Windward's Marc Chalfin and Jay Upadhyay describes the overhaul and new vision at Groupon $GRPN
Summary
Windward argues that Groupon ($GRPN) is a damaged but recoverable local-services lead-generation platform whose post-COVID collapse reflected an extraordinary supply-demand inversion. Gross profit had held near $1.25 billion annually from 2015 through 2019, but reopening left consumers “demand starved” and flush with stimulus while merchants lacked workers and capacity. A salon booked for weeks would not discount a $100 massage to $50 and then net roughly $35 after Groupon’s cut: “The cure was worse than the disease.”
Prior management turned that operating shock into a liquidity crisis by retaining a bloated cost structure and failing to push out its debt. Negative EBITDA combined with an approximately $300 million working-capital outflow as old merchant payables exceeded new billings, draining cash from roughly $400–500 million to below $100 million. The resulting covenant breach and going-concern warning further deterred merchants and prospective enterprise partners from trusting Groupon.
The bull case rests heavily on Pale Fire’s alignment, turnaround record, and willingness to rebuild the company in public. Pale Fire accumulated roughly 22%, while CEO Dušan Šenkypl left a multibillion-dollar investment firm, accepted no salary, and received equity awards beginning at $14.86 and extending through $82. Management has removed approximately $800 million of costs; Chalfin believes that makes $200–300 million of EBITDA achievable without anything “Herculean.”
There are tangible operating improvements, but repeated technology failures have obscured them. Marketing payback reportedly fell from 18 months to seven days, North America Local grew for three consecutive quarters for the first time in eight years, and active customers grew for two; an anti-fraud integration then blocked checkouts in March and April, while the cloud migration disrupted traffic attribution beginning in July. Groupon expected its average Google rank to improve from roughly sixth to fourth, but the new site initially fell to seventh or eighth instead.
Andrew Walker’s strongest pushback is that Groupon’s inventory and targeting still feel visibly broken. Manhattan inventory appeared sparse, while ads served him restaurants near Seattle, Chicago, Miami, and Washington despite his being in New York. Chalfin flipped that criticism into the opportunity—Šenkypl said inventory was “not even 80% of where I want it to be”—while Upadhyay pointed to new merchant attribution tools, Starbucks activity, and SiteMinder’s Expedia inventory integration.
At roughly a $400–450 million market capitalization, Windward sees valuation, asset monetization, high short interest, and operating leverage combining into an unusually convex setup. Chalfin estimates Groupon’s sub-2% SumUp stake could be worth at least $125 million, versus the company’s earlier approximately $90 million expectation for SumUp and Giftcloud together, and normalizes current-year EBITDA from $70–75 million to about $100 million after roughly $30 million of identified disruptions. Add almost 50% short interest and prospective buybacks, and “the upside here is explosive” if billings turn.
The thesis has a clear expiration window: Groupon must demonstrate durable improvement from late summer through year-end. Upadhyay would reconsider if peak-season traffic were still declining high single digits or worse and North America Local billings were not growing, calling that outcome a “perpetual melting ice cube.” Chalfin allows for the logged-out customer cohort through June and a difficult June comparison, but says that without a material rerating from July through December, “we’re probably wrong.”
Deep dive
1. Groupon’s stable marketplace was broken by an extraordinary reopening mismatch
Chalfin defines Groupon simply as a lead-generation tool for local service businesses: a barber discounts a $100 haircut to $50, Groupon takes roughly one-third, and the merchant receives about $35 in exchange for acquiring a customer. That produces a high-margin, negative-working-capital business when the marketplace functions normally.
The historical stability matters more than Groupon’s shrinking headline revenue: gross profit ran around $1.25 billion in each year from 2015 through 2019. In 2021, investors assumed an 80% recovery toward that base, attached an eight-times multiple to roughly $150 million of prospective EBITDA, and drove the stock rapidly toward $60 before the anticipated recovery failed.
