Taxi driver retention — how UK fleet operators cut driver turnover in 2026

The visible cost is the £400-£800 acquisition spend repeating itself — application processing, the 4-8 week licensing-flight retainer or guaranteed-hours period most operators carry for an incoming driver, and the admin time to onboard a replacement. The less visible cost is the vehicle sitting idle or under-utilised during the gap between a driver leaving and a replacement reaching full productivity, plus the corporate-account and regular-customer relationships a departing driver takes with them if the fleet runs any account or regular-rider work tied to a specific driver rather than the platform.

A fleet that has never calculated its own turnover rate is not unusual — most track headcount, not churn, because headcount is what a rota needs day to day. The number worth calculating is drivers who left in a rolling 12 months divided by average active driver count over that period; anything above roughly 25-30% for an owner-driver or commission fleet is worth investigating, and anything above that for a PAYE-employed model is a materially higher signal given the added recruitment overhead of employment status.

Commission rate gets blamed for turnover more than the exit data usually supports. Drivers comparing operators do look at commission or rental cost, but operator-reported exit conversations point more consistently at three other factors: unpredictable earnings week to week (not low earnings, unpredictable ones), a dispatch process that feels arbitrary about who gets which jobs, and settlement problems — late payment, disputed line items, or unclear deductions that a driver cannot audit themselves.

A driver who earns less but can predict it, and can see exactly how a settlement was calculated, routinely outlasts a driver on a marginally better rate who cannot tell whether the dispatch algorithm is treating them fairly or whether this week's payout will match their expectation. This is why fleets that improve settlement transparency and job-allocation clarity report retention gains that a straight commission-rate cut rarely produces on its own.

A driver who quits usually gives a polite, low-friction reason on the way out — "moving on", "family reasons" — because there is no incentive to give the operator an honest account of what actually pushed them, especially if they plan to reapply or maintain goodwill in a small local driver community. A structured exit conversation, run by someone other than the driver's direct dispatcher and asked as specific questions (earnings predictability, job allocation fairness, settlement clarity, vehicle condition, support responsiveness) rather than an open "why are you leaving?", surfaces patterns a polite one-line resignation never will.

The value is in the pattern across departures, not any single exit. One driver citing settlement confusion is an anecdote; six drivers citing it over two quarters is a fixable process gap that is materially cheaper to address than the recruitment cost of replacing all six.

Drivers talk to each other, and a driver who believes the good jobs consistently go to someone else — whether or not that belief is accurate — will act on that belief by looking elsewhere, regardless of how the allocation actually works. A dispatch system that can show a driver (not just tell them) how a job was offered — proximity, acceptance-rate history, rotation fairness — closes the trust gap that a verbal assurance from a dispatcher cannot.

This matters more in fleets running any AI-assisted or algorithmic dispatch, where a driver has even less visibility into why a job went to someone else than in a manual radio-dispatch model. An operator moving to modern dispatch software should treat allocation transparency as a retention feature, not just an efficiency one — see our AI vs human dispatcher comparison for how allocation logic differs across dispatch models.

Settlement is the part of the operator relationship a driver checks every single week, which makes it one of the highest-frequency trust touchpoints in the entire relationship — far more frequent than a rota conversation or a licensing renewal. A settlement a driver can audit line by line, paid on a predictable schedule with no surprise deductions, does more for retention per pound spent than most driver-facing perks an operator considers.

Operators moving off legacy settlement processes (manual reconciliation, disputed deductions resolved case by case) toward an audited, itemised settlement run typically report the dispute volume itself drops sharply within the first few settlement cycles, independent of any change to the underlying commission structure — the trust problem was the opacity, not the rate.

Driver app quality shows up repeatedly in operator-reported retention data and in what drivers raise at recruitment fairs when comparing operators (see our driver app comparison guide) — sub-second push notification delivery, a clean offer-accept flow, and a settlement view the driver trusts. A driver evaluating whether to stay with an operator is, in effect, also evaluating whether to keep using that operator's app every working day, and a slow, unreliable, or confusing app is a daily friction point that compounds.

This is not an argument that app polish alone retains drivers — pay structure and job allocation fairness matter more directly — but a poor app is a retention tax on top of whatever else is working, and it is one of the cheaper problems to fix relative to redesigning commission structure or rota policy.

Track active driver count and departures monthly, segmented by tenure band (under 3 months, 3-12 months, over 12 months) rather than as one blended figure — a fleet losing drivers mostly in the first 3 months has an onboarding problem; one losing drivers mostly after a year has a structural pay or trust problem, and the fix for each is different. Blending both into one turnover percentage hides which problem is actually driving the number.

Pair the quantitative number with the exit-interview themes from section 3 — the number tells an operator whether turnover is getting better or worse; the themes tell them what to actually change. A fleet that only tracks the percentage without the reasons behind it ends up reacting to symptoms (raising commission, running a one-off bonus) rather than the underlying process gap driving the departures.