Trust the number on the screen. That is the advice for anyone who accepts work through a platform: the offer is the deal, the screen shows what you will earn. The offer and the settlement are controlled by the same party. That is not a flaw in the system. That is the system.

A Spark driver in a parking lot looks at his phone. One delivery, one dollar amount, one tap to accept. He drives, delivers, and checks his payout. The number is smaller. In early 2026, the FTC settled with Walmart for $100 million over that gap. Christopher Mufarrige, the FTC's director of consumer protection, put it plainly: labor markets cannot function efficiently without truthful and non-misleading information about earnings and other material terms. The Spark settlement is the largest federal case documenting what happens when people follow the advice to trust the number.

"Trust the number" does not fail people. It fails the system they are trying to run.

The problem is not that Walmart was careless. The problem is that the party making the offer also controls the formula that sets the payout.

Internal Walmart documents, cited in the federal complaint, show this gap was known and left in place for years. One manager called the incentive failures a "train wreck." Another flagged the base pay gaps as a "serious issue" and a legal risk in November 2022. By January 2023, Walmart was fielding roughly 2,000 driver complaints per week about base pay alone. Drivers said the company was stealing from them. By May 2023, a Walmart employee put the failed incentive payouts at $600,000 per month.

The anchor number is worse. An internal audit found $199,000 in tips charged to customers but never paid to drivers, in a single week. Walmart told customers that 100 percent of tips go to the driver. The money was collected. It did not arrive.

There is a name for this in the literature. In 2025, Kadolkar published a systematic review in the Journal of Organizational Behavior. He names three forms of asymmetry that platforms build into algorithmic management: information, power, and calculative. The third one drives this case. The platform controls both the offer screen and the settlement formula. The worker sees the first number. The worker never sees the math that produces the second.

A 2025 Human Rights Watch report found the same pattern across the field. Six of seven major platforms use pay formulas workers cannot read until after the job is done. This is not one company's scandal. It is a structural pattern.

You are not a gig driver. But if you run an operation, you have accepted a number from a party that also controls the payout math. Revenue-share contracts where the split shifts after the deal closes. Marketplace payouts set by a formula you never approved. Performance bonuses built on metrics that change after the quarter ends. The mechanism is the same. The offer looks clean. The settlement runs through a formula you cannot see.

The better principle is short: audit the settlement math before you accept the offer math.

Once that is clear, three moves follow from it.

Pull up each deal where a second party controls how you get paid. Find the actual math that turns your work into their payment, not the rate card, not the promise in the pitch. If you cannot find it, that gap is the answer.

Most operators skip this step because the offer looked fair when they signed. That is the part worth sitting with.

For each deal, compare what was offered to what you received, in dollars and in percent. A gap under five percent may be rounding. A gap above ten percent is a pattern. A pattern means the formula is doing work you did not agree to.

When you sign or renew, write the payout formula into the terms, with inputs named and math spelled out. If the other side will not put it on paper, the gap between the offer and the payout is where they take their margin from your work.

Running this for one quarter does something the old advice never did.

You see which deals have a gap between promise and payment. You see how large it runs. You see whether it grows over time or holds flat. And you see which relationships let you read the math before you commit, and which ones are built so that you cannot.

At the end of the quarter, ask three things.

→ Which deal paid closest to the offer number, and why?

→ Which deal looked like good terms but left money behind after settlement?

→ Which formula did you have to ask for more than once before anyone produced it?

That is the difference between advice that sounds right and a system that proves itself.

The driver is still in the parking lot. The number on his screen still looks clean. The question is whether you can read the formula behind it before you tap accept.