In October 2008, a Chicago startup sold two-for-one pizza vouchers at a bar downstairs from its office. Sixteen months later, Groupon was worth a billion dollars, and according to PitchBook, it is still the fastest any startup has ever reached that mark.
Then something quieter turned up in the IPO paperwork. Revenue per sales rep fell from roughly $172,000 in Q1 2011 to $138,000 in Q2 2011. One quarter. Same playbook, same pitch, less output per person.
The slide explains what happened to Groupon far better than the valuation ever did.
Groupon reported one side of a two-sided marketplace
The S-1 became famous for a single cohort chart. Groupon spent $18 million acquiring 3.7 million North American subscribers in Q2 2010, then pulled $61.7 million in gross profit out of that group over the next four quarters. Investors read it as proof the engine worked. Buy a customer cheap and monetize them for years.
There was no equivalent chart for merchants. Nowhere did the filing say, "Here are the businesses we signed in Q2 2010, and here is the share that ran a second deal." A marketplace with two customer types published a retention story for one of them.
Outsiders measured the supply side first
Utpal Dholakia at Rice University surveyed 150 businesses that ran Groupon promotions between June 2009 and August 2010. Two-thirds called the deal profitable. About a third lost money on it. More than 40 percent said they would not do it again.
His follow-up research, spanning close to 500 businesses across the deal sites, found 55.5 percent made money, 26.6 percent lost money, and 17.9 percent broke even. Around 36 percent of voucher buyers spent beyond the face value, and fewer than 20 percent returned to pay full price.
Andrew Mason had said publicly that 97 percent of Groupon's merchants asked to be featured again. The distance between that claim and the survey data is where the business quietly cracked.
The pitch sold customer acquisition. The product delivered a discount.
A merchant running a deal was buying new regulars. The price of that experiment was steep: half off the menu, roughly half of the remainder to Groupon, and the payout landing in installments over weeks.
Deal-seekers came once, tipped lightly, spent close to voucher value, and moved on to the next email. The owner absorbed the discount, the commission, the chaotic Saturday, and the staff burnout, then looked at the repeat rate and passed on deal number two.
Every city had a countable number of merchants
Consumers replenish. Local supply does not. Any metro area holds a finite list of spas, studios, salons, and restaurants with real spare capacity and enough margin to give half of it away.
Once a rep worked through that list, three options remained: sell to businesses that fit the profile poorly, push deeper discounts, or re-sign owners using deal cash to cover payroll. Each choice lowered deal quality, which lowered buyer satisfaction, which made the next merchant pitch harder to close.
That is what a supply-side CAC problem looks like before anyone names it.
Headcount masked the leak for two years
Groupon grew from 37 employees in June 2009 to 10,418 by September 2011, including more than 4,800 sales reps covering 175 North American markets and 45 countries.
Gross, merchant, Add looked spectacular. Over 66,000 featured merchants across 2010. More than 190,000 in the first nine months of 2011. Hiring faster than churn keeps a top-line curve pointing up, and it lets a company book new logos while the old ones quietly stop answering the phone.
The damage shows up in per-rep productivity long before it hits revenue, which is precisely what that $172,000 to $138,000 drop was flagging in the middle of an IPO roadshow.
Groupon was listed in November 2011 at a $12.65 billion valuation. Revenue peaked slightly above $3 billion in 2016. Active customers peaked above 50 million in late 2014.
Fifteen years later, the fix is the thing nobody measured
Groupon has guided to $513 million to $523 million in revenue for 2026. Active customers sat at 16.1 million in Q2, up 2 percent, while unit sales fell 7 percent as buyers shifted toward higher-value local inventory.
Under CEO Dušan Šenkypl and the team backed by Pale Fire Capital, the stock has climbed hard off its 2023 floor, and the current work reads like a supply-quality project: better local inventory, higher average order values, and fewer throwaway deals. The company is finally optimizing the side of the marketplace that broke first.
What founders can pull from this
As a founder, here are a few things you can pull from Group’s story:
1. Build cohort charts for both sides.
If you run a marketplace, marketplace-style pricing, or a partner channel, plot supplier cohorts the same way you plot user cohorts. Sign-up month on one axis, percentage still transacting on the other. Do it this week, even if the answer is ugly.
2. Treat output per rep as your early warning system.
Revenue per rep, deals per rep, and time-to-second-order fall before total revenue does. A sales team that is growing while per-head output shrinks is a team burning through finite supply.
3. Ask what your customer's second purchase actually requires.
Groupon's model needed merchants to profit on a discounted cohort, which required repeat visits the platform never provided. Write down the assumption that has to hold for reorder number two, then test it directly.
4. Count your total addressable supply, not just demand.
Demand-side TAM slides are easy. The harder number is how many qualified suppliers exist per city and how many stay after one bad experience.
5. Verify your own success claims from the outside.
A 97 percent satisfaction claim and a 40 percent never-again rate cannot both be true. Get an independent read on your churn before an academic, a journalist, or a short-seller publishes one for you.
Forward this to a founder who is celebrating gross adds and ignoring churn.