Chalfin’s reopening explanation: in some places, consumers had gone 18 months without haircuts, massages, or restaurants and were flush with stimulus, while many service workers never returned. Merchants became “short supply, long demand” in an intensely inflationary environment, eliminating any reason to discount full-price capacity through Groupon.
Management compounded the macro shock by leaving costs intact. With EBITDA negative and billings falling, Groupon still owed merchants for older, larger transaction volumes; working capital consumed roughly $300 million, cash plunged from approximately $400–500 million to below $100 million, and a current revolver plus breached covenants produced the damaging going-concern designation.
2. Pale Fire supplied unusually aligned operators to a distressed asset
Windward began buying just below $3 after studying Pale Fire’s earlier 13D and settlement with Groupon. Pale Fire’s partners had become billionaires through European e-commerce turnarounds, and the group took roughly a 22% stake at $1.90. Chalfin viewed Šenkypl leaving a multibillion-dollar fund to run a roughly $100 million market-cap company as a powerful signal: “That to me is very interesting.”
Šenkypl receives no salary and must create substantial equity value to realize his performance awards, whose disclosed thresholds include $14.86, $20.14, $31.01, and a top tranche at $82. Windward says it seeks at least a 5-to-1 risk-reward and reports a 76% hit rate over six years. Windward itself owns about 2.2 million shares—nearly 6%—and calls Groupon its most asymmetric idea in almost 25 years.
The first achievement was removing approximately $800 million of costs from a business once producing $1.25 billion of gross profit and now running below $500 million. Chalfin’s conclusion is that Groupon does not need a heroic revenue recovery to regain $200–300 million of EBITDA.
3. Better incentives and technology created green shoots—and new outages
Windward’s reference calls found that roughly 80% of legacy inventory was not what customers wanted, while commissions paid to salespeople could exceed the gross profit generated by their deals. The turnaround opportunity therefore includes better merchant selection and salesperson incentives that reward economically useful inventory rather than merely filling the site.
Marketing has shifted from broad “treasure hunt” acquisition—with one dollar of gross profit taking 18 months to repay—to bottom-up targeting such as offering a Virginia-bound family a Busch Gardens deal. Management says payback is now seven days and increased marketing from the mid-20s to the mid-30s as a percentage of gross profit in June.
The underlying architecture comprised around 100 technology stacks that struggled to communicate; Groupon could not readily add video or let Spanish-speaking customers toggle languages. Management rebuilt the front and back ends and migrated to the cloud—“open-heart surgery while we’re public”—while North America Local grew for three quarters and active customers for two, each for the first time in eight years.
The turnaround then encountered two major execution issues. A third-party anti-fraud installation prevented willing customers from checking out during March and April; beginning in July, cloud-related instability impaired attribution to Google and other traffic partners just as business and marketing accelerated. A subsequent SEO reset demoted Groupon’s new pages, including on more than a third of traffic; management had to work through tens of thousands of landing pages and redistribute them to Google, a process Chalfin described as about 60 days.
4. Weak inventory is both the central objection and the largest operating option
Walker tested the product in Manhattan and found thin inventory plus almost comically irrelevant advertising: restaurants outside Seattle, Chicago, Miami, and Washington. His challenge was concrete—“It looks great in a spreadsheet”—but the customer-facing marketplace still appeared far from delivering the targeting and selection underpinning Windward’s projections.
Chalfin agrees with the observation but reaches the opposite conclusion: if Groupon can produce current results with poor inventory, better supply becomes incremental upside. Šenkypl told him, “Our inventory is not even 80% of where I want it to be,” with improvement expected through 2025; the team is “drinking out of a fire hose and doing a hundred different things at once.”
Upadhyay adds that merchants previously needed a Groupon employee to obtain ROI reporting because they could not deploy their own attribution tools. The consolidated backend makes enterprises more willing and able to integrate directly, with Starbucks cited as a recent example and SiteMinder feeding Expedia travel inventory into Groupon.
