We guarantee that what we build will return at least twice what you spend on it, in the first year.
People assume this is a marketing device. A risk-reversal play, the kind of thing you put on a landing page next to a stock photo of a handshake. And I understand the assumption, because that's what most guarantees are.
This one is pointed the other way. It's not primarily a promise to you. It's a constraint on us — and it changes almost everything about how we run the company.
What it forbids
Start with what we can't do anymore.
We can't sell a project we haven't priced. If we're going to guarantee a 2x return, we have to know what the problem is worth, in dollars, before we quote the work. That means every engagement begins with a real quantification — hours, frequency, loaded rates, error costs, opportunity costs — and if a client won't participate in that exercise, we can't work with them. Not won't. Can't. We'd be guaranteeing a return on a number we don't have.
We can't take exciting projects with weak economics. This is the painful one. There are builds we would genuinely love to do — technically interesting, fun to talk about, great for a case study — where the underlying workflow just isn't worth enough. Under a guarantee, we have to walk away from those. We do. It costs us revenue every quarter.
We can't overbuild. Every additional feature is additional cost on our side of the equation, and the return has to clear on the total. So the guarantee ruthlessly enforces scope discipline — not because we're virtuous about scope, but because gold-plating is now our problem, not yours. If we build a Ferrari to solve a Corolla problem, we eat it.
And we can't hand over something nobody uses. A system that returns nothing because it sits unused returns nothing. So adoption isn't a nice-to-have we mention in the SOW; it's the thing standing between us and a refund. We are extremely motivated to make sure your team actually uses what we build, which is a healthier incentive structure than any amount of good intentions.
What it does to the sales conversation
It kills half of it.
A normal build-shop sales call is a persuasion exercise. The prospect is uncertain, the vendor is enthusiastic, and the vendor's enthusiasm is doing the work that evidence should be doing. Everyone knows this on both sides. It's why buying software feels the way it feels.
Under a guarantee, the sales call becomes an arithmetic exercise, and arithmetic is not persuasive — it's just true or false. We show up with a number for what the problem costs, a number for what the fix costs, and the ratio between them. If the ratio is comfortably above 2, we say build it. If it's below, we say don't, and we mean it, and we've said it enough times now that people believe us when we say the opposite.
That's the whole trade. We give up the ability to talk anyone into anything, and we get something better in return: when we say a project is worth doing, that sentence carries weight.
The thing that makes it survivable
Here's the mechanic that makes a guarantee like this possible rather than reckless, and it's worth understanding because it applies to how you should evaluate any automation, ours or anyone's.
We only guarantee returns on workflows that are measurable, repetitive, and already happening.
Measurable: there's a baseline. We know it takes 5.2 minutes today because we watched and counted. Repetitive: it happens 400 times a month, every month, and it happened last year too. Already happening: we are not projecting a return on a behavior change, a new revenue line, or a market that might exist.
That's a narrow class of project. It's also, at owner-led companies, where the overwhelming majority of the real money is. We don't need to guarantee returns on speculative work because we're not doing speculative work — we're doing the unglamorous, deterministic, expensive-in-aggregate work that your business is already paying for by hand.
If someone offers you a guaranteed return on a project that depends on your salespeople changing their behavior, or on a new product finding a market, be skeptical. Those returns are real sometimes, but nobody can guarantee them, and anyone who does is either confused or lying.
What this should change for you
Even if you never work with us, take the frame.
Before you commission any automation, ask the vendor to state the annual dollar value of the problem, and ask them to show their work. Then ask them what return they expect, and by when. Then ask what happens if it doesn't hit.
The answers will sort the market for you in about four minutes. Most will not have a number, because most have never been asked for one — and the reason they've never been asked is that nobody's business model required it.
Ours does. That's not moral superiority; it's structure. We designed the company so that we lose money when we build the wrong thing, and it turns out that when you set it up that way, you stop building the wrong thing.
The result is a very short list of projects and a very high hit rate. Which, if you're an owner deciding where to put fifteen or thirty thousand dollars this year, is exactly the list you want to be on.