Sy Jacobs — Haymaker March Webinar: the cheap auto-insurance lead-gen marketplaces

David Hay's intro — the SII multi-year-breakout chart

• Hay opens on Sprott (SII), "stirring" in late 2023 then breaking out from a tight trading range in 2024 — "our dream scenario," especially with a strong fundamental story (assets grew drastically, hugely accretive to margins as overhead barely rose).
• Sy recommended SII "way back when"; Hay credits him with A-Mark, Banco Popular, Allstate "and on and on" — "one of the best investors… super investor… not being hyperbolic." (SII's ticker can be read "Si" — a happy coincidence with Sy's name.)
• The webinar's main theme: the bull case for "lead generation companies" — an esoteric, small, and "extremely cheap" part of the market.

Sy's background — a financials-only stock picker in an index world

• 30 years managing money in a private partnership; a dozen years before that as an equity research analyst, starting at Salomon Brothers in the mid-'80s, randomly assigned the financials group.
• Went through the 1985 Salomon training program with Michael Lewis "by chance" — which is how he ends up named in The Big Short, "in chapter one, in fact" (left on the movie's cutting-room floor).
• Survived "the gale-force wind against stock picking" in an ETF/index market and can "still smile about it."

The mortgage-securitization epiphany — a technological shift in lending

• In the mid-'80s, watching Lou Ranieri "birth" the mortgage-backed-securities market at Salomon, he saw MBS as a game-changer — "almost like a technology," a massive shift away from S&Ls/portfolio lenders toward securitization, that then rippled through every lending market.
• His first taste of the core method: champion the companies that benefit and short the ones in harm's way — "I don't really need to know which direction the stock market's going to go… If I just get this right about this change in fundamentals." Forty years later, "that's still what I'm doing."

The method — contrarian, theme-and-cycle, long/short

• The only thing he knows how to do: follow financials, form opinions on which are undervalued/underappreciated vs overvalued/overappreciated, pay attention to themes and cycles, and "be suspicious when everybody loves something… when everybody hates something."
• Track record of playing both sides: right on the big short in mortgage, while sitting on the board of a high-quality subprime-auto-finance company he later took private, sold to PE, bought back during COVID, and sold to Fiat Chrysler/Stellantis.

Precious metals inside a financials-only fund — SII and AMRK

• Post-COVID — with all the stimulus, money printing and QE — he leaned harder into precious metals, but his fund is financials-only, so he expresses it through Sprott (SII) and A-Mark (now gold.com) — both financial-services companies.
• "Longs that came out of a sense of doom about the US dollar and interest rates going up and social and political polarity in the United States" — i.e. longs born of macro skepticism, expressed through stock-picking.

The carriers' post-COVID blunder — and the cyclical recovery

• The auto underwriters (Progressive, Allstate, State Farm) were hugely profitable during lockdown (less driving, fewer accidents), then "stepped on the accelerator," grew like mad into a price war, and got slammed by post-COVID revenge-driving and inflation — a sharp down-cycle of profitability in '21–'22.
• What carriers do best is fix their own mistakes: cut spend, push through state-approved price increases (regulators grouse but grant them), restore margins, then spend again — which played out as big price increases in '23–'24. He bought ALL and PGR "a beat or two too early" at the trough.

Discovering lead gen — the operating-leveraged way to play the up-cycle

• Lead-gen companies are the operating-leveraged way to play the coming auto-insurance up-cycle with greater return potential than the carriers: cyclical, but with a big secular growth tailwind.
• They run online pricing-and-binding marketplaces — harvesting leads from Google and other searches, then selling them to carriers/brokers or, increasingly, letting you click through a comparison site to a real quote and bind straight through to (e.g.) the Progressive mainframe.
• It looks commoditized but isn't: no two driving records/cars are alike, there are 50 states' regulatory pricing regimes, and carriers guard their pricing algorithm closely — divulging it only to a handful of trusted exchanges, not to websites at large or to LLMs.

The 20-year demographic super-cycle — the non-cyclical tailwind

• "Every time one of us takes the keys away from their father or grandfather… it is one less customer that gets their insurance through a live human being. And every time one of our kids turns 16… this is like a baby boom generation fading away and turning into generations X, Y and Z" — who are increasingly comfortable buying auto insurance online, never speaking to an agent.
• That demographic tailwind "is not at all cyclical" — a super-cycle that's run 5–10 years and "is going to last another 20."

The cyclical trough turning — "the really good times are upon us now"

• The cyclical part: every 5–10 years carriers make mistakes or get into price wars, tap the brakes and buy fewer leads. A few years ago he judged the cycle at its trough just as the tech-change tailwind was accelerating, and these stocks were depressed — he played it via both carriers and lead-gen.
• The cycle has now turned: prices up, margins restored. Progressive accelerated a year ago; the rest got their pricing in the last 3–9 months and are now interested in growing again — he sees a 2–3 year growth cycle ahead, blowing at the same time as the demographic/technological cycle. "The really good times are upon us now at the beginning of 2026."

