Fiverr Faces Proof Test On AI Traffic Drain After Slashing Its 2026 Outlook

Fiverr enters its third-quarter report with the hard part already announced. Three months ago management cut its full-year outlook, guided to a revenue drop of as much as 26%, and blamed a fresh slowdown in marketplace traffic on large language models and Gemini's arrival inside Google Search. The question for this report is not whether the quarter will be weak, since that is spoken for. It is whether the damage stopped where management said it would, or whether the AI-driven erosion of the top of the funnel is still outrunning the company's push upmarket.

Wall Street expects non-GAAP earnings of $0.36 per share on revenue of $84.5 million. That revenue figure would be down about 21.7% from $107.9 million a year ago and sits almost exactly at the midpoint of management's $80 million to $88 million range. Analysts have fully absorbed the reset, which leaves little cushion either way. The earnings estimate is more curious. It sits well above last quarter's $0.27 even as revenue is expected to fall by more than $13 million sequentially, which implies the Street is counting on aggressive cost control. The Earnings Whisper of $0.34 sits a bit below consensus, a sign that the more skeptical crowd doubts expenses can be cut fast enough to protect profit on a shrinking top line.

The full-year math also matters. Fiverr booked roughly $203 million in revenue in the first half, so the new $356 million to $372 million annual range implies a fourth quarter of only about $73 million to $81 million if the third quarter lands as guided. In other words, management is telling investors the decline gets worse before it gets better. Any change to that range will reveal whether the recent traffic deceleration has stabilized, deepened, or been overstated. A second cut would suggest the company still cannot see the bottom, while a reaffirmed range with third-quarter revenue near or above the midpoint would be the first evidence that the reset was conservative enough.

Beneath the headline numbers, the bullish case rests on mix. Spend per buyer rose 15.6% last quarter, and $1,000-plus projects in programming and graphics grew strongly. But growth in clients completing large projects slowed to 13% from 18%, and active buyers fell 22% to 2.7 million. If spend per buyer keeps accelerating while the buyer decline moderates, the upmarket strategy is gaining traction. If high-value growth keeps slowing, the transformation story loses its main support. Services is another tell: management now expects that former growth engine to exit the year shrinking by double digits, so further softening in Ads, Seller Plus and AutoDS would be a bad sign. Early read-throughs on the new fulfillment layer, the matching model that cut cancellations, and the TikTok, Amazon FBA and Shopify merchant campaign could help, though none is large enough yet to move revenue.

The market has delivered its verdict so far. Shares have fallen 8.9% since the last report, trailing the S&P 500 by nearly 13 points, and at $8.55 the stock sits far below its 200-day average of $11.12 and only about 3% above its post-earnings low of $8.31. The entire range since the last report sits beneath the prior quarter's trough. Sentiment is bearish and essentially unchanged from three months ago, so pessimism is steady rather than building.

Ultimately, this report hinges on one thing: whether management can show that the AI-driven traffic decline has found a floor. Hitting the guided range and holding the annual outlook would keep the extended, at-least-six-quarter turnaround plausible. Another downward revision would suggest the structural headwind is still setting the pace.

This earnings preview was originally published on Earnings Whispers.

This article is for informational purposes only and is not investment advice. Earnings Whispers and its staff may hold positions in the securities mentioned. It was generated automatically, with the help of AI, from Earnings Whispers and company data, and checked by Earnings Whispers before publication.