When evaluating company documents and filings, a hedge fund manager friend of mine has a simple mantra: "read the footnotes". So when the Groupon S-1 was posted last night, I was eager to pour over it and find a few hidden nuggets that might explain the company's success and provide some indication of what the future might hold.
I think I found it on pages 74 (not the footnotes, I grant you, but buried quite deeply in the 110 page document). The company presents two case studies for its two oldest cities, Chicago and Boston, and provides a few datapoints that allows one to infer two of the most important metrics for the business – customer churn and customers per merchant.
Let's look at the longest-tenure city – their hometown of Chicago. Here's the chart they provide in the S-1 for performance over time:
All the numbers are growing, which is terrific. But what matters in these situations is the ratios of growth, not just the absolute growth. So let's do some simple calculations. First, subscriber (defined as someone who receives an email from them) growth per quarter and percentage growth:
As you can see, quarter over quarter subscriber growth is slowing considerably in Chicago. In Q1'10 it was 81%. In Q1'11 subscriber growth was only 37%. This is incredibly important because there are going to be inactive subscribers – also known as attrition or churn – and that number will likely grow over time. Groupon needs to fill this leaky bucket with new subscribers every quarter. If not, growth will slow, flatten or decline.
Now, let's look at customer growth. Typically, I would expect the definition of a customer to be someone who purchases a Groupon in that period. But a footnote on page 8 provides a different definition. A customer is defined as anyone who has purchased a Groupon since January 1, 2009. That is, it is a cumulative figure representing anyone who has purchased in Groupon's history. Thus, the change in customers is far more important than the absolute number. Here's what that looks like:
As with the subscriber numbers, this figure is slowing in quarter-over-quarter growth. In Q1'10, the change in cumulative customers (that is, new customers) was 69%. In Q1'11, it was 35%.
Now let's look at things from the merchant's view. The number of Groupons sold per merchant and the number of Groupons sold per customer per merchant are two important ratios of merchant success and customer activity. Here's what that looks like:
As you can see, these ratios are also in decline, with Groupons per merchant per 1000 customers dropping 10x from Q3'09 (21.3) to Q1'11 (2.3). This ratio is a good proxy for how active the cumulative customer base is. The lower the ratio, the more likely there are large pools of inactive customers. Again, if customer engagement attrits over time, then there is a leaky bucket that can only be filled with new customer acquisition. In saturated markets, it is typical for new customer acquisition to slow down and become more expensive over time.
To be clear, I'm incredibly impressed with this company and what it has achieved. But a careful read of the rations is instructive when thinking about persistence and durability of the model. Groupon has so many assets at their disposal and is so early in their innovation cycle, that I'm sure we haven't seen anything yet in terms of targeting, loyalty and other techniques to provide sticky customer relationships. But the ratio data above suggests that the basic model they're executing on right now has some weaknesses that will become more acute over time.