How we deliver
Standards you can hold us to.
Every firm says enterprise-grade. Almost none of them say what it would mean to fail. These are ours, written so you can check. [Draft copy]
Why this page exists
"Enterprise-grade" is doing a lot of work in our positioning, and a claim that cannot be checked is marketing. So this page states what we actually hold ourselves to, in terms specific enough that you could point at one and say we did not do it.
It is also the page we would want if we were buying. A firm that will not write its standards down is asking you to take the standard on trust, which is the opposite of what we are selling.
The standards
-
Every design decision is recorded, with the option we rejected
A decision without its alternative is an assertion. Short architecture decision records, written when the decision is taken rather than reconstructed later, so the reasoning survives the people.
-
Evidence is produced during delivery, never assembled afterwards
Control registers, test results, approval records and configuration baselines fall out of the work as it happens. Evidence written the week before an audit describes the estate somebody hoped for, not the one in production.
-
Controls map to your obligation, not to a framework we prefer
We start from the regulation, the questionnaire or the finding that put you here, and map backwards. A control set chosen for our convenience is a control set you will have to redo.
-
Nothing we build requires us to run it
Conventional structure, no in-house cleverness, documentation written during the build. A competent engineer you hire next year should be productive in it without calling us.
-
The person who scoped the work is on the work
Not as an escalation path. In the standups.
-
We say no to work we cannot staff properly
The constraint is real and we would rather decline than staff you thinly and discover it in month three.
[Draft copy] — these are the commitments the firm would be held to publicly. Each one should be reviewed by the principals before launch, because a standard you cannot keep is worse than no page at all.
How to check we did it
Standards that cannot be tested are decoration, so here is what to ask for. You do not need our permission and you should not wait until the end.
| Ask for this | At this point |
|---|---|
| The decision record for a choice you disagree with | Any time. It should already exist. |
| The evidence pack, in its current incomplete state | Mid-engagement, not at handover. |
| A walkthrough of something by someone from your team | Before we finish, while we can still fix the gap. |
| The list of things we have not done and why | Every steering conversation. |
If we cannot produce the artefact on the day you ask for it, it did not exist and we were going to write it later. That is the thing worth catching early.
What we will tell you
Three commitments about bad news, which is the part of a delivery relationship that actually differentiates.
- If the work is not what you need, we will say so before you commission it, including when that ends the conversation.
- If we get something wrong, you will hear it from us before you find it.
- If we cannot put a founder on your work, we will tell you in the first conversation rather than in month three.
[Draft copy]
Who owns it after
Every engagement has an exit condition, and it is not the invoice. It is that a named person on your side owns each deliverable and can run it without us. That handover is scoped and scheduled at the start, not negotiated at the end.
The Academy exists for the same reason. A dependency on us is a failure state, not a business model. [Draft copy]
Tell us what you are being asked to prove.
Ask us for the standard we would be least comfortable being held to.
Start a conversation