The 24-Point Checklist for Choosing Taxi Dispatch Software

Dispatch software is a five-year decision that most operators make in about three weeks, based on two demos given by people who give demos for a living. That is not a criticism of them — it is just an uneven contest unless you arrive with your own list. Print this, take it into the call, and ask to be shown rather than told.

Before you book a single demo

Twenty minutes of preparation changes what you get out of every call afterwards. Write down, on one page:

1. Your real numbers — active drivers, vehicles by type, bookings per month, percentage taken by phone versus app, and how many are account or contract work.
2. The three things that go wrong most often now. Not features you want; problems you have.
3. What must survive the move — your passenger list, your account customers, your historic job data, your phone numbers.
4. Who has to use it. A system that the two controllers love and the drivers hate will fail.

Dispatch and operations (1–6)

1. Phone booking speed. Time it. Under 20 seconds for a repeat passenger with saved addresses is achievable; over a minute will be resented within a week.
2. Caller recognition. Does an incoming number pull up the passenger, their usual pickup and their history before the controller speaks?
3. Auto-dispatch behaviour. Ask them to simulate a job nobody accepts. Watch the offer roll driver to driver, then cascade to the next zone. Ask how long each offer window is and whether you can change it.
4. Manual override. Reassign a live job, force it to a named driver, and put it back in the queue. If any of those needs support to do, that is your answer.
5. Zone building. Draw a zone on the map during the call. Ask how adjacency works and whether zones can have their own tariffs.
6. The unallocated job. How does a job that nothing could take surface to the controller, and how loudly?

Driver experience (7–11)

Drivers change firms over app quality more often than operators expect. These five are worth more than any feature on the office side.

1. Poor signal behaviour. Ask what happens when a driver goes through a tunnel or into a rural notspot mid-job. Does the app queue state changes and sync, or lose them?
2. Battery drain. Continuous GPS is expensive. Ask what a 10-hour shift costs in battery on a mid-range Android.
3. Offer visibility. Can a driver see the job's pickup, destination and fare before accepting, or only after?
4. Earnings transparency. Can a driver see their own day, week and settlement without ringing the office? Every call this prevents is a call your controller does not take.
5. Onboarding a new driver. How long from "he starts Monday" to a working app with the right documents on file?

Money (12–17)

1. Fare priced before dispatch. The passenger and the driver should see the same number before the journey starts.
2. Tariff complexity. Airport surcharges, waiting time, soiling charges, school-contract rates, bank-holiday multipliers. Ask them to build one of your awkward ones live.
3. Card payments. Whose Stripe account, whose rate, and how quickly does the money reach you?
4. Pay-by-link. For phone bookings from passengers with no app — can you text a payment link, and what does the text cost?
5. Driver settlements. Show me a week's settlement for one driver, including commission, account jobs and card takings. Ask how much of it is manual.
6. Account invoicing. Can a corporate client get one monthly invoice with a cost-centre breakdown, generated rather than assembled?

Compliance and risk (18–20)

This section is short and non-negotiable. Licensing failures cost operators their licence, not their margin.

1. Document expiry alerts. DBS, badge, MOT, insurance, PHV plate. Alerts must fire *before* the date, to someone who acts, and must be able to block a driver from going on shift.
2. Audit trail. Can you reconstruct who changed a booking, when, and what it looked like before? Ask what happens in a complaint or an insurance claim.
3. Data protection. Under UK GDPR you are the controller of your passengers' data and the vendor is your processor. Ask for their data processing agreement before you sign, not after — and read the sub-processor list.

The commercials (21–24)

Ask these four in writing. They all have short factual answers, and having them in an email rather than a memory of a call is worth a great deal if anything is ever disputed.

1. Total annual cost at my real numbers — including setup, app build, SMS and the support tier you would actually need.
2. Minimum term and notice period. These vary widely across the market, from monthly rolling to multi-year with a long notice period. Ask rather than assume.
3. Price at renewal. Ask whether the quoted rate is introductory, and what governs increases. A rate fixed or capped in writing is worth more than a lower opening number.
4. Cost and format of a full data export. How easily you could leave is worth knowing before you arrive.

One last test

Ask for the phone number of an operator roughly your size who switched to them in the last year, and ring that operator. Ask what the first month was like. Vendors who are proud of their onboarding hand this over immediately.

If you are working out budget, the guide to what dispatch software actually costs covers the pricing models and the fees that only appear on the invoice. If you are already on a system and weighing a move, switching without losing a day's work covers the migration itself.