What blew it off course — first tariffs, now the AI/LLM scare

• His doubled-or-more carrier longs (ALL, PGR) have actually outperformed the lead-gen names over the last 6–12 months, even though he expected lead-gen to do better — the two diverged despite the operating leverage showing up in the businesses.
• First it was a "tariffs-victim" scare last spring (fear that pricier foreign auto parts would raise repair costs); that didn't play out — carriers stayed profitable and restored marketing budgets, to the lead-gen names' benefit. Now it's the AI/LLM scare: "the companies are doing great and the stocks are doing poorly."

Why these are orphan stocks — mis-covered, sub-$1B

• Sub-billion-dollar caps, thinly covered (three to five analysts each, typically the IPO underwriter plus a couple of random tech/business-services analysts).
• Taken public as "internet services plays" at big multiples; the tech analysts were "dismayed to learn there's an insurance cycle" — and a bad one in '21–'22 — so they became orphans: not covered by insurance or financial-services analysts. "I'm coming at this much differently than most."

David Hay's price-to-sales-vs-PE "double-cheap" interjection

• Hay's readers are "indoctrinated" to look at price-to-sales in cyclical industries because earnings are volatile — a very high PE with a very low price-to-sales is often the perfect buying opportunity; here it's the rare "double win," low price-to-sales AND low PE.
• He's pivoting out of US stocks generally, but if you're in US stocks, into "the value stocks that are still value stocks" — former low-PE names like Walmart, Eli Lilly, Caterpillar and Deere are now "priced for perfection." The big opportunity: AI-fear has crushed a long list of fine companies, so correctly identify the ones that are AI-neutral or, better, AI-winners.

Sy on AI/LLMs — a tailwind, not a threat (the regulatory moat)

• Why LLMs aren't a threat: carriers have great fear of their proprietary pricing/underwriting models becoming a public utility — "their secret sauce," hundreds of thousands of employees over decades. They have no interest in handing it to an LLM, and state regulators wouldn't want that either. "It's a state-regulated business, which I think is actually better protection."
• So LLMs (like Google search) become more useful to consumers but end up delivering leads to these companies — "an ally," "just another new media source of leads." The names addressed the LLM threat on recent calls and "very eloquently made the case that LLMs are helping them right now" (efficiency internally; ChatGPT price-comparisons that dump into their lap). "They got lumped in with the software and SaaS scare, but it's so different."

The names — EVER the purest, MAX the best capital manager, QNST the favorite

• EverQuote (EVER): purest auto lead-gen and "so inexplicably cheap because it's so simple" — ~$585M cap, $171M cash, no debt → ~$410M EV on ~$112M EBITDA (he thinks low); EBITDA ≈ FCF ≈ GAAP, so 3.7× cash flow. Hoarding cash; he's pushing them in Cambridge on Friday to buy back much more.
• Media Alpha (MAX): a little less pure (some Medicare/Medicaid health), ~$740M adjusted EV on ~$130M cash flow. White Mountains owns 30% + board seat; MAX more than doubled its buyback authorization to $86M last week — "the real deal when it comes to capital management."
• QuinStreet (QNST): his favorite and largest — pristine balance sheet, buyback history, plus the home-services lead-gen business he loves (much earlier in its online-adoption cycle), turbocharged by the Jan-1 HomeBuddy.com acquisition (15–20%+ grower, highly accretive). ~$666M cash/debt-adjusted cap; he sees ~$150–160M EBITDA → "four-and-change times EV/EBITDA." "If I could only own one… for the next five years."

David Hay on EBITDA — here it really is close to free cash flow

• Buffett is critical of EBITDA, but these are asset-light models with little interest/depreciation, so the usual EBITDA-vs-FCF gap is small — EverQuote's last quarter showed free cash flow, adjusted EBITDA and GAAP "very similar."

Sell discipline & the long/short edge

• His sell rule for the lead-gen names: hold them through the cycle until the market stops treating them as fully cyclical and they re-rate to "double-digit multiples of EBITDA, like most companies" — then he has to sell, because "I can no longer steal them."
• He runs long/short, ~50% short — being "ambidextrous on the long-short side allows you to be a more patient investor," because you can live with longs and shorts knowing there's an opposite book.

Carvana short — "this cycle's Enron/WorldCom?"

• He'd put Carvana on a "permanent do-not-short list" after getting his fingers burned, then took it off ~6 months ago ("I don't have a boss, I can capriciously break my own rules"), shorted it, and covered when the big short report came out — "I guess too soon."
• "Deep, deep suspicions" from having been inside a subprime-auto-finance company — "I just don't get how they do it" (Gotham and, before, Hindenburg agree); thought it could be Enron/WorldCom "for five to 10 years."

David Hay's wrap — private credit, Lincoln, and the multi-year-breakout screen

• Hay: private credit (not private equity) is the bigger threat to life insurers — float/carry-trade plays exposed to private-credit loans, including to AI-victim software firms. Lincoln (LNC), his cheap, weak personal holding, is hit by private-credit fears he thinks are "overblown" (he also holds the more-diversified AIG).
• He's building a systematized Bloomberg screen for companies making multi-year new highs — ~280 names so far; many 2023–24 breakouts went up enormously, lots of fresh utility and financial breakout candidates. He'll narrow the financials and have Sy pick his favorites. A future webinar features Robert Mullen (Marathon Resources).