension is elaborated in Section 3.1. The extension is natural and is equivalent to numerous examples in physics where theories are extended in the same way (one of the authors is a physicist by training).
• to study the initialization problem and why it exists, in particular the chicken and egg problem.
• to show that strong cross-side and same-side network effects may lead long latency periods and is, hence, a major strategic problem.
• to show that, for the same reason, the market growth follows an S-curve.

The conclusion of the paper is also revised to more clearly show the contributions of the paper. 

Round 2

Reviewer 1 Report

The manuscript has considerably improved and it now easier to follow the authors arguments and development steps. More background information has been added and the various steps are now better justified. My recommendation would be to add a section discussion as well as limitations, because the coefficients and parameters have been estimated or assumed. Alternatively, the authors should indicate that they only intend to demonstrate the validity of proposed approach to explain network effects. Contributions should be devided and further detailed in practitioners and researchers. The model intends to be used for design tasks (decision on business model dimensions and to make forecasts, probably as well post analysis and reviews of developments based on empirical data (usage data). My suggestion would be to move some of the details in section "Analysis" into the appendix. One example would suffice and then refer to the annex. Essence of this analysis should then be used to have more details on discussion (from practitioner and research point of view), as well limitations. Summary and conclusion look fine.

Author Response

Dear reviewer

Thank you for your comments. We have now added more discussion on the parameters and how they are used in the introduction. We also discuss the limitations of these parameters, and that are primarily used to test the models. 

The main contributions of this paper is the developed analytical models in chapter 5. We belive this content constitute the major part of the paper, and therefore would not move it to the appendix. If doing so, we believe the paper need to completely change focus from a theoretical to a more applied type of paper. 

Thank you.

Reviewer 3 Report

Thank you for additional article modification.

Let me recommend the possible minor improvement of two parts, current state of the art description and methodological part.

As concerns the description of current understanding of research topic, it could be based on a wider literature review, with appropriate research gap identification. At this moment research gap is not clear, for this reason presented research steps are not clear too. As concerns the second part for possible improvement - it could be still improved methodological part. Let me propose to clearly describe the research goal on the page 2, the lines 59-64. The modification of current text will be enough and helpful for easier research goal identification.

The article should clearly describe the current research gap and the researchers´ contribution to cover this gap, presented in the article. At this moment is not clear, what are the standard procedures, generally known, and what are the researchers´ contribution.

Thank you.

Author Response

Dear reviewer

Thank your for your comments.

We have added some more text explaining the research gap, goals, and contribution in the introduction of the paper. We have also added a new reference on Polya Urn models, and briefly discussed how this paper relates to the use of Polya Urn models, which is an alternative approach to describe such markets. 

Thank you.