The driving-school question is a values question dressed up as a money question. People ask, "should I go to a driving school?" but they mean, "do I trust myself, my family, and my own discipline enough to skip one?" Those are different questions, and they have different answers.
I'm going to argue that for roughly half of Kenyan learners, school is the right choice, and for the other half, it isn't. The trick is knowing which half you're in before you spend the money, not after.
The honest case for school
A good driving school sells you three things, only one of which you can replicate at home.
A car with dual controls. A second brake pedal on the passenger side is the only reason instructors live to teach a second cohort. You cannot install one on your uncle's Vitz. This is non-negotiable for the first few hours of practical lessons; it is the difference between a stalled engine in a quiet yard and a stalled engine on a roundabout.
A neutral instructor. "Neutral" is the operative word. A driving school instructor has taught the same yard reverse 8,000 times. Your father has taught it twice, both to you, both ending in shouting. The transactional, repeated, slightly bored teaching of a school is, frankly, what most learners need.
Practice volume. Schools front-load lessons. You'll do 10 to 15 hours in three weeks, much of it on the same maneuvers (reverse-into-bay, three-point turn, hill start). This is the entire point. Most self-taught learners average four hours of actual driving across the same three-week period and wonder why they failed.
The thing school doesn't sell you, despite the brochure, is the theory test. The theory portion is rote material. You will spend more time on it in a school than you would alone with the right materials. We say this knowing it makes the math more confusing, not less.
The honest case for self-study
Self-study works when three things are true at the same time.
One: you have access to a car you can use weekly. Not "sometimes". Weekly. The single biggest predictor of practical pass rate is hours behind the wheel, and four hours every Saturday for six weeks is more than most school packages.
Two: someone with patience can sit beside you. Patience is doing more work in that sentence than "skill". A 20-year licence holder who cannot keep their voice down for 90 minutes is a worse practical instructor than a 5-year licence holder who can. Look at the temperament; the skill is downstream.
Three: you are honest with yourself. Most self-study failures are not skill failures. They are scheduling failures. The car was available, the relative was patient, the learner moved the session three Saturdays in a row, and then walked into the test underprepared. If you'd skip three Saturdays, pay the school.
The cliff edge
The worst outcome in this market is paying for school but treating it like a passive subscription. Three lessons in, the car is fine, the instructor is fine, but the learner is not driving every week and the muscle memory never forms. School makes you pay; school does not make you practise.
A decision rule that actually works
Score yourself on three lines, honestly, before opening WhatsApp.
- Car access: weekly (3 points), every other week (2), monthly (1), none (0).
- Patient co-driver: yes and they live with me (3), yes and I see them weekly (2), maybe (1), no (0).
- Discipline: I'll book and keep a weekly Saturday lesson for 6 weeks (3), I'll do most of them (2), I'll miss some (1), I'll start strong and fade (0).
Add the three. 7 to 9: self-study is honestly fine. 4 to 6: borderline; consider a small school package (5–7 lessons) just for the dual-control hours. 0 to 3: pay for a proper school. You'll save money in the long run.
We didn't make this rule up. It mirrors what we hear in the feedback we collect on the site, and it lines up roughly with the actual pass rates we see across the Pelican trainer's user base.
What's the same either way
Whichever route you pick, two things stay constant. The theory test does not care how you studied. And the road signs do not care either.
If you're going to learn signs, you might as well learn them efficiently. The Pelican trainer handles recall under mild time pressure, and 3D sign recognition handles the bit of the test where the sign is rendered at an angle. Both are free, both work without signup, and both will get you 80% of the way to the theory pass mark in about a week.
When you're ready to sit a simulation, the quick NTSA test runs in three minutes and tells you what you don't know yet.
The piece nobody includes
There's a third option people don't usually consider: hybrid. Take a five-lesson package from a school just for the dual-control hours and the yard maneuvers, then drill the rest with a relative on quiet roads. It's the lowest cost-to-confidence ratio I've seen, and it works particularly well for learners who scored 4–6 on the rule above.
The schools won't sell you this directly. They'll sell you a 15-lesson package. Ask if they'll do 5 lessons at the per-hour rate, and many of them will.
Pick, then go
If you've picked: the step-by-step NTSA walkthrough is the next read.
If you're still weighing it on money: the full cost breakdown has the numbers.
If you're picking the class first: Class B, C or D.
The right answer is not the cheapest answer. The right answer is the answer you'll actually follow through on. Start there.