Chapter 1

Bought Traffic

NerdWallet matches consumers with financial products and is paid by the providers. Revenue reached $836.6 million in 2025, up 22% [1], producing $96.0 million of non-GAAP operating income [2] and $113.4 million of cash flow after all software spending [3]. Half of that revenue went to buying the traffic that produced it, a share that has risen every year since 2022 [4] [5]. The market values the whole company at roughly $584 million.

Sources: FY2025 Form 10-K, Consolidated Statements of Cash Flows and Non-GAAP reconciliation [6] [7]; share count from the Q1 2026 Form 10-Q balance sheet [8]; price as reported.

What the company is

NerdWallet runs a personal-finance website and app. Its editorial team publishes reviews, comparison tables and calculators covering credit cards, banking, insurance, loans, mortgages and investing; readers who arrive for the guidance are routed into marketplaces where they can compare and apply for products [9]. The company earns nothing from the reader. It is paid by the financial institution on the other side — per click, per lead, per completed application, or per funded loan, depending on the product [10]. It operates in the United States, Canada and the United Kingdom, and also owns regulated subsidiaries that broker mortgages and provide advisory services [11].

For most of its life the acquisition engine was free. NerdWallet ranked at the top of Google for the queries people type before they buy a financial product, and the audience arrived at no marginal cost: 19 million monthly unique users on average in 2021 [12], 23 million in 2023 [13]. The company has not disclosed the metric in either the 2024 or the 2025 Form 10-K.

Tim Chen, who co-founded the business in 2009, is Chairman and CEO. He and his affiliated trusts hold every one of the 31,685,652 Class B shares, which carry ten votes each, plus 1.5 million Class A shares — 50% of the economics and 90.72% of the vote [14]. The practical consequences run both ways. At $8.88 his stake is worth about $295 million, against total 2025 compensation of $1.54 million — a $700,000 salary and a performance bonus, with no stock or option award at all [15]. He is paid by the share price, not by the grant. Equally, no outside party can force a sale, a board change, or a change of strategy on him. NerdWallet is a "controlled company" under Nasdaq rules and is exempt from several governance requirements as a result [16].

The Class A shares listed on Nasdaq on 4 November 2021 and closed that first day at $28.30 [17]. They closed at $8.88 on 24 July 2026. A $100 investment at listing was worth about $49 at the end of 2025, against $151 for the Nasdaq Composite over the same period [18]. No dividend has ever been paid [19].

Source: daily closing prices through 24 July 2026, as reported; the 4 November 2021 first-day close of $28.30 is confirmed in the FY2025 Form 10-K stock performance graph [20].

Where the revenue comes from

The revenue mix has been rebuilt in two years. Credit cards — the highest-margin, most organic-search-dependent category, and the largest single line as recently as 2023 — has fallen from $209.7 million to $133.4 million, a 36% decline that the company attributes squarely to "continued pressures in organic search traffic" [21] [22]. Insurance has gone the other way, from $45.0 million in 2023 to $280.8 million in 2025, driven by auto insurers expanding acquisition budgets [23] [24]. Insurance was 8% of revenue in 2023 and 34% in 2025. It is also the one category paid per click rather than per completed application or per funded loan, and clicks are the unit of volume that can most readily be bought.

Source: FY2025 Form 10-K, Note 2 Revenue [25]. The 2023 split shown here is the company's restated presentation, which carves Insurance out of the Emerging verticals line originally reported for that year [26].

From the first quarter of 2026 the company stopped publishing this split. Revenue is now presented in two lines, Consumer and SMB, on the basis that it matches how management reviews the business [27]. Insurance, credit cards, loans and everything else consumer-facing now sit inside a single $198 million quarterly aggregate [28]. An outside analyst can no longer see the auto-insurance line separately, in the quarter management first flagged a problem with it.

The cost of the traffic

Management has been explicit about the cause. On the fourth-quarter 2025 call, Chen said that in 2025 "we faced headwinds as consumers increasingly turn to AI overviews and large language models over traditional search, resulting in steep organic search declines," and that growth in performance marketing, direct and non-search referral channels "more than offset the declines in organic search" [29]. The 10-K risk factors carry the same point in formal language: reliance on Google is a named dependency, and "the introduction and acceptance of AI-assisted technologies could further impact search engine relevance, causing declines in our ranking and decreased platform traffic" [30].

What the substitution costs shows up in one line. Performance marketing — money paid to acquire visitors — was $168.4 million in 2022 and $416.9 million in 2025 [31] [32]. Subtracting it from revenue gives the money the business keeps before it pays for anything else — editorial, engineering, brand, overhead.

Source: derived from reported revenue and performance marketing expense, FY2023, FY2024 and FY2025 Forms 10-K [33] [34] [35].

Reported revenue compounded at 15.8% a year from 2022 to 2025. Revenue after the cost of buying traffic compounded at 4.2%. In 2024 that second figure went backwards, from $392.9 million to $390.2 million, while headline revenue rose 15%.