5. Headline traffic can conceal improvement in the profitable local business
Windward tracks Similarweb, credit-card data, and the site itself, but Chalfin warns that Groupon trades like “a vehicle for alternative data” even when that data is misleading. As much as one-third of off-peak traffic can come from casual Goods browsing, although Goods contributes zero EBITDA and has reportedly been falling 40–50%.
That mix alone can create roughly a 12-point total-traffic headwind even if Local performs materially better. Conversion improvements also create a spread between traffic and billings, so Chalfin argues that undifferentiated web visits cannot reliably measure the part of Groupon that matters economically.
Seasonal periods provide more useful tests because the average user uses the site only about 2.4–2.5 times annually. Evidence cited included approximately 50% year-over-year Halloween traffic, low-teens Christmas traffic, and low-single-digit North America Local growth from Black Friday through Cyber Monday. Valentine’s Day, Mother’s Day, and the June–August things-to-do season are forward-looking signposts.
6. Valuation offers several ways to win, but execution must close the gap
Chalfin frames Groupon at roughly a $400–450 million market capitalization, with no net debt expected pro forma for fourth-quarter reporting. He estimates its just-under-2% SumUp holding is worth at least $125 million, citing a Reuters article saying Goldman Sachs was shopping a sizable SumUp stake worth about €400 million at a valuation above €8 billion; Chalfin thinks Groupon is probably part of that process.
Walker presses the discrepancy with Groupon’s earlier expectation of approximately $90 million from SumUp and Giftcloud together. Chalfin attributes that disclosure to going-concern conservatism and depressed European payments valuations, adding that SumUp subsequently grew sharply; nevertheless, the $125 million remains Windward’s estimate, not a disclosed sale price.
Windward expects current-year EBITDA around $70–75 million despite approximately $30 million of anti-fraud and web-stability damage, implying roughly $100 million normalized. Chalfin says the guidance was conservative because of a convertible offering and concerns about debt holders, and expects more than $75 million of free cash flow.
Further levers include another $50 million of legacy costs, purchase frequency rising from roughly 2.4–2.5 toward five, with each additional user-engagement turn worth about $100 million of incremental gross profit, checkout shrinking from 12 steps to nine, alternative payments, video, and higher conversion from existing traffic.
The larger bridge is deliberately expansive: recovering to two-thirds of 2019 activity could support roughly $400 million of EBITDA under the new cost base, while Slevomat’s precedent suggests gifting might become meaningful—potentially $200 million of incremental gross profit in Chalfin’s framing. International comparisons are also easing, with performance already improving excluding Italy.
7. The turnaround becomes falsifiable between late summer and year-end
Walker invokes the grim base rate for legacy consumer platforms such as Tripadvisor, QV—which he says rhymes with QVC—and Yelp: investors repeatedly see low multiples and optionality, yet the businesses keep deteriorating. Chalfin’s rebuttal is that Groupon is “spring-coiled” at roughly one-third of its former gross profit, with far more low-hanging fruit than a platform already operating near historical levels.
Chalfin would not use the next three months as the verdict because the logged-out-user cohort from the site transition affects comparisons through June, which itself had double-digit growth last year. His real test is July through December: absent a material rerating—or if technology suffers another major failure—“we’re probably wrong.”
Upadhyay’s end-of-summer test is more operational: after Valentine’s Day, Mother’s Day, and the June–August attractions season, traffic should not still be falling high single digits or worse, and North America Local billings must grow. Failure on both measures would suggest Groupon is “just too difficult” to restore and probably remains a “perpetual melting ice cube.”
Timing is central to the trade because almost 50% short interest, a potential SumUp sale, and buybacks could amplify any inflection. Chalfin also argues that Groupon has no real competitor and could benefit from a recessionary environment. He warns against waiting for perfect reported confirmation—the stock can, in his words, “rip $67 in a week” when it starts to improve—while Walker reminds him that a similarly promising setup before the third quarter was derailed by an unexpected reversal.