CONSULTING
Review your code. Teach your people. Or build what's next, right the first time.
Usually that's three different firms, and each one's answer is that the other two got it wrong. Here it's one, so whoever reads your code is the same person who has to live with what gets built on it. Twenty-eight years of that, and more than $5M taken out of other companies' running costs.
01 · Review
Technical debt, in the code and in the process around it.
You already have software. Somebody built it, and nobody has read it since.
Found the way a building inspector finds mold: get rid of it, then put systems in place so it can't come back. The code is half of it. The other half is how work reaches your customers — how long a release takes, how many environments you are paying for, what breaks at seven on a Friday, and who gets called.
- A graded dossier. Every finding cites a file and a line and ships with a prompt that fixes it.
- What it costs to run, itemised, against what it needs to cost.
- The three things worth doing first, in order, with a number on each.
- The fixes cleared for you, as reviewable pull requests, if you would rather not do it yourselves.
02 · Train
AI-directed development, taught to your team.
Your team is using AI already. Nobody can tell you whether it is helping.
Directed badly, it produces 40,000 lines nobody has read. Directed well, it compresses the work far enough to change what a budget can reach. We have put AI-assisted development across four teams and measured a 40% productivity gain, and delivered twelve hours of it to other software leaders at a company conference.
- A working agreement your team will actually follow: what the model may touch, and what it may not.
- Review standards written for code nobody typed.
- The guardrails in your pipeline rather than in a slide deck.
- A before and after measured on your repository, not on ours.
03 · Build
Built from your business, not from the wish list.
You are building something new, or rebuilding what a previous vendor left.
On a new build the job inverts: we already know what the technology can do, so the time goes into your business. What a client asks for and what their business actually needs are frequently two different things, and the gap between them is where most software budgets die. Companies also have hard and easy the wrong way round, bracing for what turns out to be two days of work and waving through the thing that eats three months.
- With your team, in your repository and your standups, on the part nobody internally has time or background to take.
- Or built and handed over, audited by the same system we sell to everyone else, with the report attached.
- Written to be read by whoever comes next, not to demo well.
- Handover includes whatever is left on the remediation list and the prompts to clear it.
WHERE THE MONEY IS
The money is rarely where anyone expects.
Four places, in more or less this order, on nearly every engagement we have taken.
Sized for a peak that never arrives, running at full capacity to serve a Tuesday.
Being done by hand, every week, by people who could be building something instead.
It fit the business four years ago. The business moved and the architecture didn't.
Quietly turned into an insurance premium somebody in finance now pays every year.
WHO YOU'RE WORKING WITH
Twenty-eight years of other people's software.
Rüadán Archer, principal. Started in 1998 at Nortel, a 100,000-person telecom company and Cisco's biggest rival at the time. Since then: healthcare platforms carrying millions of transactions a year, cloud infrastructure for Fortune 500 names, a consultancy from 2017, and a director's chair with 23 people and millions in budget.
- Nortel Networks
- Verizon
- Party City
- Advantix
- Alegeus
- Samaritan Ministries
- Hattori Hanzo Shears
Employee at some, consultant at others, founder with a team behind him from 2017 to 2021. Which number below came from which name is between him and them.
Measuring0/9
HOW AN ENGAGEMENT WORKS
We look before we promise.
Nobody can price work they haven't seen. Anyone who quotes you before looking at what you already have is guessing, and you are the one who pays for the guess.
01
Read
The software, what you are spending to run it, how long a change takes to reach your customers, and what has been going wrong. Scope and cost are settled before anything starts.
02
Name the three
Not forty findings. The three that pay, in the order we would do them, with a number on each. In plain English, in a document that is yours.
03
Do the work, or hand it to your team
Some clients have us lead it. Some hand the document to their own people and never call again. Both are fine. What we will not do is leave you dependent on us.
What the read produces is yours, whether or not anything follows it. Not as a way to start billing you before the work does. It is the only honest way to price the work, because a consultancy that quotes before reading your code is quoting a guess and you are the one carrying it.
LET'S TALK
What you are actually buying is judgment.
Which things are possible. Which are expensive. Which will not work at any price, no matter who is telling you otherwise. Knowing that before the money is committed is what nearly three decades and a mind built for systems buys you.
The date isn't real. The vendor isn't delivering. The plan is being protected instead of the outcome. Everyone in the room already knows, and the reason nobody says it is that saying it costs something. I say it, and I try to say it in love rather than for sport. It has cost me work. It has never cost me a client.
Rüadán Archer, principal
Either one is enough. The call is free, and we will tell you on it if we are the wrong people for the job. If we are, we usually know who is right.