Verisk Analytics, Inc. VRSK

Key data

The business

Verisk sells data and analytics almost entirely to property and casualty insurers. The core product is the underwriting and claims information a carrier needs to price a policy and settle a claim: loss-cost data, catastrophe models that estimate hurricane and wildfire damage, and software that routes claims. Most of it arrives as hosted subscriptions, billed a year in advance and set to auto-renew, which is why the revenue is steady and the customer rarely leaves. After selling its energy unit (Wood Mackenzie) and its financial-services business in 2022 and 2023, Verisk is now a pure play on insurance data. The moat is the dataset: decades of pooled industry loss data that no single insurer holds and a new entrant cannot recreate. An underwriter at a mid-size carrier opens Verisk's ISO forms every morning, and switching would mean rebuilding the actuarial backbone of the business.

The numbers

The direction that matters is the top line, and it is slowing.

Revenue grew 7.8% in Q2 2025 against the prior year, 5.9% in Q3, and about 3.9% in Q1 2026. That deceleration is the whole story, and it is clean: the divestitures that muddied 2022 and 2023 are behind the numbers now, so the roughly 4% is the real organic pace. Q2 2026, reported July 29, came in at adjusted earnings of $1.98, five cents ahead of estimates.

Over four years revenue compounded about 5.7% while diluted earnings per share compounded about 12%, from $4.08 to $6.48. The gap is the tell. Net income actually fell in 2025, to $908M from $958M, so the per-share growth comes from a shrinking share count, not a growing business. In the first quarter of 2026 Verisk spent $1.4B buying back stock, more than double any prior quarter, and funded it by taking long-term debt from $3.1B to $4.8B. Net debt now sits at about 2.4 times EBITDA. That is the machine: mid-single-digit organic revenue, 44% margins, and a buyback doing the heavy lifting on earnings per share. The variant is plain: the market prices Verisk as a durable compounder, and the compounding is increasingly financial, not volume. The organic subscription growth rate in the next two prints settles it.

For a data business the real question is whether pricing power survives slower volume. So far it has.

Management

Lee Shavel has run Verisk since March 2022 and completed the pivot to pure-play insurance data. The capital record is aggressive: buybacks have run ahead of free cash flow, funded partly with debt, which works while the stock is cheap against its own history and stops working if leverage climbs and growth does not return. Insiders have been net sellers, about $6.4M sold against $0.8M bought over the past year across 20 sales and 6 buys. Director Samuel Liss sold about $2.1M in early June, and Shavel himself sold $696K on August 3, with plan status not disclosed on either. Stock compensation is low, under 2% of revenue, and the latest option grants struck at $220.59 sit underwater against today's $185.79.

How it fails or surprises you

Growth stalls below 4%. If carriers push back on annual price increases, consolidate, or build analytics in-house, organic subscription growth slips under 4% and the buyback can no longer paper over it. The first tell is the organic subscription growth figure on the next earnings call. A print near 3% breaks the compounder label.

The balance sheet loses its cushion. Verisk took debt to $4.8B, net 2.4x EBITDA, to fund a $1.4B quarter of repurchases. If growth disappoints, there is less room to keep shrinking the count. Watch net debt to EBITDA and the quarterly buyback pace together. A pause in repurchases would say the model is tapping out.

AI turns the dataset into a new product (right tail). Verisk's pooled loss data is exactly what insurers need to train and run underwriting and claims AI. If that converts flat subscriptions into higher-value usage revenue, growth reaccelerates past the mid-single digits the market expects. The tell is transactional revenue outgrowing subscriptions for two straight quarters.

Closing thoughts

The market has already compressed the multiple: at 17.7 times EBITDA the stock sits near the cheap end of its twelve-year range, down 43% from its high, so the price no longer pays for the acceleration the company once delivered. What it does pay for is durability, mid-single-digit growth at 44% margins for a long time. If organic subscription growth holds near 5%, the price is fair and the buyback quietly compounds earnings per share. If growth drifts toward 3%, the debt-funded repurchase looks like a stretch and the multiple has further to fall. The left tail is a slow leak, not a collapse. This business does not break, it fades.

The bet is still that property and casualty insurers keep paying Verisk every year for data they cannot rebuild, and keep paying more. What breaks it is those price increases stalling. The one pair to watch is organic subscription growth against the quarterly buyback: as long as the first stays mid-single-digit, the second is a tailwind, and the day growth sags while the buyback keeps running on borrowed money, you are paying 28 times earnings for financial engineering.