Insurance · Deep dive

Verisk Analytics

The insurer-owned ratings bureau founded in 1971 that became a for-profit near-monopoly on U.S. property-casualty data — ISO standardized policy forms, industry loss costs, catastrophe models, and the Xactimate claims-estimating platform that sits inside nearly every U.S. property claim — now a ~$3.1B-revenue, ~56%-EBITDA-margin pure-play after shedding energy and financial-services units, but wearing a rich multiple into an AI-disruption debate.

well positioned

Verisk owns irreplaceable, industry-pooled insurance data assets that every major U.S. P&C carrier depends on, sold on ~80%+ recurring subscriptions at ~56% EBITDA margins with enormous switching costs — a compounding, regulator-embedded moat whose only real threats are valuation and a still-unproven AI-commoditization thesis.

My take ★★ Very interesting ★ Interesting − Maybe ✗ Not interesting

Snapshot

Verisk Analytics is the data-and-analytics utility the U.S. property-casualty insurance industry built for itself and then lost control of. Founded in 1971 as the nonprofit Insurance Services Office, it converted to for-profit in 1997 and IPO’d in 2009, and today it sells the standardized policy forms (ISO forms), regulator-filed loss costs, actuarial and statistical data, catastrophe models, anti-fraud tools, and the Xactimate claims-estimating platform that virtually every major carrier and a large share of the claims ecosystem depend on. After shedding its energy (Wood Mackenzie), financial-services, and environmental units in 2022-2023, it is a focused ~$3.1B-revenue (FY2025) insurance pure-play running ~56% adjusted EBITDA margins on largely recurring subscription revenue. It matters as the reference case for a genuine data monopoly — near-100% penetration of the top 100 U.S. P&C insurers, regulator-embedded standards, and switching costs so high the debate is not whether the moat is real but whether AI can erode it.

Founding story

Verisk’s origin is a piece of insurance regulatory history, not a garage. Under the McCarran-Ferguson Act of 1945, Congress gave insurers a limited antitrust exemption to pool loss experience and file joint rates — the only way, at the time, to price risk credibly in a fragmented market. On March 22, 1971, a consortium of P&C insurers and state rating bureaus merged their statistical operations into Insurance Services Office, Inc., a New York nonprofit whose mandate was to standardize policy language and produce industry-wide loss-cost data so member carriers could set adequate rates.

The transformation came in two moves. In 1997 ISO converted from a nonprofit cooperative into a for-profit corporation, and its founding carriers were gradually required to relinquish control — the moment an industry cost-sharing arrangement became a business with pricing power over its own creators. Then, on October 6, 2009, the renamed Verisk Analytics went public in the largest U.S. IPO of the year, pricing 85.25 million shares at $22 and raising about $1.9 billion — but for the selling insurer-owners, not the company: a pure secondary that let them cash out and set a market price.

How it works

The mechanics rest on a data flywheel almost impossible to reconstruct from scratch. Verisk collects granular, policy- and claim-level data contributed by hundreds of insurers — billions of records on premiums, exposures, and losses — and normalizes it into products no single carrier could build alone: ISO standardized policy forms (the contract language most homeowners, commercial, and auto policies are written on), advisory prospective loss costs filed with and approved by state regulators, and actuarial and rating tools carriers plug into their pricing. Because the data is industry-wide and the forms regulator-blessed, adopting them is the path of least resistance; the alternative is building a credible loss database alone and defending non-standard language before 50 state regulators.

On the modeling side, Verisk’s Extreme Event Solutions (the former AIR Worldwide) runs probabilistic catastrophe models that simulate tens of thousands of synthetic hazard years to output the expected losses insurers and reinsurers use to price risk. On the claims side, Xactimate (from the 2006 Xactware deal) is the estimating engine: an adjuster or contractor builds a line-item scope of repair, and the platform prices each task from Verisk’s independently researched, region-specific reconstruction-cost database, refreshed against material and labor surveys. Because carriers and their vendors settle a huge share of U.S. property claims through Xactimate, it effectively sets the reference price for what a repair “should” cost — exactly where the controversy lives.

Product and business overview

Post-divestiture, Verisk reports two core areas. Underwriting Solutions is the historic ISO franchise: standardized forms, loss costs, rating and actuarial data, underwriting attributes, property and address data, telematics, and marketing analytics. Claims Solutions spans Xactimate/Xactware property estimating, anti-fraud analytics, claims-workflow and casualty tools, and a growing set of European claims-tech businesses. Cutting across both is Extreme Event Solutions (catastrophe modeling) plus specialty and climate data. The through-line is proprietary data plus the software that consumes it — Verisk sells both fuel and engine, which is why customers rarely leave.

