Constructive Distractions
Backseat Driver
A car-maintenance advisor that works for the owner, not the shop.
US auto repair is a $199B market built on 289M vehicles with an average age of 12.8 years — the oldest ever recorded. 78% of drivers distrust the shops they depend on. The app that actually fixes this doesn't exist yet.
Build in progressThe moment that matters
Picture it: you're in the oil-change waiting room. The service advisor comes out with a clipboard. Your cabin air filter is “really dirty.” Your brake fluid is “discolored.” They recommend a flush — $340 all-in. You have no idea if any of it is true, urgent, or fairly priced. So you either pay and feel played, or decline and spend the next week wondering if your brakes are quietly failing.
83% of drivers consult someone else after a mechanic's recommendation — a parent, a friend, a frantic Google search in the parking lot. The behavior already exists. The product doesn't.
You don't know your schedule
Most owners have no idea what the manufacturer actually recommends at a given mileage — what's real vs. what's marketing.
You can't see what's coming
There's a timeline of upcoming work — some expensive — with no tool that surfaces it proactively.
You're captive at the shop
Upsells arrive when the car is on the lift and leaving costs time. No low-pressure way to evaluate what you're being told.
You can't evaluate a recommendation
Is it real? How urgent? What's fair to pay? Does declining risk the warranty? No app answers this in real time.
“My 2013 Hyundai Sonata engine seized one mile from the dealership. I'd just gotten married, joined a pre-revenue startup, and my wife was the breadwinner. The dealership claimed the seizure was due to faulty maintenance and asked for records I didn't have — most of the work had been done by a family friend with paper receipts only. Credit card statements and a signed letter weren't enough. The dealership had no incentive to push the claim through.
Seven months later, a service advisor twenty miles away told us something the first one didn't: you can photograph the engine internals to show the oil passages are clean — proof the seizure wasn't from neglect, regardless of paper records. We were lucky. Without him, I had no way to prove I'd taken care of my car.”
(The engine was notorious for seizing — Hyundai had already faced a lawsuit — which is why the extended warranty existed. The records problem was still nearly enough to sink a claim I was legally entitled to.)
What's already out there
Fragmented. Each app solves a slice. None solve the moment that matters.
| Know your schedule | See what's coming | Real-time at the shop | Evaluate the rec. | |
|---|---|---|---|---|
| Log & reminder apps | ||||
| OBD dongle apps | ||||
| CARFAX Car Care | ||||
| RepairPal | ||||
| On-demand mechanics | ||||
| Car super-apps | ||||
| Backseat Driver |
“Real-time at the shop” is completely vacant.
No product answers, in the waiting room, in real time: “Is this recommendation real, fair, and urgent?” Problem C has no incumbent.
Why the gap persists
Every app that gets close to the trust problem gets economically captured by the supply side it's supposed to police.
FIXD
“Know what your warning light really means”
→ Devolved into dark patterns and subscription traps — optimized for revenue, not user trust
CARFAX
“Passive service history you can trust”
→ Consumer app steers toward dealers — the supply side it surfaces is also its distribution partner
RepairPal
“Price transparency for any repair”
→ Shops pay to be listed as "RepairPal Certified" — the recommender is funded by the recommended
The product that solves problem C can't be funded by shops without poisoning itself.
Conclusion: the consumer must be the customer. That's not just an ethical stance — it's the only model where the incentives stay clean enough for the product to work.
The product
Three moments. Three distinct jobs.
At home
Enter your VIN → your actual maintenance schedule in plain English. What each item is, why it matters, what it costs. Reminders adapt to mileage.
At the shop
They recommend something → open the app → legitimacy, urgency, fair price range, and exactly what to say to the advisor. The second opinion that already exists as a behavior.
After
Snap the receipt → record vault updated. Warranty-grade documentation whether the work was done at a dealer or by a family friend with paper receipts.
Now — MVP
- VIN → maintenance schedule
- Smart reminders (mileage + time)
- Second Opinion flow
- Record vault
Next
- Community pricing (Glassdoor for mechanics)
- Warranty-aware guidance
Later
- ML on uploaded photos
- Shop-side verified records
The business model
The incentive structure is the product. Every revenue decision flows from that.
Consumer subscription — ~$5–10/mo
Schedule and reminders are free. Decision support and the record vault are paid. No shop relationships. No referral revenue. The subscription is the only line of business — which means there's nothing to compromise the recommendation engine.
Data flywheel — community pricing
Upload receipts → see what others paid for the same job, same region, same vehicle. The give-to-get mechanic solves cold-start for price data and builds a compounding moat — consumer-sourced, not shop-sourced.
Shop/dealer promotions
Deliberately parkedThere's an obvious revenue line in connecting shops to users who are due for service. It's parked. It recreates the exact conflict of interest the product exists to fix. If ever revisited: opt-in only, clearly labeled, structurally firewalled from any recommendation output. Reasoning about why not to take revenue is the judgment the product depends on.
Tradeoffs & risks
The honest version of a case study includes what could go wrong.
Accuracy is existential
Wrong advice about a brake job is worse than no advice. Mitigation: conservative urgency framing, OEM schedule data as ground truth, clear "advisory, not diagnosis" language throughout.
Cold start on pricing data
Fair-price ranges need data before the community exists. Mitigation: seed from public estimator sources; community data refines over time.
Engagement is inherently episodic
People think about car maintenance a few times a year. Retention design must respect that — reminders + vault as the heartbeat — rather than manufacturing fake daily engagement.
Willingness to pay is unproven
Consumers say they distrust mechanics; whether they'll pay $7/mo to fix it is the core market risk. This is a hobby project — the honest framing is: this is the riskiest assumption.
Liability surface
Advice that intersects with safety and warranties needs careful language standards and probably legal review before anything is customer-facing.