Every services company we benchmarked before launching Relay had tiers. Junior, mid, senior. Or bronze, silver, gold. Or T2, T3, T4. Customers picked a tier when they signed; the engineer they got depended on what they had paid for. The economics make sense on a spreadsheet. We don’t do it.
We have one bench. Every engineer on it is senior by every reasonable measure. A press routes by stack and by availability, not by what plan the customer is on. The customer who pays the lowest price gets the same engineer the customer who pays the highest price gets.
A tier is a way of telling some customers their problem isn’t worth a software engineer. We don’t believe that.
Three reasons we do it this way. They’re ordered by how much we believe in them.
One. The press is short.A median Relay session is twenty-eight minutes. Inside that window, the cost difference between a senior engineer and a junior one is the cost difference between the customer shipping and not shipping. We can’t price the difference cleanly enough to charge for it; the spread we’d introduce by tiering is larger than the spread we’d capture in revenue. So we don’t.
Two. The promise breaks the moment you tier.“Click the green dot. A real engineer joins in seconds.” If we tiered, the promise would have an asterisk. A real engineer of the seniority you have paid for joins in seconds. The asterisk would make the product worse, in a way the spreadsheet wouldn’t catch.
Three. The bar is easier to maintain at one level than at three. We can hire for one bar and write down what that bar is. We tested the alternative on paper and concluded it would slow our hiring, complicate our matching, and create a tier-of-engineer-to-tier-of- customer routing problem we did not want to solve. The flat bench is a hiring discipline as much as a customer promise.
The cost of this choice is real. Some presses don’t need a senior engineer; they could be solved by a competent mid-level one for half the price. We absorb that delta. We have decided we’d rather absorb it than write a routing system that asks customers, every time they press, what version of help they would like to pay for.
We’ll say what would change our mind. If we found that more than a third of our presses were genuinely well-served by a less senior engineer, and we found a clean way to identify those presses before routing, we would consider it. We have not. Our internal data suggests the opposite: a press that looks junior at first turns out, in two thirds of cases, to need senior judgment by minute four. We prefer to route to the seniority the press will need by minute four, not the seniority it looks like at minute zero.
That’s the engineering reason. The customer reason is simpler. Every customer we have wants the software engineer. They have always wanted the software engineer. The reason there are tiers in our industry is that the software engineer was scarce. We have organized the company around making the software engineer not scarce. There is nothing else to charge for.
Engineering, Relay. April 2026.