Failure modes and hard truths

Every course on this site teaches failure as a real lesson, and this guide earns that promise here, because gig work has four failure modes that end bridges, and all four are quiet until they are sudden. You have met pieces of each. Now they get a lesson of their own.

The first is deactivation, the platform firing you by software. The triggers are numeric, and most are published: Uber says drivers can lose access for ratings below the minimum average rating in their city, computed over the last 500 rider ratings, and publishes no national number; operators put the line near 4.6 in many markets, and drivers report losing access right at it, with new accounts most exposed because a handful of bad ratings moves a small sample. DoorDash gates your account on completion rate; its documentation now describes dropping below 90 percent as deactivation risk where 80 percent was once the line, and drivers report warnings even above it. Every platform also reserves permanent deactivation for fraud, account sharing, and severe safety reports, and the appeals run through forms and queues, not phone calls. The operator-reality summary: treat your rating, completion, and acceptance history as your credit score, check it weekly, screenshot your trip records monthly, and never build a household budget that one algorithmic decision cannot survive. That last clause is the entire argument of this guide.

The second is pay erosion, and it is not a risk, it is the default weather. You have seen the data: customers paying nearly 10 percent more per trip while platform fees per trip rose more than 33 percent and driver pay per hour rose 4.1 percent in 2025. Rates you calibrated in March quietly buy less in October, and the forums fill with posts about the same phenomenon from the inside, drivers who were pulling $22 hours last year finding $15 ones this year. The defenses are real but limited. Work peaks, not averages. Multi-app so no single algorithm sets your prices. Track your net hourly weekly, the way "A real week on the bridge" showed, so erosion gets caught in weeks, not years. And remember what the erosion actually is: a pricing system where you hold no pricing power. The permanent fix is the far bank.

The third is the car death spiral, and it deserves the name. AAA put all-in ownership of a new vehicle at $11,577 a year in its 2025 study, about 77 cents a mile. Bridge mileage consumes the car in proportion to your hours, and the failure pattern is always the same: the transmission goes in month seven, the payout did not include a depreciation reserve, the repair goes on a credit card, the interest raises your cost per hour, so you drive more hours to cover it, which consumes the car faster. The countermeasures are mechanical. Price your hours on the full-ownership ledger, not the marginal one. Build a repair fund from day one, a fixed cut of every payout, before the set-aside account if you must choose. Buy used and unglamorous for bridge duty. And the moment repair costs exceed the car's value twice in a year, that plank is over; switch to a low-mileage plank rather than feed the spiral.

Two hard truths to close, both structural. The platforms can change pay structure, fee structure, and deactivation policy whenever they like; you agreed to that in a clickwrap contract, and drivers in several cities have learned it overnight. And the bridge pays you precisely because it does not grow: every hour is sold once, at a rate someone else set, with no equity, no compounding, and no customer list. That is not a moral complaint. It is the specification of the machine. Machines built to those specifications are for crossing, and a far bank worth crossing to is one whose pricing power survives software; the AI-exposure method is how this catalog checks.

You now know what the machine is, what it pays, what it costs, and how it breaks. What remains is the reason this guide exists: the far bank.