A Procedure to Set Prices and Select Inventory in Thinly Traded Markets Using Data from eBay
Round 1
Reviewer 1 Report
In the introduction, please indicate a specific goal, possibly a research hypothesis. Provide an overview of the theoretical and empirical literature on the analyzed topic. Indicate which researchers have previously used similar tools, such as the authors of this article. What other research tools could possibly be used. At the end, please clearly present the answer to the set goal and the verification of the research hypothesis as well as the directions of further research.
Author Response
The introduction has been substantially modified to clarify the goals of the research. A conclusion paragraph was added at the end elucidating what we have done.
Reviewer 2 Report
Please find attached pdf file.
Comments for author File: 
Comments.pdf
Author Response
- When the authors are describing the benefits of this method, the authors are always comparing to historical data. You show your selected portfolio generates better profit compared to the entire portfolio and claim that this method is better. This doesn’t make sense to me. You are using full information after the sales are completed, it is natural to select a better portfolio compared to at the time before the sales.
The revised ms. now clarifies that we used historical data to develop a methodology to optimize the portfolio of items with low volume. The focus is on the method, and the portfolio improvement is a demonstration of the method.
- Moreover, how exactly are you going to do forecasting? Let’s assume I have a number of candidate items to sell, how are you going to use this method to select a portfolio? At the time of portfolio selection, you don't know how long the items will be listed before sale, what the selling prices are, etc. Basically, in order for the method to have practical use, all information you can utilize for forecasting must come from time before listing, if you add any information after the listing time, you can only use an assumption. I think the manuscript lacks consideration in this line.
The method we use is partially drawn from finance in which the underling properties of a new asset are unknown but can be estimated using historical data from similar assets. The practical application of this communication identifies which data one needs to forecast in order to choose a portfolio of items to sell to reach profit goals.
- Even if you can do a valid forecasting, how can you claim your portfolio is better without a control group? In a proper experiment design, you will need to randomly divide the candidate items into two groups. Your method is applied to one group and the other group uses the traditional method. And you track the performance of these two groups and compare the difference to see if they are statistically significant before making any claims. This part is not considered in the manuscript.
As in our reply to R1, our purpose here is to present a methodology to choose inventory for thinly traded markets. The comparison we make in the paper is to the store owner's "guesses" for what to carry and our systematic approach. We now clarify this in the Introduction and conclusion so that readers understand that this not an RCT.
- The introduction is not sufficient. How is thin market defined exactly? What are the methods investigated by others, pros and cons? Because of the uniqueness of items in thin markets, I feel it is hard to find one method that fits all. What is the scope of your method, does it work for real estate, antiques and art work? What is the common method used for auto parts? What is the issue with it? The research question is not properly set up.
More information was added in the introduction that seeks to address these points. The paper is a short communication so brevity is required.
- One thing I don’t understand about your proposal is that what do you propose for the items you don’t select into your portfolio. Are you suggesting not to list them or are you suggesting not to purchase them in the first place? If former, why not list? If latter then it is another question. Basically the question becomes when facing an item (thin market), you need to determine whether or not to purchase it, instead of selecting items from a pool.
The conclusion of the revised ms. now discusses that the methodology we developed using a cash constraint can be adapted to account for capacity constraints on warehouse space.
- The Poisson distribution is not clearly described. How exactly was lambda calculated? How can you validate this distribution assumption?
More explanation was added.
- There are grammar issues and typos in the manuscript, a technical editor would help.
Grammar issues were corrected.
Reviewer 3 Report
The paper "A Procedure to Set Prices and Select Inventory Thinly-Traded Markets Using Data from Ebay" covers a very interesting topic: price-setting in low-volume or "thin" markets. The structure is the usual one. It is a very synthetic paper, in which all the sections are very well written and have enough content. The quality of the English language is quite good, nevertheless I recommend a miner spell check.
Author Response
I recommend a miner spell check.
The entire ms was edited for clarity and a spell check was performed.
Round 2
Reviewer 2 Report
The abstract has some grammatical errors. I can see there are multiple grammatical issues even in the abstract. For example “prices can then assigned”, “applying it a store”. This indicates that there are probably more issues in the text itself. The manuscript should go through a technical editor. I browsed through author’s response, I do not have strong objection to have it published.
Author Response
We thank this reviewer for his/her patience with our poor writing. We have revised the entire ms for appropriate English grammar, clarity, and missing words. We believe this version is error free. The ms with tracking is attached.
Author Response File: 
Author Response.docx
