What we decline

An agency that says yes to everything is optimising for its own next quarter, not for your outcome. Here is what we walk away from, and why.

Staff augmentation, priced as a project

If the real ask is three bodies for six months with a delivery manager between them, that is staffing. We will say so, and we will not dress it up as a fixed-price engagement — the two have opposite incentives, and pretending otherwise is how resentment starts in month two.

Work with no owner on your side

Every engagement needs someone who can answer a question within a day. Without that, schedules slip in silence and the first sign of trouble is the deadline. We ask who that person is before we quote, and "we will find someone" is not an answer.

Rewrites with no stated reason

"We want it in Rust / Go / whatever is current" is not a reason. A rewrite with a written reason — this cost us this, that will cost us that — can be excellent work. The same rewrite without one is a gamble with your budget.

Systems we cannot instrument

If telemetry is refused, metrics are out of scope, or production access is permanently off limits, we cannot tell whether what we built works. We would be selling you confidence we cannot verify.

Engagements that begin mid-crisis with a fixed date

Real emergencies deserve a different shape: a short, capped diagnostic first, then a plan. Committing to a delivery date before anyone has seen the system is how both sides end up disappointed.

What we say instead

Usually a diagnosis — a couple of days' work that tells you what the actual problem is and what it would take. It is cheaper than a bad engagement for you and more honest than enthusiasm from us.

If we have said no to you, it was probably one of these five. If it was something else entirely, we would rather hear about it than not.