Perion Network 2025: the platform is broadening, but search still sets the pace

Getting to Know the Company

Perion at the end of 2025 is no longer just a search distribution business with an AI layer on top. 79% of revenue now comes from Advertising Solutions, and the growth in CTV, DOOH, and commerce shows there is a broader ad-tech platform here. But anyone looking only at the Perion One rebrand can miss the real issue: in 2025 the company still depended partly on a shrinking search base, and the hole left by Microsoft Bing was larger than what the new growth engines could yet fill on their own.

What is working is easy to see. CTV revenue rose 42%, DOOH grew 36%, and Retail Media & Commerce topped $109.9 million. TAC fell from 57% to 54% of revenue, so a slightly larger share of every sales dollar stayed in the model. The company also ended the year with $312.9 million in cash, cash equivalents, short-term deposits, and marketable securities, with no financial borrowings. This is not a balance sheet under funding pressure.

What is still messy is the income statement. Total revenue fell 12% to $439.9 million, search revenue dropped 44% to $91.0 million, selling and marketing expense rose 12% to $76.5 million, and the company moved from $12.6 million of net income to a $7.9 million net loss. In other words, Perion has shown that it can build a broader platform, but it has not yet shown that this platform can replace the old search engine at the level of clean profitability.

That is why 2026 looks less like a breakout year and more like a proof year. For the read to improve, Perion needs to show four things at once: that search stabilizes at a smaller base under Yahoo, that CTV, DOOH, and commerce keep growing without relying on easy comparisons, that Perion One starts to create real operating leverage, and that Greenbids and Outmax move from strategic framing to visible economic contribution.

There is also an important market screen here. On April 3, 2026, the TASE line traded at NIS 30.67, implying a market value of roughly NIS 1.2 billion based on year-end shares outstanding. Local short-interest data does not show an extreme bearish position, but it does not show blind confidence either. This is not a market treating the story as broken, and it is not treating it as proven. That is exactly why the next few quarters matter so much.

The most important conceptual point is that the accounting is already unified, while the economics are not. Perion now reports as one operating segment and one reporting unit. For this company, that makes the channel disclosure more useful than the headline segment reporting, because the real economic gap is between the new ad-tech engines and the shrinking search business.

Perion's Economic Map

73% of revenue still comes from U.S. end users, so this remains heavily tied to the U.S. advertising market. At year-end, Perion had 511 employees and 104 contractors, down from 528 employees a year earlier. That means the company has already cut layers, but revenue declined faster than headcount.

Events and Triggers

First trigger: the launch of Perion One on February 3, 2025. This is more than a branding change. Perion is positioning itself as an execution layer, with Outmax as the AI-native engine meant to manage budgets, pacing, and allocation in real time across CTV, DOOH, walled gardens, commerce, and the open web. The problem is that the organizational and accounting shift has already happened, while the operating leverage that is supposed to justify it has not yet fully arrived in the P&L.

Second trigger: the end of the Bing tail period. The legacy Microsoft agreement expired on December 31, 2024, and 2025 was the tail year. That sits at the center of the 44% decline in search revenue. Starting in 2026, search is expected to operate primarily with Yahoo. This matters for two reasons. First, the revenue base is already much smaller. Second, the company says nearly all search revenue in 2026 is expected to come from the current provider. So concentration is lower at the total-company level than it was in 2024, but potentially higher inside the remaining search business.

Third trigger: the Greenbids acquisition in May 2025. Strategically, it makes sense. Greenbids brings AI-based bidding optimization for walled gardens and major DSPs. Economically, the picture is more demanding. The transaction value was $49.6 million, including $30.0 million in cash and $19.6 million of fair-valued contingent consideration. On top of that, there are $15.0 million of retention arrangements in cash and RSUs, which run through compensation expense rather than purchase consideration. And despite all that weight, Greenbids remained small enough that management's internal-control assessment excluded it, while its 2025 revenue contribution was only 0.6%. That is the key point: strategically meaningful, financially still early.

Fourth trigger: the buyback. The board expanded the authorization to $200 million during 2025, and by year-end the company had executed $118.1 million in total, including $71.2 million during 2025 alone. That reduced shares outstanding by 12.9% versus the end of 2024. It supports the per-share picture, but management also explicitly ties part of the decline in financial income to lower cash balances after the buyback and the Greenbids acquisition. So the program helps ownership concentration, while also reducing some of the balance-sheet protection the company had enjoyed.

Efficiency, Profitability and Competition

What Actually Improved

The clean positive in 2025 is better mix. Cost of revenue plus TAC fell from 66.7% of revenue to 65.5%. That does not look dramatic in isolation, but when total revenue is down 12%, every point matters. It means the newer channels, especially CTV, DOOH, and commerce, are taking a bigger share of the business and bringing somewhat healthier economics with them.

That is why Advertising Solutions still grew 4% even as the group contracted. CTV rose 42%, DOOH rose 36%, and Retail Media & Commerce grew 36% to more than $109.9 million. Management also says DOOH now represents 22% of display advertising revenue. That is a sign the newer engines are no longer optionality, they are already part of the real business mix.

