Experian Takes a Mortgage Haircut — and Quietly Proves the ChatGPT Bet
A 7% organic compounder gives back half a point of guidance to rising rates, while its LLM distribution partnerships move from pilot to measured reality.
EXPN.L · Earnings Call · 2026-11-18
The Arithmetic of a Narrowed Guide
Experian's fiscal first half was, by its own standards, boringly excellent: organic revenue growth of 7% in both quarters, benchmark EBIT margin up 50 basis points to 28.8%, benchmark EPS up 15% at actual rates to $0.975, operating cash flow up 16% to just over $1 billion on 81% conversion, and return on capital employed of 17.3% on a growing capital base. Then management took the top of its organic revenue range off the table — narrowing full-year guidance to 6%-7% from a 7% midpoint.
The cause was not demand. It was arithmetic, and Lloyd Pitchford was unusually plain about it:
With rising interest rates, we now expect mortgage volumes to decline by about 20% in the second half.
Mortgage is roughly 5% of the group. Moving from a flat volume assumption in H1 to down 20% in H2 costs about a point in the second half, half a point for the year. Everything else was held constant. Yet the guidance cut collides with the tape's dominant grammar: rising interest rate environments and the impact of high rates both sit high in the market-wide keyword list this quarter, alongside a cluster of Fed-hike language. Experian is no longer a special case; it is a transmission channel for the same rate shock every other reporter is describing.
The irony is sharp. In H1, North American mortgage revenue grew around 40% on flat volumes — almost entirely price — and the prior call had guided to only a "slight downtick" in volume. A downtick became a one-fifth cut. On the margin "at least 50 basis points," Pitchford was asked whether there was room above: “at least means at least” — Lloyd Pitchford, Chief Financial Officer · 2026-11-18.
The Disruptor Turns Out to Be a Channel
The genuine news of the half is distribution. For two years the bear case on Experian's consumer business was that a chatbot would eat the credit-education funnel. In November 2025, Brian Cassin framed LLM traffic as explosive in growth but immaterial in scale: “a really significant increase in traffic from the AI platforms in excess of 1,000% growth over the last year. But the traffic coming from there is still very small” — Brian Cassin, Chief Executive Officer · 2025-11-12. By May 2026 the language had moved to pilots — “We've integrated the first score display on ChatGPT in U.K.” — Brian Cassin, Chief Executive Officer · 2026-05-20 — and a Google Gemini app-integration programme announced "in the last 24 hours."
Now it is live and measured. OpenAI and Gemini integrations are running, and: “more than 350,000 consumers have already used the experience, and encouragingly, around one in five then go on to engage with our own Experian properties.” — Brian Cassin, Chief Executive Officer · 2026-11-18 Direct audiences still grew 8 million in the half to 224 million; subscriptions kept growing; management sees no cannibalisation. Attribution is genuinely hard and the base is small, but the direction matters more than the number: consumer relationships are being pushed into third-party environments rather than defended inside an app — a strategic answer to the personal AI agents wave rather than a rear-guard action against it. Cassin resists overselling: “We are not assuming all of these opportunities will succeed. We do not need them to.” — Brian Cassin, Chief Executive Officer · 2026-11-18
The FICO Clock Is Still Ticking
The recurring tension of the last three calls remains unresolved. FICO royalty pass-through added roughly 1.5% to organic revenue and cost 40 basis points of reported margin — under the hood, underlying organic margin progression was +110bps against a 20bps acquisition drag. It is a zero-margin toll. And FICO has been signalling alternative forward distribution models. Experian again assumes none of it happens: “we continue to assume the traditional FICO distribution method and consistent pricing benefit.” — Lloyd Pitchford, Chief Financial Officer · 2026-11-18 Compare that with May: “we're not assuming in our guide any structural changes to the market in FY '27” — Lloyd Pitchford, Chief Financial Officer · 2026-05-20. Same sentence, six months apart. Suhasini Varanasi pushed for an FY2028 view; Pitchford deferred to "by the time we get to May, we might get a little bit more clarity."
This has stopped being an Experian-only question. The phrase FICO Preferred Program appears in the market-wide top keywords this quarter — the score-distribution debate is now a sector-level theme. Meanwhile the hedge is quietly compounding: VantageScore is now being used to price actual mortgage originations, a step past last call's "testing." The distribution model question, tracked on the company's own keyword history under FICO royalty, is the single largest swing factor in the medium-term revenue and margin optics of this business.
What Went Quiet
Two themes have faded, and the fade is the story. Data breach support — once the company's single largest momentum spike, and still a top-five keyword as recently as the last quarter — is now simply a roll-off drag. North American consumer services fell 4% reported, and 3% ex-items. Cassin conceded the obvious: “it's unlikely you're going to see something of that size to replace that revenue.” — Brian Cassin, Chief Executive Officer · 2026-11-18
The other silence is the dual-run cloud cost tailwind that used to anchor the margin bridge — 20 basis points a year for four or five years. It barely rated a mention. Productivity has replaced it. Labor cost has fallen from 35% to 30% of revenue since H1 FY2024, roughly $400 million annualised, with about 30% productivity gains across the product development population from AI tooling. That is the new underpinning of margin guidance, and the reason management can absorb royalty dilution and still promise at least 50 basis points.
The half also carried $2.5 billion of buybacks announced since January plus $0.6 billion of dividends — around 9% of market capitalisation — with net debt to EBITDA at 1.9x, just under target. Ascend platform B2B revenue grew 17% with all ten largest US financial services clients live on it, and EMEA/Asia Pacific margin jumped to 8.5% from 4.9%. LatAm consumer services grew 34%.
This is not a story of deterioration. It is a compounder whose growth rate is being temporarily set by the US mortgage market, and whose next leg — proprietary data monetised through AI platforms — is real, evidenced, and still too small to size. Excellent execution, one macro headwind disclosed honestly, one structural unknown punted to May.