Business model and pricing

The model is why the stock trades where it does. The large majority of revenue — Verisk describes it as overwhelmingly recurring, ~80%-plus subscription/long-term-agreement — is booked as multi-year subscriptions renewed at high retention, plus transactional revenue that scales with claim and policy volume. Verisk publishes no enterprise price lists; deals are negotiated per carrier by data set, seat, and volume, and Xactimate is sold to contractors and adjusters on tiered subscriptions. The economics are extraordinary: FY2025 revenue of ~$3.07B converted to ~$1.73B of adjusted EBITDA (~56% margin) and ~$1.19B of free cash flow (up ~30%). Margins expanded from ~54.7% in 2024, and the company returns cash aggressively — a dividend raised 11% to $0.50/quarter and a buyback lifted to $2.5B in early 2026. Once data is collected, distributing another subscription is nearly free, so incremental revenue drops to the bottom line.

Traction over time

The arc is a company that got smaller in revenue (shedding energy and financial units) but far higher in quality — margins up, cash flow up, revenue almost entirely recurring. The soft spot is growth: organic revenue decelerated from ~7.5% (2024) to ~6.6% (2025) to ~3.9% in Q1 2026, the crux of the valuation debate.

Market analysis

The addressable market Verisk anchors — insurance data and analytics — is estimated at roughly $17.5-19.6 billion in 2025 across research firms, with low-to-mid-teens CAGRs pushing it toward ~$50B by the early 2030s (Straits Research: ~$17.5B in 2025 to ~$51B by 2033 at ~14.4%). The structural forces are favorable: rising climate/catastrophe volatility (a tailwind for cat models), regulatory demand for rate adequacy (which entrenches ISO’s filed loss costs), escalating fraud, and a push to automate underwriting and claims — all pulling carriers toward more third-party data, not less. The countervailing force is AI: if models plus a carrier’s own data could someday substitute for pooled industry data, the moat narrows. Verisk’s rebuttal is that regulators require representative, industry-wide data to approve rates, and no carrier’s internal data is broad enough — so AI becomes a tool that runs on Verisk’s data, not a replacement for it.

Competitive intel

Verisk’s monopoly is real in its core but contested at the edges. LexisNexis Risk Solutions (RELX) is the strongest broad rival, dominant in contributory driver/claims histories and identity/fraud — the underwriting-data and anti-fraud segments where pooled third-party data is most contestable. Cotality (the former CoreLogic, taken private ~$8B in 2021 and rebranded 2025) competes in property risk and reconstruction-cost data that overlaps Xactimate’s inputs. Moody’s RMS (acquired ~$2B in 2021) is the co-standard in catastrophe modeling and the most credible threat to Verisk’s cat franchise. CCC Intelligent Solutions owns auto claims estimating the way Xactimate owns property — the reason Verisk’s estimating dominance is concentrated in property lines. TransUnion and the credit bureaus nibble at insurance scoring, and Guidewire/Duck Creek are adjacent core-system vendors accumulating policy and claims data that could over time disintermediate some feeds. Where Verisk wins is the combination no rival replicates: regulator-filed loss costs, standardized forms every carrier already uses, and industry-scale pooled data with decades of history. Where it is vulnerable is any single vertical — auto, cat modeling, property data — where a focused competitor has parity.

History and evolution

• 1945 — McCarran-Ferguson Act grants insurers a limited antitrust exemption to pool loss data and file joint rates.
• March 1971 — A consortium of P&C insurers and state rating bureaus forms Insurance Services Office, Inc. (ISO), a nonprofit.
• 1997 — ISO converts from a nonprofit cooperative into a for-profit corporation; founding carriers relinquish control.
• 2002 — Acquires AIR Worldwide, entering catastrophe modeling.
• August 2006 — Acquires Xactware, the property-estimating engine that becomes Xactimate.
• October 6, 2009 — IPOs as Verisk Analytics at $22/share, ~$1.9B raised, the largest U.S. IPO of the year (a pure secondary; no primary proceeds).
• 2010s — Diversifies aggressively via ~40+ acquisitions into energy (Wood Mackenzie, 2015, ~$2.8B), financial services, and environmental data under CEO Scott Stephenson.
• May 2022 — Lee Shavel becomes CEO and pivots to an insurance pure-play; Verisk sells Verisk Financial and 3E.
• February 2023 — Closes the ~$3.1B sale of Wood Mackenzie to Veritas Capital, completing the refocus.
• 2023-2024 — Bolt-on European claims-tech deals (Krug, Mavera, Morning Data, Rocket Enterprise Solutions) expand claims/underwriting digitalization.
• February 2026 — Reports FY2025 revenue ~$3.07B, ~56% adjusted EBITDA margin, raises buyback authorization to $2.5B; guides cautiously as growth slows.

