When a fired Snapchat employee sued the company earlier this year, claiming it misled investors and the public at large, the world mostly yawned. The shareholders who bought Snap stock after its giant IPO last month seem to have cared particularly little about the lawsuit.
Now, investors have reason to pay attention. Snap (snap) released previously redacted portions of the ex-employee’s complaint this week, saying it “has nothing to hide” as it had already revealed much of the once-confidential information when it became a publicly traded company. Still, the unsealed lawsuit contains some key statistics that Snapchat has never before disclosed, not even to investors.
And while there are some doubts about the credibility of Snapchat’s accuser, Anthony Pompliano, who worked at the company for all of three weeks before Snap says he was “fired for poor performance,” he was right about one important thing: Around the time Snap began selling ads on the platform, Snapchat was not as popular as most people thought it was—including the company itself. Some of Pompliano’s other allegations could also cast suspicion on whether Snap’s growth is healthy enough to justify the unprofitable company’s lofty valuation.
The crux of Pompliano’s suit alleges that the company inflated some of its numbers, telling advertisers and investors it had surpassed the coveted milestone of 100 million daily active users in the summer of 2015, when in fact the real tally was closer to 95 million or less. Snap vociferously denies misrepresenting its user metrics, and Pompliano, who was hired to help the company grow its user base, doesn’t offer hard evidence to corroborate his assertion.
Yet what Snap does not dispute is that the discrepancy underlying Pompliano’s argument actually existed. His claims revolve around a divergence between two different systems for measuring active users, Flurry Analytics and another one called Blizzard. Flurry vs. Blizzard may sound more like a milkshake contest between McDonald’s and Dairy Queen than a variation in computer programs. But ultimately, Snap discovered the same problem Pompliano did: that Flurry, which is owned by Yahoo (yhoo), was overstating the number of users, counting people multiple times if they returned to the app from different devices in the same day, and recording as an active user someone who merely received a push notification without ever opening Snapchat.
Snap says that Pompliano’s lawsuit overstates the significance of that discrepancy. A former Facebook (fb) employee who is described in the lawsuit as a “decorated war veteran” who served in Iraq, Pompliano goes by the nickname “Pomp,” according to his LinkedIn and Medium profiles. And though he claims that he tried to be a whistleblower about the unreliable metrics, Snap tells Fortune that it had already switched to a more accurate, in-house method of counting users by the time Pompliano arrived at the company in August 2015, and in no way did he inspire the change.
In fact, Snap’s IPO prospectus has a special section explaining that transition, which led the company to restate its daily active user figures from before mid-2015, lowering them by almost 5%. But investors who bought into Snap at its IPO received only the corrected information, (The timeline isn’t the only thing Pompliano got wrong: his lawsuit also misspells the name of Snap’s CFO, Drew Vollero, instead referring to him as Drew “Boller.”) “Put aside for the moment that this musty, two-year-old allegation about a minor metrics deviation hardly measures up to Pompliano’s gasping rhetoric about Snap being built ‘on a house of cards,'” Snap’s lawyers wrote in a legal filing when they unsealed the complaint. “The more important point is that the allegation is false.”
Still, it’s easy to see why people might have believed Snap had reached the 100-million mark long before it was true. A cloud of hype and high expectations has clung to Snapchat ever since it rejected a $3 billion takeover bid from Facebook (fb) in 2013. As early as 2014, The Wall Street Journal reported that Snapchat, then valued at nearly $10 billion, had more than 100 million users, but that was on a monthly basis; the number who were active daily was significantly lower. And in the spring of 2015, Snap CEO Evan Spiegel announced at a Re/code conference that his disappearing message app was “approaching 100 million daily active users in developed markets.”
For investors, though, whether or not Snap exaggerated the user number matters less at this point than the fact that Pompliano’s lower number turned out to be right. That’s because his lawsuit also points to several other Snapchat metrics which, if true, could change the way investors value the company.
For example, Pompliano claims that he found “glaring inaccuracies” in statistics such as the percentage of people who begin the process of setting up a Snapchat account and also complete their registration: While Snapchat had said this figure was 87%, according to Pompliano, the real number was less than 40%.
Potentially more troubling, Pompliano alleges another discrepancy that could mean that Snapchat’s user engagement, which Wall Street generally considers to be unusually high, could actually be just average. In the world of mobile apps, when fewer than 25% of people ever return to one a second time, Snapchat was said to have a 48% user retention rate a week after the first use, greater than the rate for Twitter (twtr), The Information reported in 2015. A few years earlier, Spiegel swore the rate was even higher, telling a friend, according to Business Insider, “This thing is a rocketship…Our 7-day retention is 60% right now (target is 30%) and we’re growing.” Pompliano, however, says that while Snap was quoting 40%, the real retention rate after seven days’ time was “closer to 20%.”
A Snap spokesperson says the company disputes the accuracy of Pompliano’s figures in those two cases. But a source familiar with the matter says the metrics were only meant for internal use, not to be shared with investors, and that there is no universal or industry standard for measuring such types of engagement, so it’s possible that the disparities are just a matter of apples and oranges.
To Pompliano, however, they are symptoms of what his lawsuit calls “a wide-spread, systemic failure in Snapchat’s internal controls over its user data.” At the very least, they are a reminder to investors of the inconsistencies and ever-changing nature of the measurements being used to evaluate a company like Snapchat. After all, with an app whose photos and messages disappear, which, unlike Facebook and Instagram, doesn’t even show how many people have “liked” or viewed any piece of content, investors have little to go on but their assumptions when it comes to determining how engaged users really are.
And with Snapchat, those assumptions may be subject to change.
Snap stock fell 2% Wednesday.