Every shop owner trains their first installer the same way: standing next to them, correcting technique in real time, catching mistakes before they reach the customer. It works — right up until you're not the one standing there. Somewhere between installer number two and installer number four, the informal apprenticeship model that got you this far quietly stops scaling, and the shop's quality starts depending on which technician happened to work on a given car that day.
This guide is about the fix: turning what's in your head into written, structured systems a new hire can actually follow. What to document first, how to build an install checklist that gets used instead of ignored, when video beats text, a real 30/60/90-day ramp plan for new installers, and how to catch quality drift with spot-checks and comeback tracking before a bad job turns into a bad review.
1. Why "I trained them myself" stops working
When you're the only installer, or one of two, quality control is trivial: you did the work, or you personally watched someone you trained do it, on cars you inspected yourself. There's no gap between "how it's supposed to be done" and "how it's actually done," because there's no distance between the standard and the person enforcing it.
That breaks the moment you add a third person, or the moment your first two hires start training your third and fourth. Now there are two, three, four slightly different versions of "how we do it" floating around, each shaped by who trained whom and what shortcuts crept in along the way. None of it is written down, so none of it can be checked against a standard — it just is the standard, until a customer's car proves otherwise.
The fix isn't more supervision. You can't personally inspect every car once you're running four bays, and trying to is how owners burn out. The fix is writing the standard down once, in enough detail that someone who's never watched you work can follow it, then building a light system to check that the standard is actually being followed. That's the rest of this guide.
2. What to document first — variance and risk, not everything
The instinct when you decide to "finally write SOPs" is to try to document everything: how to answer the phone, how to sweep the bay, how to cut film. Don't. A shop that tries to write forty SOPs on day one finishes zero SOPs, because the project is too big to start.
Instead, sort every recurring task in your shop along two axes: how much it varies between your best two people, and how much it costs you if it's done wrong. Document the ones that are high on both, first.
- High variance, high risk — document these immediately. The steps where two competent installers would genuinely do it differently, and where the difference matters to the customer or the warranty: the contamination check before laying film, curved rear-glass shrink technique, ceramic coating prep and dwell times, PPF stretch and relief-cut placement, chemical mixing ratios, what actually counts as "clean enough" before final inspection.
- High risk, low variance — document these next. Steps everyone already agrees on, but where a slip is expensive: pre-existing damage documentation on intake, VIN and vehicle verification against the work order, anything touching electrical systems or trim removal that can crack if forced.
- High variance, low risk — worth documenting eventually, not urgently. Scheduling preferences, tool cart organization, which playlist runs in the bay.
- Low variance, low risk — don't bother yet.
A simple test: pull your last 90 days of comebacks, warranty claims, and any "the customer wasn't happy" conversations. Whatever step those problems trace back to — that's your first SOP.
3. Structuring an install checklist that actually gets used
A checklist built around failure modes — not a sequence of steps — is what catches problems. "Inspect the install" becomes "no visible bubbles or debris under film at any edge, checked in direct light from three angles." Split into two lists: pre-install (VIN/vehicle match, pre-existing damage photographed, product tier confirmed, contamination check, materials staged) and post-install (the failure-mode items specific to the service, plus a customer walkthrough and signature). Keep each under 12-15 items.
If you're running SalesThumb, Settings → Quality Control lets you build separate pre- and post-install inspection templates with your own checklist items and a default template per kind, and the Gates tab can require a signed post-install inspection plus before/after photos before an invoice's pay link goes out — so the checklist you design actually gets enforced instead of living in a binder nobody opens.
4. Written vs. video SOPs — and where they live
Written SOPs for judgment calls, ratios, sequences, and decision trees. Video for hand-skills that are hard to describe and obvious in 30 seconds — squeegee angle, film stretch on a relief cut, a properly shrunk curved window. Pair a short written SOP with one or two 60-90-second clips rather than over-producing either format. Where they live matters less than whether the habit of checking them gets built during onboarding.
5. The 30/60/90-day new-installer ramp plan
Days 1-30: study the checklists, shadow, prep work only, no unsupervised customer cars. Days 31-60: supervised jobs, working the checklist item by item rather than from memory. Days 61-90: solo work, every job spot-checked against the same standard, ending in a graded day-90 benchmark job that's the graduation gate — pass and they move to full independence, fail and they get two more weeks rather than being pushed forward on a deadline.
6-8. Catching drift, spot-checks, and comeback tracking
Quality drifts quietly even with a good checklist and ramp plan. Catch it with two systems: a scheduled (not reactive) spot-check cadence — weekly for Level 1-2 installers, every other week for seasoned Level 3s, jobs picked at random and scored against the identical checklist customers get — and a comeback log that captures root cause (not just symptom) and gets reviewed monthly for patterns by installer, service, and vehicle type. If you're running SalesThumb, the AI Install QC tool (AI → Install QC) gives a fast, advisory second set of eyes on finished-work photos for spot-checks, and the warranty claims queue (each claim ties back to the original warranty, vehicle, and job) doubles as a ready-made comeback log if reviewed on the same monthly cadence.
9-11. Keeping SOPs alive, common mistakes, and the year-one picture
Assign an owner per SOP, set update triggers (new product, a comeback root cause, a near-miss) instead of a calendar nobody follows, and keep a one-line changelog. Avoid treating checklists as a compliance exercise, making spot-checks punitive, skipping the day-90 benchmark under schedule pressure, or letting a top performer's undocumented habits become the real standard. A year in, the shop looks the same from outside but runs on one checklist instead of three tribal versions, a scheduled spot-check cadence, and a comeback log reviewed monthly — with the owner touching fewer cars personally without quality depending on who's in the bay that day.
[Full guide includes complete prose for all 13 numbered sections, a 12-15-item checklist framework, the full 30/60/90 day-by-day plan, spot-check and comeback-tracking systems with SalesThumb feature references, a mistakes section, a Frequently Asked Questions section matching the faq array, and a Related Reading section — see apps/web/src/lib/guides-data.ts for the verbatim ~3,540-word body as published.]