There is also some evidence of cost discipline outside go-to-market. R&D fell 5% to $34.7 million, and G&A fell 6% to $36.4 million. Management ties this to the Perion One unification, lower payroll and consulting costs, and higher software capitalization in 2025. So some efficiency gains are beginning to show up.

Where the Leverage Is Still Missing

This is where the yellow flag appears. Better mix has not yet translated into cleaner profitability, because selling and marketing is rising too fast. S&M reached $76.5 million, or 17% of revenue, up from $68.5 million and 14% of revenue in 2024. That is the cost of repositioning the company around Perion One, but as long as it stays at this level, it is difficult to argue that the unified platform is already producing operating leverage.

Labor productivity has not improved enough either. Headcount fell from 528 to 511, but revenue per employee dropped from about $944,000 to about $861,000. That is a simple but useful reminder that the company has streamlined more slowly than revenue has fallen. The unification effort matters, but it has not yet shown up as a clear productivity gain.

It is also worth paying attention to how Perion defines its competitive position. It describes itself as a neutral, cross-channel execution layer. Conceptually, that is attractive. In practice, this is still a very hard market. The company itself lists competitors such as The Trade Desk, Zeta, Criteo, PubMatic, Magnite, and Teads, as well as giants like Google, Meta, Amazon, and Microsoft, many of which are simultaneously partners and competitors. So the real question is not whether Perion can tell a strong infrastructure story. It is whether it can defend pricing power and margin quality while the most powerful platforms still control the core ecosystems.

Technically, the move to one segment and one reporting unit makes the Perion One story cleaner. Analytically, it also makes life harder for investors. The company gives less direct separation between the new growth engines and the old search engine just when that separation matters most. Here, the channel note is one of the most important windows into the business.

Cash Flow, Debt and Capital Structure

First the normalized picture, then the all-in one

The right way to read Perion's 2025 cash flow is to hold two cash bridges at once. The first is normalized / maintenance cash generation. Operating cash flow rose to $41.9 million from $6.9 million in 2024. If you deduct reported capex of $3.8 million, capitalized internal-use software of $1.9 million, and lease cash payments of $1.1 million, you get roughly $35.1 million. That is a reasonable view of what the existing business generated before acquisitions and buybacks.

But that is not the full story. This cash flow did not come from clean operating earnings alone. The bridge includes $31.1 million of stock-based compensation and $17.7 million of depreciation and amortization. At the same time, accounts receivable increased by $23.0 million to $187.9 million even as revenue fell, and the allowance for credit losses more than doubled from $2.5 million to $6.4 million. So the cash improvement is real, but it is not pristine.

The second frame is all-in cash flexibility. Once you take operating cash flow of $41.9 million and deduct $3.8 million of capex, $1.9 million of capitalized software, $1.1 million of lease cash, $26.6 million of net cash paid for Greenbids, and $71.2 million of share repurchases, Perion ends up with an all-in cash picture that is negative by roughly $62.7 million. That is not a liquidity crisis. It is a clear capital-allocation choice. The company bought strategic time and per-share support at the cost of part of its cash cushion.

The balance sheet is still strong, but no longer unlimited

Even with that outflow, the balance sheet remains solid. Perion ended 2025 with $90.0 million of cash and cash equivalents, $151.0 million of short-term deposits, and $71.9 million of marketable securities. That is enough to support operations and absorb volatility. But the strength is no longer costless. There are also $27.7 million of acquisition-related obligations and $22.4 million of lease liabilities on the balance sheet. And financial income fell by $8.6 million, partly because the company is earning less interest on a smaller cash pile.

The buyback changed the equity picture materially. Shares outstanding fell from 44.8 million to 39.0 million. That helps the per-share setup, but it also reduces flexibility if 2026 requires more commercial investment, more M&A, or simply more patience. Management has clearly decided the stock was cheap enough to justify aggressive repurchases. That can prove correct, but it raises the price of any execution misstep in the next year.

Outlook and What Comes Next

First finding: 2026 is a proof year, not an automatic recovery year. The comparison base is already lower after the search collapse, so nominal growth could look good quickly. What matters is whether that growth comes from the new engines or merely from stabilization after a difficult year.

Second finding: Perion One has already changed the way the company reports, but it has not yet proved it can change the economics quickly enough. The move to one segment and one reporting unit makes the story cleaner, but S&M shows integration still costs real money.

Third finding: Greenbids is strategically more important than financially material today. It added $35.5 million of intangible assets, $19.0 million of goodwill, $20.1 million of contingent consideration liability at year-end, and, together with Hivestack, drove $7.6 million of retention-based compensation expense in 2025. That is a lot of accounting and economic weight for a business that contributed only 0.6% of revenue in the year.

Fourth finding: search may be smaller, but it has not stopped being critical. Microsoft represented 23% of revenue in 2024. The current provider generated 16% of revenue in 2025, and the company says nearly all 2026 search revenue is expected to come from it. Perion has effectively swapped one concentration point for another, just on a smaller base.

