CarGurus (CARG): debt-free, 26% operating margin, but the 9.9% FCF yield is ~6.6% once you normalize tax and SBC

**Thesis.** CarGurus is the most-visited car listings site in the US, with no financial debt, a 26.5% operating margin and a buyback that cut the diluted share count by 29% since 2022. After shutting down CarOffer in 2025, what's left is a clean subscription business. My problem is the growth: US dealer volume has nearly stalled, so growth now comes from price increases, and the published free cash flow overstates what repeats. I'd buy below $28. At $29.35 it's close, about 5% above, but not there.

**Dates.** Today (6 Oct 2026) the stock is at $29.35.

Against the Monte Carlo median intrinsic value of $40.21: 40.21 / 29.35 − 1 = +37.0%

Against the median of the five valuation models ($34.42): 34.42 / 29.35 − 1 = +17.3%.

Figures are from the FY2025 10-K and the 2026 10-Qs, $ million unless stated.

**How it makes money**

Dealer subscriptions first, then OEM advertising, then finance partnerships. At 30 June 2026: 34,629 paying dealers (US 26,151) and average quarterly revenue per subscribing dealer of $6,771 (US $8,134). Year on year that's +4.6% volume and +6.6% price, which explains most of the 13.1% revenue growth in Q2 2026. Trailing four-quarter revenue from continuing operations: 967.3.

The moat is a two-sided network: 46.8 million monthly unique users (37.8 million in the US, +11.1%), which pulls in dealers, whose inventory pulls in consumers. The "Deal Rating" algorithm adds consumer trust a competitor can't buy with ad spend.

**Balance sheet and cash**

* No financial debt. A $400 mn revolver sits undrawn. The $185.1 mn some screens call "debt" is operating leases.
* Cash fell from 231.2 to 122.1 in a year because buybacks ($369 mn over four quarters) exceeded operating cash flow.
* Buybacks of $908.7 mn in 3.5 years, 31% of the market cap. Diluted shares: 128.2 mn (2022) to 91.1 mn (Q2 2026).
* Receivables grow at half the speed of revenue, DSO steady at about 16 days. Accruals are negative in every year: CFO over the last four quarters was 318.7 against net profit of 175.9.
* Book equity is down to 264.4 from 541.7 at end-2024, so P/B (12.6) and ROE (53.5%) describe the buybacks, not the business.

**The number I'd adjust: free cash flow**

Published trailing FCF is 291.3, a 9.9% yield at $33.17. Two things inflate it:

* Cash tax paid over four quarters was $2.1 mn against $53.8 mn of book tax, as capitalized R&D tax treatment reverses and prepaid tax comes back. The deferred tax asset behind this fell from 106.6 (2024) to 76.1 and is running down.
* Stock comp of $51.1 mn a year (5.9% of revenue with the capitalized part) isn't deducted.

291.3 − 51.1 − \~45 tax normalization ≈ 195 of owner earnings, a 6.6% yield at $33.17. GAAP net profit from continuing operations over the same period was 187.1, so two separate routes land in the same place.

**The bear case**

*CarOffer.* Management bought it in January 2021, the month Jason Trevisan became CEO. It lost $219.0 mn in 2023-2025, with $144.4 mn of 2024 impairments on top (CarOffer plus a closed online-buying pilot). Value destroyed: over $300 mn, more than 10% of today's market cap. In Q3 2024 Digital Wholesale lost $25.3 mn on $27.3 mn of revenue. To its credit, the board shut it in August 2025, and discontinued losses in H1 2026 were zero. I still charge the record in the discount rate.

*Dealer count and pricing.* US paying dealers grew 2.6% y/y in Q2 2026, down from 5.5% in 2025. US revenue per dealer rose about 8% a year for three years: $6,532 (Q4 2023) to $8,134 (Q2 2026). Contracts are month-to-month with no renewal obligation, and dealers also pay Autotrader, [Cars.com](http://Cars.com) and CARFAX. If price increases stop working, cancellations show up within a quarter.

*Competition.* Autotrader belongs to Cox, bundled with Manheim auctions and Dealertrack. CarGurus won't build that stack after CarOffer. Carvana competes for the transaction itself rather than the dealer's ad budget. The longer risk is AI search moving the first step of a car search away from listings sites.

*Margins.* H1 2026 vs H1 2025: sales and marketing +17.4% on revenue +13.9%, gross margin 92.2% vs 93.1%, operating margin ex-impairments 24.8% vs 25.7%. Amortization of capitalized software doubled to $4.2 mn in Q2.

*Used-car cycle.* Dealers are under margin pressure. The allowance for doubtful accounts rose from $600k to $850k in six months. Small, but it moves first. My third pre-mortem: tax normalizes (published FCF falls to about $230 mn with nothing operational changing) at the same time a slowdown cuts dealer ad budgets 10%, as in 2020 and partly 2022.

*Governance.* The CEO has doubled as interim CFO since March 2025; a permanent CFO starts 19 Oct 2026. The founder holds 64.6% of votes through Class B shares and has been converting B to A, the step that precedes selling. The February 2026 breach (about 12.4 million records, class actions since withdrawn without prejudice) isn't named in either 2026 10-Q.

**Valuation**

Owner earnings $195 mn, 9% growth for years 1-5, 10.5% discount rate, 2.5% terminal, net cash $122 mn, 90 mn shares. Five models: DCF bear $25.91, EPV $21.99, own 5-year median EV/EBITDA (10.4x) $34.42, DCF base $39.94, DCF bull $62.10. A 20,000-run Monte Carlo on the same DCF gives a median of $40.21, with P10-P90 at −23.4% to +94.9% at the $33.17 reference. The model's spread comes mostly from the growth assumption feeding the terminal value. Treat lease liabilities as debt and the base value drops about $3.41.

On the scorecard (Graham, Buffett and Lynch blocks, the last two weighted double) it gets 67.8%: Graham 5.5/9, Buffett 7.0/9, Lynch 5.5/9. That lands in my "interesting" band, which is a watchlist rating, not a buy.

**What would change the verdict**

* Upward: US paying dealers growing above 5% y/y at the 5 Nov 2026 report, or operating margin holding above 26% once AI spending eases.
* Downward: gross margin below 91% for two straight quarters, sales and marketing growing above 18% y/y without faster revenue, revenue growth below 8% y/y, or operating margin below 20% for two quarters.
* On 5 Nov I'm also watching cash tax paid (a jump to $30-40 mn over nine months means normalization has started) and the bad-debt allowance (above $1.1 mn would be real dealer stress).

Under $28, the DCF base value gives +43% and the Monte Carlo median +44%. If the November numbers hold up and the price gets there, it moves from watchlist to buy. Corrections on the tax normalization are welcome; that's the part that moves the number most.

Full report, with the FCF bridge, quarter-by-quarter balance sheet and pre-mortem: [https://waitingisthework.com/stocks/carg/?ref=reddit](https://waitingisthework.com/stocks/carg/?ref=reddit)

*Disclosure: I run Waiting Is the Work, the non-commercial site where this analysis is published. I hold no position in CARG. This is research, not investment advice.*