The first quarter of 2026 sharpened it. Revenue of $222.2 million was up 6% year over year; performance marketing of $121.7 million was up 25%, to 82% of all sales and marketing spend and 55% of revenue [36]. Revenue after that cost was $100.5 million against $111.6 million a year earlier — down 10% while the headline grew.

Sources: performance marketing expense, FY2025 Form 10-K [37]; FY2024 Form 10-K [38]; Q1 2026 Form 10-Q [39]; 2022 and 2023 revenue from the FY2023 Form 10-K [40]; percentages derived.

The strongest fact against reading this as deterioration is that total marketing intensity has barely moved. All sales and marketing was 67% of revenue in 2023, 70% in 2025, and 67% in the first quarter of 2026 [41] [42]. What has changed is the split inside marketing: brand spend fell from $106.3 million in 2022 to $65.1 million in 2025 [43] [44], and in the first quarter of 2026 non-performance marketing dropped from $62.1 million to $27.4 million, principally because the company did not repeat the prior year's Super Bowl advertisement [45] [46]. Performance marketing is also genuinely variable — the filings state plainly that "we are able to adjust our marketing spend to reflect changes in external factors and consumer behavior" [47]. A company that can turn the tap off is not trapped by it.

The reading that fits the evidence best is that both things are true: NerdWallet is converting bought traffic into profit today at a rate that is measurably improving, and the pool of free traffic underneath it is shrinking, so each incremental dollar of revenue costs more to source than the last. The trade is only worth making while the spread holds. A brand-spend cut that funds one year of operating leverage cannot fund the next.

What is left over

Whatever the mix, the profit is real cash. Non-GAAP operating income — stated after stock compensation and after deducting the $20.4 million of software the company capitalized rather than expensed, so nothing is added back twice — doubled to $96.0 million in 2025, with adjusted EBITDA of $145.0 million [48]. Operating cash flow was $131.6 million against $16.9 million of capitalized software and $1.3 million of equipment, leaving $113.4 million of genuine free cash flow, up from $50.5 million in 2024 [49].

Two cautions on the reported profit. GAAP net income is distorted by tax: 2024's $30.4 million of net income sits on $5.0 million of pre-tax income, courtesy of a $25.4 million tax benefit, while 2025's $48.7 million follows a $19.5 million tax charge [50]. And the headline free cash flow figure quoted in some data feeds treats capitalized software as free — it is not; the $113.4 million above deducts it.

The balance sheet carries no funded debt. The $125 million revolver was undrawn at both year-ends, and the only borrowing is a $6.9 million warehouse line at the mortgage subsidiary, secured by the loans it funds [51]. Cash was $56.3 million at 31 March 2026, down from $98.3 million because the company spent $66 million on its own shares and $17 million on an acquisition in a single quarter [52] [53]. Repurchases have run $20.0 million in 2023, $80.1 million in 2024, $70.2 million in 2025 [54] and $66 million in the first quarter of 2026 — about $236 million, or 40% of today's market value, in nine quarters. Shares outstanding fell from 71.3 million to 65.8 million in the March quarter alone [55].

What the price implies

Sources: shares and cash from the Q1 2026 Form 10-Q [56]; non-GAAP operating income and free cash flow from the FY2025 Form 10-K [57] [58]; 2026 guidance from the Q1 2026 earnings call [59]; price as reported; multiples derived.

The derating is recent and dated. The shares closed at $11.20 on 6 May 2026, the day of the first-quarter release, and at $7.58 six sessions later — a 32% fall — before recovering to $8.88. Two things were said on that call. Chen disclosed that "monetization from one of our large partners started running below our expectations" in auto insurance, with a greater effect expected in the second quarter, and separately that the company would "be more aggressive in placing our long-term bets" [60]. The CFO then guided full-year non-GAAP operating income to $85 million to $110 million against $96.0 million delivered in 2025, holding the top of the range and cutting the bottom [61]. A company that had just doubled its operating profit guided the next year flat, and named single-partner concentration in its largest category as the reason.

The $528 million is not a distress price: there is no debt to default on, and the business generated $113 million of cash last year. At 4.7 times that figure it is a price consistent with 2025 having been the peak of a spread that closes.

The question this report exists to answer

The question is whether the profit NerdWallet earns on traffic it now buys is a durable return, or a narrowing spread left over from the free organic-search audience that built the business — and what an enterprise value near $528 million against $113 million of 2025 free cash flow implies about which.

Three things would settle it in the bull's favor: revenue after performance-marketing cost turning back up on a full-year basis; the insurance line proving to be a broad carrier-demand cycle rather than one partner's budget; and direct, non-search traffic growing enough to be worth disclosing again. Three would settle it the other way: revenue after traffic cost declining for a second consecutive year; operating leverage that comes only from cutting brand and fixed spend, which is a one-time source; and further reductions in what the company chooses to disclose.

Evidence in this chapter runs through the Q1 2026 Form 10-Q filed 6 May 2026 and market prices through 24 July 2026. Second-quarter 2026 results, guided to revenue of $186 million to $200 million and non-GAAP operating income of $6 million to $14 million [62], were not yet in the record.