Open in interactive viewer → charts, metric popovers & call review

Serko Raised Guidance and Quietly Wrote Off Its Own New Growth Bet

An H1 FY27 that delivered the first real free cash flow — and a U.S. corporate-segment story that went from guidance swing factor to rounding error in two quarters
SKO.NZ · Earnings Call · 2026-11-15

The upgrade nobody asked about, and the admission nobody wanted to give

On the surface, Serko's H1 FY27 is the cleanest half the company has produced. Total income grew 17% to NZ$72.3 million, total spend grew 12% to $66.4 million, EBITDAFI rose 37% to $8.4 million, and the net loss narrowed from $9.5 million to $0.9 million. Free cash flow came in at $8.9 million against $3 million a year earlier — a 196% swing — and the balance sheet carries roughly $62–63 million of cash and no debt. Gross margin touched 90%. Management lifted FY27 income guidance to $135–141 million from $128–134 million. Underneath that, though, sits a genuine reversal. For two quarters, the corporate segment experiment in the U.S. was pitched as the swing factor in guidance. In May, CFO Shane Sampson said of the U.S. targeted segments: “they're contributing a meaningful portion of both the low and upper end of the range” — Shane Sampson, Chief Financial Officer · 2026-05-20. The conceit was that a month of timing slippage in signing those customers could move real revenue into FY28. This week that evaporated into a single Q&A answer: “we indicated that we're not expecting meaningful revenue from the U.S. corporate segments” — Shane Sampson, CFO · 2026-11-15. Five customers are signed. Volume has not followed.

Despite these sign-ups, adoption has been limited to date, and turning sign-ups into meaningful volume is taking longer than we would have expected.

Matthew Gerrie, COO · 2026-11-15
COO Matt Gerrie was frank about why: supply. These mid-market buyers need specific attributes — parking, availability depth — that Booking.com's inventory doesn't always satisfy. The tension is that the keyword stayed loud in the narrative even as the substance drained out of it; corporate segment still ranks near the top of the company's theme set precisely because it is now a problem being explained rather than an opportunity being sold. That is what makes this report interesting. Serko beat, raised, and simultaneously de-risked a promise. The upgrade happened despite zeroing out the thing that was supposed to be the upside.

What actually carried the half: a compounding loyalty loop

Booking.com for Business did the lifting, and the second derivative finally turned. Active customers reached 363,000, up 38%. Completed room nights hit a record 3 million, up 47%. Revenue rose 33% to $47.9 million. Crucially, the pace accelerated within the half: completed room nights rose 13% sequentially in Q2 versus 2% in the comparable period a year earlier, and active customers grew 12% quarter-on-quarter against 9% a year ago. What management is really selling is operating leverage. Product and technology spend on Booking.com for Business has fallen from 39% of that unit's revenue in H1 FY24 to 19% today — each incremental dollar of volume needs less reinvestment. Over three years the loop compounded at 106% in customers, 128% in room nights and 116% in revenue. Call it a growth loop, and the mechanics are credible. The nuance investors should hold onto is the tiered commission structure. Revenue per completed room night fell 12% as volume pushed into lower-marginal tiers — by design. Sampson characterised the mix as sitting in the second and third tiers for most of the half, with September the first month below the 50% tier. Analysts questioned whether the second-half guide looked conservative; the answer was that “the guidance still implies I think growth in the second half in terms of completed room nights above 50%” — Shane Sampson, CFO · 2026-11-15, with the usual southern-hemisphere seasonal hole in managed travel offsetting part of it. Elsewhere the legacy book behaved better than feared. Australasia travel revenue rose 5% to $14 million with revenue per booking up 12% to $6.23, flattered by the Australian dollar. GetThere's U.S. bookings fell 15% to 1.5 million — a shallower decline than guided, helped by one customer recovering volume and by departing customers lingering longer than expected. Modest churn has, however, begun to appear in that base.

Serko.ai: agent-first, monetisation-later

Serko spent $14.5 million building Serko.ai in the half, offset by $7.3 million of efficiency gains elsewhere, and pushed it into open beta in the U.S. in mid-October, inviting 800 waitlist users. The positioning is deliberate: most incumbents bolt chat onto existing tools, whereas Serko built an agent-first product where, in Gerrie's words, “many agents work in concert and carry out hundreds of actions” — Matthew Gerrie, COO · 2026-11-15 behind a single conversation. This is where Serko meets the broader market rather than leading it. The global theme tape has been consumed all year by personal AI agents and agentic AI — from the proliferation of consumer shopping agents to enterprise agentic stacks. Serko is a small-cap riding a very large wave, and its wedge is genuine: policy guardrails built into every search, and live spend visibility rather than month-end estimates. The addressable framing is the U.S. SMB travel market at roughly US$76 billion a year for companies under 500 employees, or about US$84 billion including micro-businesses. The unresolved piece is the same one that was unresolved two quarters ago. In May, the commercial model was still being tested. This week, CEO Darrin Grafton confirmed: “I don't think the AI industry has settled on what are the ideal commercial models for their products, and we have not settled on one either” — Darrin Grafton, CEO · 2026-11-15. Today they charge a slice of total transaction value as a placeholder. The explicit ambition is to move away from the flat managed-travel fee that has capped industry economics for a decade. Until that pricing is proven, investment in Serko.ai remains a call option funded by the core, not yet a business.

The tell is in the cash

Perhaps the most revealing shift is tonal. Six months ago, asked about a $100 million free cash flow target, Sampson pushed back on the premise: “that's our key focus versus arbitrary focus on generating positive free cash flow” — Shane Sampson, Chief Financial Officer · 2026-05-20. This week the company led with it — $26.8 million of core cash generation, $3.9 million of GetThere outflow, $14 million into Serko.ai, and still $8.9 million positive at the group level. The old cash burn worry has flipped to a discipline narrative. FY27 free cash flow is guided positive but below the first-half figure, reflecting second-half seasonality and further AI investment. That is a company that has stopped arguing about whether it deserves to be cash generative and started arguing about how fast it can reinvest. The bear case writes itself: a NZ$93 million market-cap software company whose flagship new initiative has already missed its first commercial test, whose second-half guide leans on Booking.com's willingness to keep porting business bookers across, and whose AI monetisation is a hypothesis. The bull case is equally clean: 47% room-night growth accelerating into the second half, 90% gross margins, no debt, and an aspirational $250 million revenue line by FY30 that now needs less help from the U.S. corporate detour than it did in May. Watch two things — whether the 800 waitlist converts into repeat booking behaviour by March, and whether the tiering drag on revenue per room night stabilises before volume growth has to carry the whole load.