Before we start

Ten questions, and why we ask them

Most software projects go wrong because the questions at the start were wrong. Ask fifty about screens and you still miss the handful that decide whether a build takes three months or three years. These are that handful. None of them are technical, each one says plainly why it matters, and a sentence is enough.

15 minutes
Ten questions, then four about how you want to buy it.
Nothing technical
If you can describe how the work gets done today, you can answer these.
A real reply
A considered read on your project within one business day, from an engineer.

What happens next: we read your answers, come back with what we think it is and what we would want to look at, and then — if it still makes sense — spend ninety minutes with you, ideally where the work actually happens. That conversation is the one that produces a specification. This page is what makes it short.

Who you are

So we can reply. Nothing else is done with it — see our privacy policy.

The ten

The questions that decide the build

Each of these changes what we would build and how long it takes. A sentence is enough — we go deeper together afterwards.

About 15 minutesNothing technicalYour draft is kept in this browser

Who will use it — and is anyone outside your organisation?

An internal tool and a system your clients, customers or vendors log into are different products. External users bring sign-up, password resets, permission boundaries and a support load.

Who will use it — and is anyone outside your organisation?

What are you using today, and are we replacing it?

Years of records in a format nobody controls is a project in its own right, so one of our first jobs is proving your data can come out of it cleanly.

What are you using today, and are we replacing it?

Does anything here answer to a regulator or an auditor?

Pharma manufacturing (GMP, 21 CFR Part 11), financial services, insurance, health records and personal data under the DPDP Act each carry obligations on audit trails, retention and validation. Those have to be designed in from the first day — they cannot be added to a finished system.

Does anything here answer to a regulator or an auditor?

Does it touch money?

Invoices, GST, payments, payroll, reconciliation and refunds demand a different level of rigour — every figure traceable to a document, and periods that will not close while the books disagree.

Does it touch money?

Does it have to talk to something you don't control?

This is the schedule risk nobody can manage. A government portal needs a licensed intermediary; a bank API needs their onboarding; a vendor's closed database needs a licence that is a negotiation, not a download.

Does it have to talk to something you don't control?

What happens if it is unavailable for a working day?

Going back to paper for a day and stopping a production line need very different things built around them — how much redundancy, how much monitoring, how quickly a restore has to work.

What happens if it is unavailable for a working day?

How many users, records and transactions — now, and in two years?

We need the order of magnitude, not a precise number. Fifty invoices a month and fifty thousand are different architectures.

Who signs this off? Who else can say no?

The commonest reason a project goes sideways is not technical. It is a second decision-maker who appears in week six with a different idea. We would rather meet them in week zero.

Who signs this off? Who else can say no?

Is there a date? What happens on that date?

A deadline with a reason behind it — a financial year, a rule change, a trade show, a contract ending — is real and we can plan to it. We need the reason more than the date.

Is there a date? What happens on that date?

If we could only build a fifth of this, which fifth?

The most useful question here. It tells us what to ship first, and that part is usually the one that starts being useful on its own.

Fit

How you want to work

These decide as much as the ten above: how soon you need something in your hands, and what you have to work with. We ask now because they change what gets built — rather than later, when they change what gets re-built.

Three minutesSay “rather not” to anything

Would you rather have part of it working soon, or all of it later?

There is no right answer, and the two lead to genuinely different projects. Starting with one useful piece means you are using something while the rest is built, and you can change your mind about the rest. Waiting for the whole thing means fewer moving parts to live with in the meantime.

Would you rather have part of it working soon, or all of it later?

What have you set aside for this?

We ask so that what we propose is something you can actually say yes to. A build we would scope at four months is not worth describing to someone who has set aside six weeks of work, and finding that out at the proposal stage wastes your time more than ours. A range is enough, and “rather not say” is a complete answer.

What have you set aside for this?

How would you rather buy this?

A fixed-scope project suits a clear, bounded requirement. A monthly fee including hosting, support and changes suits something that will keep evolving. Our engineers inside your team suits a long programme.

How would you rather buy this?

Who should own the software at the end?

For a bespoke build you own the source, the domains and the cloud accounts, in your name from day one. For a licensed platform we keep the code and you own your data, with export whenever you ask. Both are legitimate; they are different contracts, and far simpler to settle now than later.

Who should own the software at the end?

Will you use this yourselves, or sell it on?

If your own clients will use it, that is a platform rather than an internal tool — billing, onboarding and support tiers all have to be built, and the contract between us is a different one.

Will you use this yourselves, or sell it on?

Anything you already know will be a problem?

An IT policy, a sceptical partner, an approval cycle, a previous project that went badly. Telling us now saves us both a wasted month.

Optional

What you can send us

A description of a process is never as good as watching it. Tick whatever you can get to us — every one of these takes guesswork out of what we come back with.

0 of 10 answered — question 10 is the one we really need

You do not have to answer all ten. Send what you know and we will ask about the rest.

Your draft stays in this browser until you send it. We use your answers to work out what your project needs, and for nothing else.