Computing Profitability with GroupOn – Relevant Costs and Cognitive Bias

Plug for my book: To Group Coupon Or Not: Small Business Guide to Groupon, LivingSocial and Others is now available.

Utpal Dholakia, Marketing professor at Rice published his results from his study, “How effective are GroupOn promotions“.  A key finding from the study is, 66% of the businesses that used GroupOn promotions reported the promotions were profitable. Let us note that (and the paper clearly points that out as well) this was measured based on survey response of business owners.

The reality could be much different from this, not just due to sampling error but due to incorrect relevant cost estimation and cognitive biases of these business owners. Note: Spreadsheet link at the end of the post.

Relevant Costs: In my previous articles (and here) I wrote about the different cost components of running a GroupOn promotion. These are not only specific to GroupOn but are true for any new promotion that is being considered. The promotion must deliver incremental profit over all its relevant costs and this profit must be better than any other option available to the business. In general, most small businesses without rigorous cost accounting are most likely to ignore:

  1. Opportunity costs: This is forgone profits  from other means to drive revenues-  including price realization, increasing customer margin, new product introduction and  other promotions.
  2. Lost profits: These are the lost profits from sales not made to current and new full price customers. Businesses may forgo these sales because their capacity is all consumed in serving the promotion traffic.
  3. Incremental costs: These are the costs incurred to add people and capacity just to serve the new traffic

Of these costs, the first two do not incur any cash outflow and hence are most likely to be ignored, “if we did not have it or spend it then we did not lose it”. But these costs are very relevant to evaluating the profitability of running the promotion and hence must be covered only by the revenues from the promotion.

In my email conversation with Utpal, he mentioned that the study did not take into account these costs (quoted with permission)

all of the things you mentioned are definitely true.. since we asked them only whether the promotion was profitable or not, they might or might not have factored in these more subtle aspects of costs

Cognitive Dissonance and Commitment Bias

You may have seen this on TV reality game shows or post-game interviews of players of losing teams.  They take huge risk but come out with very little to show for. A common quote from people who do not make it to the next round is, “It is great coming this far. We did not think we would be here when we started this. I would not exchange this for anything else”.

What is going on here is that losing team suffers from cognitive dissonance – the conflict between what we want vs. the real outcome. We take an action expecting one type of results but the reality turned out to be different. Since we can’t change the reality we change our mind by saying, “this is great, this is what actually we wanted, we are in fact better than what we thought we would be”.

In the case of business owners who reported the promotion was profitable, they may very well be addressing their cognitive dissonance. They made a conscious decision to run a GroupOn promotion. To say now that the move was not profitable is admitting their original decision to run the promotion was a mistake.

After the fact, people significantly understate their original target. Compared to the reduced targets, any revenue from a 50% promotion may look like a success.

There is also the mental compulsion to continue to act in accordance with our previous action. Psychologists call this, Commitment Bias. Once they run one promotion, the businesses are more likely to repeat the promotion because they are compelled to act in line with their previous action.

This is not to say no business made a profit but only to point out that the number 66% has more than sampling error in it.

When is GroupOn promotion profitable?

  • You have a product with high contribution margin (price less marginal cost)
  • Have excess capacity (with sunk costs and no other way to monetize it)
  • There exists a segment of customers with low willingness to pay but reducing the price to include them will deliver less profit than your current profit (even though it is still profitable)
  • There exists a segment of customers with low willingness to pay that you cannot reach through any other way
  • You can serve these low willingness to pay customers without the full price customers knowing about it (third degree price discrimination)

If any of these are not true you have hard math in front of you (Spreadsheet).

Does Presence of Customer Reviews Drive Down Product Returns?

Does presence of customer reviews and the number of reviews drive down returns by customers?

According to Internet Retailer (thanks to Gerardo for the link), that is the case. The article says, reviews has helped Petco considerably

Petco’s approach to gaining more customer reviews has paid off. On average, products with reviews have a 20.4% lower return rate than products without reviews. The return rate continues to decline as a product gains more reviews. Products with more than 50 reviews have a 65% lower return rate than products with no reviews.

Since returns eat into profits, reducing returns goes directly to the bottom line – there is no question here. But can presence of reviews drive down returns? Is there a direct causal relation or is this just incidental correlation.

Commitment and Consistency Bias: If the  case of customers who took the time to write reviews I can see that their return rate will be much lower than the return rates among those who did not write one. This is the Commitment and Consistency bias (the book Influence by Robert Cialdini has very good discussion of these biases). When the customer “commits” by writing how good they feel about the product their internal system compels them to act consistently to their previous commitment. So they keep the product.

Reason doesn’t matter: This does not mater whether or not customers wrote the review because of their LOVE for the product or because they were paid in coupons or raffles. This does not apply to negative reviews, but according to one research, most reviews are positive reviews and there is generally high product ratings. On the other hand we could argue that those who returned the products are more likely to write a negative review.

Conformity Bias: Commitment and consistency bias alone cannot explain the drop in returns because this is still a small number of reviews compared to products sold. But another cognitive bias that could be at play is conformity bias. When customers make the purchase based on many reviews by “customers just like them”, they tend to confirm to those peer reviewers. This will compel them to “like” the product and keep it – all those positive reviews cannot all be wrong, if I do not like the product it must be me.  Again, Cialdini has chapters describing Conformity bias in his book.

Cognitive Dissonance: Intertwined with conformity bias is our need to assuage cognitive dissonance. People who buy a product by doing the research, reading multiple reviews and evaluating options believe they made a rational decision of buying the best possible product. But after buying the product if it turns out that it did not live up to their expectation they  suffer cognitive dissonance – a gulf between how their feelings before and after the purchase. Customers overcome that by convincing themselves that they like the product.

On top of these cognitive biases, it is possible that there exists another common variable that both drives up number of reviews and drive down returns – for example the product experience matches its promise.

There is one way to answer many of these questions and to find out whether or not number of reviews drive down returns. It requires doing two sample test, showing some of the customers the review,  suppressing it for others and tracking the respective return rates. If the return rates are statistically significant then we can declare presence of reviews drives down product returns.

Next step, if we do the experiments by showing different number of reviews we can even find the linear causal relation between number of reviews and returns.