What people say

The case for. Analysts and investors treat Verisk as one of the cleanest data-monopoly compounders in the market: near-100% penetration of the top 100 U.S. P&C insurers, ~80%+ recurring revenue, ~56% EBITDA margins, and switching costs rooted in regulator-filed standards. Raymond James upgraded the stock to Strong Buy with a ~$260 target in early 2026, citing the value of the proprietary data assets and margin expansion; multiple firms (RBC, BMO, Baird, JPMorgan, Wells Fargo) keep positive ratings. The bull framing is that AI needs Verisk’s data to be useful — carriers cannot satisfy state rate-adequacy rules with their own data alone — so AI is a tailwind, not a solvent. Employees rate Verisk ~3.8/5 on Glassdoor across ~1,500 reviews, 78% recommending it, with strong marks for work-life balance (4.2) and culture (4.0).

The complaints. The loudest external criticism comes from contractors and policyholders, and it goes to the heart of Xactimate. Restoration and roofing contractors argue the platform’s reconstruction pricing is systematically too low and too generic — based on median surveys that lag real market costs — and that carriers and adjusters use it to underpay claims, alleging 20-40% underpayment on storm claims. Consumer advocates (United Policyholders) and plaintiff lawyers argue in cases like Sheahan v. State Farm that adjusters manipulate Xactimate at the line-item level to deflate payouts, and a 2025 legal-press piece framed it as software used to underpay wildfire claims — though Xactimate itself has never been successfully sued. Investors’ complaints differ: the valuation is rich (~28x trailing earnings) against growth that fell to ~3.9% organic in Q1 2026, and the AI-commoditization question hangs over the multiple, with Morgan Stanley and Evercore trimming targets to ~$216 and the stock down ~35% over the prior year. There is also latent regulatory risk — an insurance-data concentration this complete invites antitrust, FCRA, and state-regulator scrutiny — and employees cite recurring layoffs “despite record revenue.” The divestiture history cuts both ways: refocusing was value-accretive, but it concedes the 2010s diversification destroyed focus and had to be unwound.

Outlook: well positioned or at risk?

Well-positioned. Verisk is about as close to a structural monopoly as public markets offer in insurance: it owns industry-pooled data no competitor can rebuild, sells it as regulator-embedded standards (ISO forms, filed loss costs) and mission-critical software (Xactimate, cat models) to essentially every major U.S. P&C carrier, and books it as ~80%+ recurring revenue at ~56% EBITDA margins with ~$1.2B of free cash flow. The switching costs are not marketing — they are 50-state regulatory filings, decades of loss history, and workflows wired into carriers’ pricing and claims. Management’s 2022-2023 refocus (shedding Wood Mackenzie, Verisk Financial, and 3E) sharpened the asset rather than diluting it, and the cash returns — an 11% dividend hike and a $2.5B buyback in early 2026 — convert dominance into owner cash.

The honest risks are three, and none yet breaks the thesis. First, growth is slowing — organic revenue decelerated to ~3.9% in Q1 2026 — so a rich ~28x multiple leaves little room for disappointment, and the stock’s ~35% drawdown shows the market repricing that. Second, AI is a genuine long-term question: if models plus carrier-owned data ever substitute for pooled industry data, the moat narrows — but regulators’ insistence on representative, industry-wide data for rate adequacy is a powerful defense, and the likelier path is that AI runs on Verisk’s data. Third, the Xactimate underpayment backlash and insurance-data concentration invite litigation and antitrust exposure that could, in a harsher climate, constrain pricing power. Weigh them against the base rate — a regulator-blessed data utility every customer must use, growing mid-single digits at ~56% margins — and the call is not close: this position compounds. The debate is about the price of the stock, not the durability of the business.

