Find out what the software should be, before you commit the budget.
Pathfinder Labs Corporation is a software discovery consultancy in England. Bring us an untested product idea and we put working software in front of real conditions, then hand you the evidence: a working model your users can click through, a written record of what held up and what did not, and a specification you can put out to build.
Thirty minutes, no charge, and we say in the call whether this is work we should take on. Every email answered within two working days.
Three kinds of question, one kind of answer: something running.
Each of these ends with software you can execute and a document you can hand to somebody who was not in the room. All three are available now, and we can start one inside a fortnight.
-
01 · Feasibility probes
Answer the question a document cannot
You have a technical question a document cannot answer: whether the data is good enough, whether the technique is fast enough, whether the integration holds. We build the smallest thing that settles it and write down the result, including the negative one.
-
02 · Working models
A version you can put in front of people
A working, clickable, deployable model of the idea, running on real data with real latency and real failure modes. Good enough to test with your users, and detailed enough for a development firm to quote a fixed price against.
-
03 · Technical due diligence by construction
Implement the risky part twice
Before you pick an architecture or sign with a vendor, we implement the risky part both ways and report which one survived contact with your workload.
Four stages. Each one timeboxed and priced before it starts.
We take one engagement at a time, so the stage you are paying for is the work in front of us.
-
01 · 1 week
Frame
We write the question down and agree what would count as an answer. If we cannot agree on that, there is nothing to build yet and we say so.
-
02 · 1 to 2 weeks
Probe
We build the narrow version that settles the question. It answers one thing conclusively rather than doing many things approximately.
-
03 · 3 to 5 weeks
Model
We build the version you can put in front of people: clickable, deployed, running on real data with real latency and real failure modes.
-
04 · 1 week
Hand over
Source code, a written findings memo, and a specification another firm can quote against. All three are yours to act on without us.
Two rules that sit above the four stages
You own the code and the findings outright from day one. And if the honest answer at the end of Probe is that you should not build this, we say so and stop billing.
The findings memo, six fields, fixed before the work starts.
The format is agreed at Frame, so a result cannot be written up more favourably once it is in. You get it whether the answer is the one you were hoping for or not.
Every memo carries the same six fields
One document, written in the language of the decision you have to make, short enough that a board will read all of it.
- The question
- One sentence, written before the code, with the condition that would count as an answer.
- What we built
- What the working model does, and what it deliberately leaves out.
- What happened
- The result, stated plainly, with the measurements behind it rather than an adjective.
- What we would not rely on
- The parts of the answer that are thin, and what it would take to firm them up.
- What to do next
- A recommendation, and the specification you would hand a development firm if you take it.
- Where the code is
- A repository you own, with every component and its licence listed.
Want to see the blank format before you commit to anything? Ask us for it by email and it comes back the same working day or the next one.
We would rather hand you a short memo that says do not build this than a long one that says maybe.
Evidence, not opinion
An opinion about whether something will work is cheap. A running artefact under real conditions is different: you can point at it, argue with it, and take it to somebody else for a second reading.
That is what you are buying here, and it is what you leave with.
Timeboxed and priced before it starts
Every stage has a duration and a price agreed in writing before work begins. If a stage overruns because we misjudged it, that is our cost, not yours. If it finishes early, the next stage starts early.
You own it outright
Source code, findings, notes and any negative results belong to you from day one. There is no licence to renew, no repository we hold the keys to, and nothing that stops you taking the work to a different team the week we finish.
One engagement at a time
We run one engagement at a time, which is why the stages are short and why the dates in a scope are dates we can keep.
One person is accountable, and you meet them first.
The person who writes the code writes the findings memo and sits in the meeting where the recommendation is delivered. There is no account manager between you and the work, and nobody appears on the invoice who was not in the room.
The director of THE PATHFINDER LABS CORPORATION LIMITED is on the public register at Companies House against company number 17061706, which you can check without asking us.
Tell us the question. One paragraph is enough.
[email protected]We publish an address rather than a form. Mail lands with a person, which is quicker, and it leaves the copy of what you sent in your hands as well as ours.
- Response time
- We reply to email within two working days, Monday to Friday, UK time.
- The scoping call
- Thirty minutes, no charge, no slide deck. You leave it knowing whether there is an engagement here and roughly what it would cost.
- What helps
- The question you want settled, any deadline you are working to, and what you would do differently if the answer turned out to be no.
- Not us
- Several unrelated businesses use similar names. See the note about our name before you write.