Fifth finding: the balance-sheet story still depends on execution. At year-end, Perion carried $266.1 million of goodwill and $89.2 million of intangible assets, together 38.9% of total assets. In the annual impairment test, fair value exceeded carrying value by only 17%. That is not an immediate problem, but it is not an especially wide cushion either.

What has to happen in 2026

Perion needs to prove the new business is large enough to stand without the old one. In practice, that means CTV, DOOH, and commerce need to keep growing at a healthy pace, Web needs to stop shrinking fast enough to erase the mix benefit, and the Yahoo transition needs to happen without another meaningful leg down in search.

At the same time, the market will want evidence that operating leverage is finally getting closer. If Perion One is more than a narrative, selling and marketing cannot stay at 17% of revenue indefinitely while group revenue is still shrinking. Greenbids also needs to start showing value beyond strategic language, not necessarily as a huge immediate revenue source, but at least as something that improves margin quality or customer economics.

What could break the thesis

The biggest risk is that search keeps falling faster than expected while the newer engines continue to grow but not at a scale large enough to offset it. In that scenario, Perion could end up with a better mix on paper but weaker economics in practice. The second risk is that the commercial and integration cost of the new platform stays elevated even after the transition year. Then the story shifts from "investment before leverage" to "a structurally heavier cost base."

The market will also keep testing cash quality. Operating cash flow of $41.9 million is clearly better, but if the bridge still leans too heavily on stock-based compensation, depreciation, and working-capital timing in 2026, it will be hard to argue that true profitability has caught up with the operating narrative. Perion does not need to look perfect next quarter, but it does need to show the gap between story and economics is narrowing.

Risks

Partner risk: even after leaving Bing behind, search has not become diversified. The company states that nearly all 2026 search revenue is expected to come from the current provider. Any policy, pricing, or traffic-quality change there could cause another shock.

Structural search risk: Perion itself says AI-based discovery tools are reducing traditional search volume, page views, and clicks. This is no longer a theoretical warning. It was visible in the 2024 and 2025 numbers. If that trend accelerates, search could shrink faster than the new engines can compensate.

Regulatory and technology risk: Perion operates deep inside the privacy, data, AI, and cybersecurity stack. Management describes a more complex regulatory environment across the U.S., Europe, and Israel, and also acknowledges that the spread of AI creates new attack vectors. For an ad-tech platform, that is not side noise. It is both an operating cost and a competitive risk.

Balance-sheet and accounting risk: goodwill and intangibles account for nearly 39% of total assets. The company recorded no impairment in 2025, but it is not sitting on an unlimited valuation cushion either. If execution under Perion One disappoints, the market can quickly move from debating the story to questioning the quality of the balance sheet.

Legal risk: the securities class action and derivative action tied to search disclosures and Microsoft Bing are still alive. The U.S. appeal remains pending, and the Israeli proceeding is stayed pending the U.S. outcome. Even without a quantified expense today, this remains part of the overhang.

Working-capital risk: receivables rose, the credit-loss allowance jumped by more than 150%, and that is a reminder that digital ad growth does not automatically convert into cash. If payment quality weakens further, part of the recent cash-flow improvement could reverse.

Short Interest Read

The local short-interest data does not currently signal a collapse trade. Short interest as a percentage of float stood at 0.50% on March 27, 2026, almost exactly in line with the sector average of 0.51%. At the same time, SIR was 3.34 versus a sector average of 1.157. So this is not a stock with an extreme short consensus, but it is also not one trading with frictionless confidence. During the first quarter, short interest rose to a peak of 0.71% of float and 4.97 days to cover on March 20, before easing back. That looks more like caution around a transition year than a one-way bet against the company.

Conclusions

Perion ends 2025 with a much better strategic story than it had two years ago, but with results that show the transition is not yet complete. The newer engines, especially CTV, DOOH, and commerce, are already large enough to change the mix. The main bottleneck is that search is still shrinking quickly, while the commercial cost of the new story is still too heavy. Over the short to medium term, the market is unlikely to reward more AI framing on its own. It will want numbers that show the company can stabilize revenue, preserve cash generation, and bring operating leverage back.

Current thesis: Perion is moving from a search-centered story toward a multi-channel ad-tech infrastructure story, but 2025 shows that the economics of the transition still lag the narrative.

What changed versus the old read: search used to be almost the entire story, now it is "only" 21% of revenue, but it still explains most of the pressure in the numbers.

Counter-thesis: the new engines may still be too small, and search could keep shrinking faster than CTV, DOOH, and commerce can offset it, leaving Perion with a strong AI narrative but a business that still struggles to restore earnings power.

What could change the market read in the short to medium term: quarters that show a steadier search base under Yahoo, continued double-digit growth in video and out-of-home, and a meaningful decline in selling and marketing as a share of revenue.

Why this matters: with Perion, the key question is no longer whether the company has technology. It is whether it now has a revenue and profit structure that can live without the old search support.

In simple terms, the next 2 to 4 quarters need to show three things: search does not fall again, the new engines keep delivering growth, and the cost ratio starts to come down. What would weaken the thesis is the opposite mix, another leg down in search, persistently heavy selling expense, and Greenbids remaining mostly a strategic story with a very real economic price tag.