We build the plan with you and keep improving it
We stay in it for as long as we’re adding value, and not a week longer. Every company comes to us at a different point, so find thestep below that sounds like yours today.
One loop, repeated
We start with discovery. It is a short, focused piece of work with you: the domain, your customers, whatever already exists, and the outcomes that would make the next few months worth it. It ends with a plan that belongs to you, whatever you decide to do next and whoever you decide to do it with.
From there the plan runs through the loop from lean product practice that Eric Ries called build, measure, learn. We build the smallest thing that tests the next assumption. We measure what happened against the outcomes you set. Then we learn, which is discovery coming round again: what we found changes the plan, and the next loop starts from the changed plan rather than the original one. So each loop compounds on the last, and the plan you have in month six is better than anything we could have written in week one.
Value is yours to define
It might be revenue, retention, a funding milestone, or the hire that unblocks everything else. We write it down together and come back to it whenever the plan changes. And every week we ask one question out loud: was this week’s work worth what it cost you?
If the answer is no for long enough, we shouldn’t be there. You can step back or stop whenever the value isn’t there, without a long contract holding you in place. And I’ll say so before you have to, including when the honest recommendation is less of my time.
Alistair Stead, founder
Where you are
Each step lists what we do there and what you bring, then what gets decided before the next one. You can start at any of them. A company with an established team often comes in at the third or fourth, and a bespoke training workshop is a way in at any point.
Step 1 · Discover and experiment
You have an idea and some evidence
You know the problem is real, and you have some proof: a few conversations, a waiting list, a prototype someone built in a week. What you don’t have yet is certainty about which assumptions the whole idea rests on. We immerse in the domain with you, talk to your customers, review whatever has been built so far, and run the cheapest experiments that settle the assumptions everything else depends on.
- What we do
- Domain immersion, customer conversations, a review of what exists, and the first experiments.
- What you bring
- Access from day one, the people who know the customer, someone who owns the commercial side, and the outcomes you care about most.
- What gets decided
- What to try first, and how we will both know it is working.
Usually through Prototyping and experimentation.
Step 2 · Prove and build
You have something people are using
Real users are in the product now, which means the evidence gets better and the mistakes get more expensive. We build alongside you and keep every change small enough to measure and cheap enough to reverse.
On an AI adviser for a broadband comparison platform, we tested it against real customer conversations before any customer saw it, and the results went into a board meeting as evidence. The adviser then went live to 10% of users behind a feature flag, a switch that meant the first real-world test could be undone in a click.
- What we do
- Build with you, and run small experiments in production that we measure against the outcomes we agreed.
- What you bring
- Customers to learn from, and a willingness to act on what you learn, including when you learn to stop.
- What gets decided
- What is worth investing in properly.
Usually through Early-stage product bootstrapping.
Step 3 · Build the team
You have a product and need a team
The plan now needs more hands than ours. We help you hire the people it needs, once you have the evidence that you need them, and set up delivery practice that holds once we step back. The people we help you hire shouldn’t need us for long.
Delivery practice belongs inside this step because the loop only runs as fast as you can ship safely. Google’s 2025 DORA research found that “teams working in loosely coupled architectures with fast feedback loops see gains, while those constrained by tightly coupled systems and slow processes see little or no benefit.” So automated testing and fast, safe deployment are part of the work here, scheduled alongside the hiring rather than after it.
For a marketing group’s analytics team, we agreed a new data architecture with them on a Thursday, built it on the Friday, and handed it over the following Monday with no rework. The build was quick because the plan was already theirs.
- What we do
- Lead the technology, hire, coach, and set the delivery practice the team will keep.
- What you bring
- The budget to hire, and leaders who will own what gets built.
- What gets decided
- The shape of the team, and who leads it.
Usually through Fractional and interim CTO and Technical and product team coaching.
Step 4 · Advise
Your team is leading
Your team runs the product and the delivery without us in the work. We step back to an advisory role: a view at board level, and a standing line to whoever leads technology, so the advice comes from knowing the team rather than reading a pack once a quarter.
This is the step where the weekly question matters most, because an advisory retainer is the easiest thing to keep paying for out of habit. It has to keep earning its place like everything before it.
- What we do
- Ask the questions an investor will ask, and support whoever leads technology between board meetings.
- What you bring
- Honest feedback on whether it is still worth it.
- What gets decided
- Whether you still need us. When the answer is no, the engagement has done its job.
Usually through Board advisor.
What holds at every step
- Not everything needs discovery
- Discovery is for problems where the answer has to be found. Plenty of problems have known answers, and paying us to rediscover them would be waste. The test we use: could a competent engineer who has done this before tell us the answer in an afternoon? If yes, it is delivery, and we treat it as delivery.
- Small steps, long relationships
- Each loop is a complete piece of work with its own result, and you can stop after any of them. Most of our relationships still run for months, because once we are inside your team the next experiment costs a conversation instead of a proposal. Most consultancies sell large, fixed-scope projects because that overhead makes small ones uneconomic. We have written about why agencies that sell agile still ship waterfall.
- A bad result is worth hearing
- When a consultant’s next contract depends on this piece of work looking successful, an inconvenient result becomes awkward to report. Because we stay only while the work earns its place, telling you something didn’t work costs us nothing and saves you a loop. Ask any consultant how their incentives handle a negative result.
- Cheap software makes the wrong thing cheap too
- Software is quicker to build than at any point before. The risk now sits in whether something should be built at all, which makes discovery worth more now than it has ever been. Your runway is better measured in questions answered than months left, and the most expensive outcome is still building the wrong thing well.
- Written by default
- Updates are written down, so there is always a record you can forward to a board or an investor. How often we meet is something we agree together and change as the work changes.
What makes this work
A partnership has two halves, and this is yours. Someone on your side has to own the commercial direction, because a plan with nobody selling it measures nothing worth measuring. We need access to the code, the data, the accounts and the people from the first week, since discovery stalls on a missing login faster than on a hard question. And you have to be willing to act on what you learn, including when it shows the thing you were sure of is wrong. Bring those, and the rest of this page is what you get back.
Start with a conversation
30 minutes, no pitch, and a straight answer on whether Kodehort is the right fit.