Microlise: The £5M Restructure That Turned an OEM Leak Into a Direct-Customer Story

H1 FY26 shows direct ARR up 12% and EBITDA margin snapping back to 13.2% — but an OEM renewal cliff and two more years of negative free cash flow are the real test

SAAS.L · Earnings Call · 2026-09-23

A pivot you can now measure in pounds

Microlise Group reported H1 FY2026 on 23 September, and the headline is that the direct side of the business is finally doing the work management promised a year ago. Direct customer ARR reached £47.1M, up 12%, while direct customer net revenue retention held at 106% and churn stayed unusually low at 1.1% — circa £600k. EBITDA margin recovered to 13.2% for the half, against 5.2% in H2 2025 and 9.9% for the full prior year. That is a material swing, not a rounding error. The mechanism is a restructure executed at the back end of 2025, and CEO Nadeem Raza is explicit about the arithmetic: “We did have GBP 5 million of analyzed cost savings. We have done further savings in H1 this year, that has enabled us to then reinvest some of those things in more targeted aspects of the business that will drive that direct growth.” — Nadeem Raza, CEO · 2026-09-23 The reinvestment is aimed squarely at the mid market — a GBP 65M ARR opportunity in the U.K. alone, where Microlise sits at roughly 30% share versus 60% in the large-enterprise segment — and at TMS, which management repeatedly calls its highest-value, highest-margin module. The proof point is concrete: more TMS deals closed in H1 than in the prior two years combined, direct evidence that the lengthening sales cycle and implementation cycle that have bedevilled deployments are being attacked rather than excused.

The OEM leak: small in the numbers, large in the lesson

The direct story only makes sense against what is happening on the OEM side. CFO Nick Wightman framed it as a GBP 3.1M reduction versus the twelve months to June 2025, driven by a falling renewal rate on second-life contracts — the airtime-and-service extensions that end users buy through OEM dealers after the initial five-year term. Management now expects a further decline in FY2026, a lower and slowing decline in 2027, and a plateau toward the end of 2027. What makes this interesting is the geographic dispersion. Raza's answer to a question on OEM renewals is worth reading in full:

Some regions we can see 80% renewal rate. In some regions we can see 10% renewal rate... Our view is that the reasons why those users, who are often second or third life users, because they are not the people that bought the equipment in the first place... it is not something that they consider of significant value in those regions at that age of product.
Nadeem Raza, CEO · 2026-09-23

He then adds the admission that matters: “which is why it caught us by surprise in 2025” — Nadeem Raza, CEO · 2026-09-23. A company conceding it was blindsided — and then applying, in their words, a "hard ax" to forward assumptions — is a different disclosure than the usual sandbagging. Management stresses it has good line of sight on first-term contract expiries but the least visibility on what end users do at renewal, precisely because Microlise sits at arm's length from them, transacting only with the OEM.

Data packs, AI tooling, and the elusive moat

The language here is refreshingly specific. Microlise launched a Driver Excellence data packs product earlier this year, then found something useful when customers deployed it: “it did lead us to understand that some of our customers have very low data quality” — Nadeem Raza, CEO · 2026-09-23. That spawned a second pack that scores customer data quality and prepares the estate for AI tooling arriving toward year-end. Monetisation is dual-track — standard packs bundled into upgrade packages, premium versions charged separately. Raza is candid that AI in the products is not new (“AI isn't necessarily something new to us within our products” — Nadeem Raza, CEO · 2026-09-23), which reframes the AI tooling narrative from novelty to packaging.

The cash bill for the pivot

Here is where the story gets harder. Despite the margin snap-back and a £13.5M cash balance, no debt and an undrawn £30M HSBC facility (a £10M committed RCF plus a £20M accordion), management guided to negative free cash flow in both 2026 and 2027, turning positive only as the investment programmes mature late in 2027. The culprit is timing: a £6M working-capital inflow in H2 2025 pulled forward collections, and £2M of restructuring severance went out the door in January–February 2026. The offsetting ambition is margin enhancement married to operational leverage — procurement savings, hardware design-out, cloud-estate rationalisation, and internal AI deployment — with an initial 20% adjusted EBITDA target that is explicitly framed as a waypoint, not a destination. The stated logic is mix, not magic: direct ARR carries materially higher margins than OEM revenue, so as OEM flatlines post-2027, group margin should float upward mechanically.

Why the market context matters here

The most revealing thing about Microlise this quarter is how little it rhymes with anything else. Scan the global tape and the last twelve months of market narrative have been dominated by Los Angeles utility politics, tariff refund windfalls (echoed by CBRL, MLKN and KMD.NZ in the recent reporter set), AI data-centre capex and Middle East conflict. Microlise mentions none of it. Its keywords — mid market, data packs, renewal rate, OEM side, churn — are entirely company-unique. The only ambient risk it flags is churn-adjacent supply chain stress: component lead times so long it is placing orders for late 2027 and 2028, with engineering redesign work underway to plug 2027 availability gaps. That is the trade here. A £50M-market-cap U.K. logistics software name with a genuinely idiosyncratic pivot, a real margin turn, an honest confession about an OEM blind spot, and a two-year cash trough that management is asking investors to look through. The direct-customer engine is compounding; the question is whether the OEM reset flatlines before patience — and the balance sheet — runs out.