The first win · rung 1
Pay when it works. Not before.
The terms
- Price
- $12,000fixed
- Time to delivery
- 10 working days
- Due upfront
- Nothing
- Invoice raised
- On acceptancenet 7
We do not call this free. Free prices expertise at zero and attracts people who value it accordingly — this is post-paid, which is a different claim. The price is fixed and agreed in writing before day one, and it is the same figure whoever is asking.
The commitment
The scope limit and the remedy.
Both stated here rather than in the contract you have not read yet. A date on its own is a boast; a date with a named limit and a named consequence is a commitment.
The boundary
What fits in ten days, and what does not.
Taking a two-week job that was really six means we eat the cost and lose the deal. So we publish the second column with the same weight as the first.
✓ Fits in ten working days
- One internal tool replacing a spreadsheet everyone hates
- A single workflow: intake → approval → notification
- One report that currently justifies a BI seat
- One integration that removes a manual copy-paste job
- A migration plus a read-only view of the result
- The ~15% of your CRM you actually use, as a standalone view
✕ Doesn’t fit — so we won’t promise it
- Accounts, roles, admin and billing as a suite
- Multi-tenant anything
- More than one integration
- Data migration with cleansing and reconciliation
- Anything gated on your security review completing first
- Mobile and web together — pick one
How it runs
Five steps, and the one that matters is the first.
- 01
We agree the acceptance criteriabefore day one
In writing, before any work starts: what it does, with whose data, and what "working" means. This is the whole mechanism. A pilot with fuzzy criteria becomes an argument in week three, and the party with less money loses that argument.
- 02
We build it in ten working daysdays 1–10
One workflow, end to end. The scope does not change mid-build — not because we are inflexible, but because a ten-day promise with a moving target is not a promise.
- 03
You use it with real dataevaluation licence
It runs on our infrastructure and you use it exactly as you would in production. You get full access to the running application; you do not get the source repository yet. That is how every software trial works — you are being asked to pay for something you have used, not on faith.
- 04
You accept, or you do notday 10
Judged against the criteria from step one. If it does what we agreed, we invoice, net 7. If it does not, you owe nothing and you keep the write-up — the analysis is yours either way.
- 05
On payment, it becomes yourshandover
The repository transfers, the IP assigns in full, and we deploy it into your own cloud. One clause in the agreement, not a promise on a webpage.
The awkward questions
Why would you take this risk?
- Why would you work for ten days before invoicing?
Because the expensive thing for a young firm to acquire is trust, and this is us buying it at a capped price. The market is trust-poor and upfront payment is the blocker, so we removed the blocker and carried the risk instead.
It is capped deliberately: ten days, one workflow, and a limited number of pilots running at once. An offer we cannot staff is worse than no offer.
- What stops us using it and not paying?
During the pilot it runs on our infrastructure, so we can stop serving an unpaid pilot, and using it without paying is a breach of the evaluation licence rather than a grey area. Those two things do different jobs on purpose.
What we do not do is protect ourselves covertly. No hidden dependencies, no obfuscation, no kill switches, no undisclosed retained ownership. The pitch is that you own it outright — a pilot with a booby-trapped repo is a rental with extra steps, and it is the first thing any client engineer checks.
- We can’t let it run on your infrastructure.
That is a reasonable position and not a disqualifier. Three alternatives, in the order we would suggest them: run it in your cloud under an IP-assignment-on-payment clause with a 50% deposit; or split the ten days into two five-day milestones invoiced on each; or skip the post-paid terms and take the standard fixed-price engagement.
Never both. We do not ask for our hosting and a deposit. One protection per pilot, and you pick which risk you would rather carry.
- Who owns the data the pilot produces?
You do. The records, approvals and notifications created while the pilot ran are yours regardless of the commercial outcome — exported to you on payment, or on request if you decline. Client data is never leverage, in any dispute, at any stage.
- Can our engineers see the code before we accept?
Yes, under NDA, on request. Source access is not required to judge whether software works, which is why the default is app access — but refusing an inspection would be the covert-lock-in instinct reappearing in a different form.
- Is the pilot production?
No, and we will say so in writing. A pilot environment carries no service level and must not be relied on as production. If you want it to be production, that is what the handover and Own & Operate are for.
Pick the smallest useful thing and we’ll build it.
The teardown ranks your stack by payback, so the first win is chosen on arithmetic rather than on which tool annoys you most. Four minutes, no email.