1
$\begingroup$

Let's say I have two stores, A and B, and let's say I have two products in the same department: product 1 and product 2. Let store B have product 2 which store A doesn't have, both stores have product 1. Is there a way to calculate how much cannibalization will take place on product 1 by implementing product 2?

Or, can I do a general approach and use a chi square on their department to see if store B has a higher performance in that department and then if that's the only difference in the products offered, and if they're similarly performing stores that introducing product 2 will increase that departments performance for store A?

If you have some scholarly articles that would be great as well!

$\endgroup$
  • $\begingroup$ Your question seems to be asking multiple questions in one; and some of them are about inferring causality. Even if products 1 and 2 were in the same category, how do you know one cannibalized the other, as opposed to simple correlation or anticorrelation? Example: fishing rods do not cannibalize icecream sales in July, although they're both seasonal items. Can you ask more specific examples and post some example data? $\endgroup$ – smci Jun 10 '15 at 22:05
  • $\begingroup$ Right, sorry my question is vague. I recently started working as a data analyst for a company and I have tons and tons of POS data that I can analyze and I'm not even sure where to begin. I'm a graduate student right now but haven't had any real world experience. I guess at this point I want ideas on how to analyze this data and what conclusions I'll be able to draw from it. $\endgroup$ – Kristofersen Jun 10 '15 at 23:09
  • $\begingroup$ (getting offtopic a little, but) What is the most important business question they want you to answer? How to increase overall sales? sales in a certain product category? customer segment? store? geo region? How to segment the customer base? How to understand individual customer behavior? customer segments? How to prevent churn of loyal customers? What promotions/discounts/loyalty rewards you should offer to increase sales? Which competitors you lose sales to? If all else fails, plot them some graphs. $\endgroup$ – smci Jun 10 '15 at 23:31
  • $\begingroup$ Take a look at e.g. Kaggle Dunnhumby Shopping Challenge and Walmart Store Sales Forecasting challenges. $\endgroup$ – smci Jun 10 '15 at 23:32
  • $\begingroup$ Thanks smci, those look really interesting. I already have a lot of graphs showing how items sell at different prices, which items were sold with others, items sold the most, items sold the least. I just want to do something more insightful. The graphs I have are great at showing trends but I'm hoping I can do something to predict what will happen. Like predict how a departments performance will increase if a new item is added to the inventory. But I'm not sure how to do something like that. I can see how often it's sold at stores that currently have it but if it's added to the new store I'd h $\endgroup$ – Kristofersen Jun 11 '15 at 2:35
2
$\begingroup$

In terms of scholary articles I co-wrote a paper http://www.autobox.com/cms/index.php/news/131-102706-white-paper-on-cannibalization-qtesting-market-hypothesisq-by-john-c-pickett-david-p-reilly-view on this thorny subject which was subsequently printed by the International Business Forecasting Journal. Time series procedures (ARIMA , Transfer Function) were used to examine sales from two Staples stores and to test the hypothesis that sales for the new store cannibalized sales from the old store OR actually improved sales due to enhanced brand awareness.

| cite | improve this answer | |
$\endgroup$
  • $\begingroup$ @kristofferson You're a grad student. A grad student in what? Beware of suggestions that are software sales pitches, and not in your comfort zone or wheelhouse of expertise. They can take you down the garden of forking paths, so to speak. Consider this a research question and challenge where, as noted, you have to define the question or objective first. Go through it methodically, think rigorously. It's not an insuperable problem...use your education, training and instincts to work down to a solution. And, yes, read, read, read but read the technical literature after HBR case studies. $\endgroup$ – Mike Hunter Jun 11 '15 at 0:51
  • $\begingroup$ Hi Mike, I'm getting my MS in statistics. I'm new to statistics though. I did my undergrad in math so I don't have a ton of experience analyzing data, just the first year of grad school. The stuff I'm using is all locked in so no sales pitches. Just after ideas how to do something really insightful with this data. $\endgroup$ – Kristofersen Jun 11 '15 at 2:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.