How we work, stated plainly.

Where these values came from

None of this was written at an offsite. Each one is a rule adopted after getting something wrong, which is the only way values ever become real.

Integrity over revenue came from an early engagement where the commercially attractive answer and the correct answer diverged, and we took a fortnight to notice we had drifted towards the first. Proof over opinion came from a report that went out carrying a finding we could not reproduce when the client's developer pushed back, which was a genuinely bad afternoon and permanently changed our review process.

Accountability over process came from watching a multi vendor programme where every party had done their part correctly and the outcome still failed, because nobody owned the join. Craft over volume came from the plain observation that our best work and our busiest quarters have never coincided.

How a value gets applied when it is inconvenient

A value is only load bearing at the moment it costs something, so it matters who decides. On our engagements the person leading the work has the authority to stop it. They do not need to escalate to decline a certificate, to tell a client the scope is wrong, or to delay a report that is not ready.

That authority sits with the engineer rather than with commercial management deliberately. The pressure to ship runs downhill, and a rule that can be overruled by whoever is closest to the invoice is not a rule.

Findings are also reviewed by somebody with no commercial stake in the account before the report leaves. It is a small firm and that is easy to arrange. It would be the first thing to break at ten times the size, which is one of several reasons we are deliberate about growth.

What integrity over revenue has actually cost us

Two examples we are comfortable putting in writing. We have scoped an annual penetration testing programme down to a single quarterly application test because the client's estate did not justify the rest. And we have declined a compliance engagement where the client wanted the certificate on a timeline that did not allow the testing behind it to be done properly.

Both decisions cost real money in the year they were made. Both clients are still with us, which is the commercial argument for behaving this way, though it is not the reason.

Declining work on timeline grounds is the most common reason we lose a bid. It is also the one thing on this page that is genuinely not negotiable, because a certificate from an empanelled auditor is only worth anything for as long as it means something.

How proof over opinion works day to day

• Every finding is reproduced by a second engineer before it enters the report
• Reproduction steps are written so your developer can follow them without calling us
• Severity is argued from demonstrated impact on your estate, not a generic score
• Anything we could not prove is listed separately as an observation, clearly labelled
• Scanner output never appears in a report under its own authority
• The walkthrough is a discussion where findings can be downgraded, not a presentation

What accountability means when something goes wrong

Engagements go wrong. A test window slips. A finding turns out to be a false positive. A remediation recommendation does not survive contact with your architecture.

What we commit to is that the person who owns your engagement tells you, in their own words, without it being escalated first. There is no account manager between you and the engineer, and no version of a status report written to look better than the week actually was.

Every engagement also carries a named second who has read the scope and can pick it up, which is the difference between a two day delay and a restart when somebody is unavailable.

Why craft is a hiring and funding decision

Craft is not something you ask people to have. It is something you pay for. Our engineers hold offensive, defensive, cloud and governance certifications, and the training and examination costs sit on our side rather than theirs. An engineer funding the firm's credibility out of their own salary is an arrangement that quietly fails.

Research time is protected rather than notional, and it appears in capacity planning, which is the only real test of whether a benefit exists. An engineer who has not looked at a new technique in six months is testing last year's estate.

Hiring is where this is actually enforced, because intellectual honesty cannot be trained in afterwards. Our technical interviews include a deliberately ambiguous finding where the impact depends on context we have not supplied. We are watching whether the candidate asks for the context or rates it confidently. The confident rating is the wrong answer every time, and it is the answer most people give, because certainty feels like competence.

The four, restated as commitments you can hold us to

• If you do not need the work, we will say so before you buy it, in writing
• If we cannot reproduce it, it is an observation rather than a finding
• The person who tested your system writes your report and attends your walkthrough
• Retesting after remediation is included, not quoted separately
• If we get something wrong, you hear it from the engagement lead rather than discovering it
• We will not publish your name, logo or engagement detail without written permission

The pressure that works against all four

It is worth naming rather than pretending we are above it. Clients want the certificate before the board meeting. Sales targets exist. A finding that will delay a launch is unwelcome, and the person delivering it knows it.

What keeps the four intact is structure rather than virtue: the engineer holds the stop authority, findings are reviewed by someone with no commercial stake, and the firm is small enough that a bad decision cannot hide inside the volume.

Any firm that tells you it feels no commercial pressure is either not being honest or not being paid.