Generally yes — the miles you drive between one job and the next, while you are online and available, are business miles, and so is the drive out to collect the passenger or the order. The miles that are not are the ones driven with the apps off: the trip home at the end of the night, the errand mid-shift, the school run. Simple to state, and genuinely hard to evidence afterwards, which is the real subject here. Your log has to know which was which at the moment it happened.
What are the three phases of a rideshare driver’s time?
Drivers and preparers talk about three, and the whole question turns on them. Phase one is online and waiting: the app is on, you are available, parked or circling. Phase two is en route: a job is accepted and you are driving to the pickup or the restaurant. Phase three is engaged: the passenger or the order is aboard and you are driving to the destination.
All three happen while you are working, and in the ordinary case all three produce business miles — the drive out to a pickup is as much part of the job as the drive with someone in the car. Outside the phases sits everything else: the commute before you go online, the drive home after you go offline, and personal detours in between. Delivery work has the same shape with different nouns, and the drive from one platform’s drop-off to another’s pickup is squarely between jobs.
- Phase one — online, available, no job accepted
- Phase two — job accepted, driving to the pickup or the restaurant
- Phase three — passenger or order aboard, driving to the destination
- Outside the phases — apps off: commute, personal errands, the drive home
Why is waiting with the app on treated differently from driving home?
Because one is time spent in the business of being available for hire and the other is the private cost of getting yourself home. A driver at a rank with the app on is doing the thing the business consists of, and repositioning in that state is generally business mileage. A driver who has gone offline and is heading for their own driveway is commuting, which tax systems treat as a personal expense almost everywhere, however much the day it ends was a working day.
The line is drawn by the state you were in, not the road you were on. The same stretch of highway is a business mile at ten in the evening with the app on and a personal mile at eleven with it off. That is why "were the apps on?" is the most useful thing a rideshare mileage record can carry. Two caveats worth a preparer’s answer once: purely speculative repositioning with nothing accepted is more argued than the rest, and going online in your own driveway does not automatically convert the drive that follows.
Which miles are definitely not deductible?
Anything driven with the apps off and no business purpose, plus personal detours taken during a shift. This is the part drivers most often get wrong in their own favour, and the easiest for an examiner to test, because personal driving leaves traces — a car that supposedly never drove to a supermarket is not a believable car.
A shift is not a container. Being on shift does not make every mile inside it business; each trip stands on its own. Pull off the route for your own groceries and the leg to the shop is personal, the leg back toward work is business again, and the cleanest record lets the stop split the driving into separate trips that each take their own classification. Logging the personal trips is defensive as well as honest: a log containing personal driving was obviously kept rather than assembled, and the business entries inherit that credibility.
Does the platform’s own mileage figure cover this?
No, and this is the most expensive assumption in rideshare tax. A platform can only see the miles attached to jobs on its own system, so any figure it shows you is by construction a subset of your business driving. It cannot see the drive from a competitor’s drop-off to its own pickup, it may not see what you drove while waiting, and it certainly cannot see the trip to the parts store or the car wash.
That does not make the figure useless — it makes it a cross-check. If your own log shows dramatically fewer miles than a platform’s engaged-mileage figure, something in your capture is broken and you want to know in June rather than April. Use it as a floor and a sanity test, never as the number that goes on the return.
Why does the log have to distinguish the phases at the time?
Because the phase is a fact about your state and your app, and neither leaves a trace you can recover later. Six weeks on you can see that you drove eleven kilometres across town on a Thursday evening; you cannot see whether you were repositioning while online, driving to a pickup, or going home. The distance survives and the classification does not — and the classification is what the deduction depends on.
That is why a reconstructed rideshare log is weak evidence even when the mileage in it is honest: every phase has been assigned from memory, so the claim rests on recall rather than record. A log written at the time does not have the problem, because the answer was obvious when it was made. The standard is low — capture every drive automatically so nothing is missing, then classify the day while the day is still in your head.
What should a rideshare mileage record actually contain?
The four things every tax authority asks for, plus the one thing rideshare specifically needs. Date, distance, endpoints and business purpose are the standard set, covered in full by the audit-survival guide on this blog. What rideshare adds is the state: which platform the driving served and which phase it was in, written into the purpose so it never has to be remembered separately.
A purpose of "repositioning, apps online" or "en route to pickup, food delivery" is a handful of words and the whole difference between a trip a preparer can assess and a bare distance nobody can defend. Keep the wording consistent across the year — consistency is itself evidence.
BrewGig is built for that split: drives detected and recorded in the background on Pro and above, with endpoints, route and distance captured as they happen, leaving you a swipe queue for the one judgment a phone cannot make. BrewGig is independent and not affiliated with or endorsed by any platform named here; Uber, Lyft and the other names are the trademarks of their respective owners, and how each counts mileage is theirs to define — read your own statements and your tax authority’s current guidance.