How a challenger would attack it

Start where the monopoly has enemies. Verisk’s most attackable product is Xactimate, because one side of every transaction hates it: contractors allege its median-survey reconstruction pricing runs 20-40% below real storm-repair costs, and policyholder advocates document line-item manipulation in cases like Sheahan v. State Farm. A challenger builds the contractor-side estimating platform — live materials and labor pricing scraped from actual supplier transactions rather than lagging surveys, positioned as the audit layer that checks Xactimate’s numbers. That wedge monetizes the restoration industry’s grievance first, then becomes credible enough that plaintiff lawyers, public adjusters, and eventually regulators cite it — turning Verisk’s reference-price power into a liability. The second attack is the AI-plus-carrier-data play Verisk’s own bulls dismiss: the regulator-filed loss-cost moat protects ISO forms and rates, but it does not protect the analytics layered on top — fraud scoring, underwriting attributes, property hazard data — where LexisNexis and Cotality already have parity and where a model trained on a large carrier’s own book plus public data can substitute. Verisk’s growth already decelerated to ~3.9% organic in Q1 2026 at a ~28x multiple; a challenger doesn’t need to breach the core, only to cap the growth vector the valuation depends on.

Same playbook, new buyer

The Verisk playbook — industry-pooled data converted into regulator-embedded standards, sold back to contributors at 56% margins — is one of the great business models, and it is barely exported. The obvious transplant is other insurance geographies: no European or Asian market has an ISO equivalent with filed loss costs and standardized forms, partly because no McCarran-Ferguson-style pooling regime exists — which means the winning form there is a data consortium structured under local competition law, a build Verisk’s US-regulatory DNA and tuck-in-scale European strategy (Krug, Mavera, ~$10M deals) is not seriously attempting. The second shift is vertical: the pooled-loss-data mechanic fits any market where fragmented players individually lack credible loss history — cyber insurance is the live case, where no carrier’s book is large enough to price systemic risk and a trusted neutral aggregator could become the ISO of cyber within a decade. Verisk won’t move fast on either because its economics reward harvesting the existing monopoly — buybacks at $2.5B, dividend hikes — and because its founding asset, fifty years of US P&C history, contributes nothing to a cold-start pool.

Sources and further reading

• Verisk Reports Fourth-Quarter and Full-Year 2025 Financial Results — Verisk / GlobeNewswire, February 2026. FY2025 revenue ~$3.07B, net income ~$908M, adj. EBITDA ~$1,727M, FCF ~$1,192M, $2.5B buyback.
• ISO Parent Verisk Analytics Raises $1.88 Billion in IPO — Insurance Journal, October 2009. IPO terms: 85.25M shares at $22, largest U.S. IPO of the year, secondary sale by insurer owners.
• Verisk Analytics — Wikipedia, 2025. ISO 1971 founding, 1997 for-profit conversion, AIR Worldwide (2002) and Xactware (2006) acquisitions.
• Verisk Announces Closing of Wood Mackenzie Sale — Verisk, February 2023. ~$3.1B cash plus up to $200M contingent to Veritas Capital; pivot to insurance pure-play.
• Lee M. Shavel leadership profile — Verisk, 2025. CEO since May 2022; prior Nasdaq CFO and BofA Merrill Lynch banker.
• Xactimate is not the law: How insurers use one software program to underpay wildfire claims — Daily Journal, 2025. Line-item manipulation and underpayment allegations.
• Xactimate Demystified — United Policyholders, 2024. Policyholder-advocate view of Xactimate pricing and Sheahan v. State Farm.
• Will Verisk’s Valuable Data Moat Hold, or Will AI Commoditize It? — The Motley Fool, May 2026. The AI-disruption debate and analyst positioning (Raymond James Strong Buy ~$260).
• Verisk Analytics (VRSK) — Insurance Analytics Monopoly — Beeli Capital / Substack, 2025. Moat, ISO network effects, and switching costs; McCarran-Ferguson context.
• Insurance Analytics Market Size, Share & Growth Report — Straits Research, 2025. ~$17.5B 2025 market, ~14.4% CAGR to ~$51B by 2033.
• Verisk Analytics (VRSK) Statistics & Valuation — StockAnalysis, 2026. Market cap, P/E, EV/EBITDA, share count.
• Verisk Reviews — Glassdoor, 2024-2025. ~3.8/5 across ~1,500 reviews; work-life-balance praise, layoff complaints.