We call it the hard look. Several people go at the thing you're about to build — deliberately from different angles, deliberately not the same angle twice — and you get back the specific ways it comes apart, in order, worst first. It's a list of what breaks, not a verdict on whether you'll succeed.
By the time a build has been agreed, everyone in the room has spent a fortnight talking about what it will do when it works. That is the easy half, and it is the half that gets all the attention. The expensive half is the other one: the supplier who sends the numbers as a PDF, the second location that does it a different way, the person who has been quietly fixing it by hand for three years, the software you already pay for that won't hand its data over without an upgrade.
Those are the things that turn a four-week build into a four-month one. None of them are secrets. They're just nobody's job to go looking for, because the person who could find them is the same person who wants it to go ahead.
So the hard look is somebody whose only job is to find them, before anyone has spent anything.
One reader looking for trouble finds the trouble they already had in mind. So the plan gets attacked several separate times, by readers briefed to care about completely different things, and none of them sees what the others wrote until they're finished.
The exceptions, the manual step nobody documented, the customer who is handled differently because they've always been handled differently.
What your existing software will and won't give up, what it costs to get at, and what breaks the next time it updates.
Who has to change how they work on Monday, who wasn't asked, and what everyone quietly goes back to when it's busy.
What happens when it's wrong, when it's down, when two people do the same job at once, and how long it takes anyone to notice.
Everything that comes back is put in one pile, the duplicates are merged, and then the rule that matters: anything nobody can check does not go in the document. If a finding can't be pinned to something you told us, something we can see in your systems, or something publicly true about the software involved, it gets dropped — however clever it sounds. A pre-emptive horror list is easy to write and worth nothing.
What survives that gets ranked, and not by how likely it is. By what it costs to be wrong about it — a small chance of something that stops the whole build outranks a near-certainty you'd fix in an afternoon.
It's normally somewhere between four and eight findings. If it were twenty, most of them would be padding, and you'd stop reading before the one that mattered.
$750 for the hard look. Builds still start at $4,500, and if you go ahead, the $750 comes straight off.
That is the whole arrangement. The build price is the same number it was before this page existed — we didn't put the build up to pay for the credit, which is the usual trick and the reason most people rightly assume a credited deposit is theatre. If you have the hard look and then build with us, the hard look has cost you nothing. If you have it and build with somebody else, or don't build at all, it has cost you $750 and you keep every page of it.
About forty-five minutes. Mostly us asking, and you'll know by the end of it whether this is worth doing at all.
That's the actual work. You don't hear much from us in the middle of it.
Sent before the call about it, so you've read it properly and can argue with it.
No retainer attached, nothing that renews, and no obligation to build anything.
The case where it costs you
It happens, and when it does there's no build for the $750 to come off. You've paid $750 to not spend several thousand more, which we'd argue is the cheapest version of that lesson available — but we're not going to pretend it's free, because it isn't.
And the same rule as everywhere else on this site applies before you've paid anything: if the honest answer is that a setting in software you already pay for fixes it, you'll hear that on the first call and there'll be no invoice.
Worth being clear about this before you pay for it, rather than after. Four things this is not, and one limit we can't get around.
We don't predict outcomes and we won't pretend to. Anybody selling you a confidence score on your own business is selling you a number they made up.
No revenue forecast, no payback period, no "this will save you X hours a week". We don't know your margins and we're not going to estimate them at you.
Nobody can hand you every risk. We can hand you the ones we can find and stand behind, and tell you plainly where we stopped looking.
Anything we can't check gets cut, so the list is shorter than a scary one would be. That is a deliberate trade and you're paying for it.
The one thing it does tell you: where this goes wrong, what that costs, and what you'd do about it. Everything else on this page is us declining to sell you something we can't actually do.
The read is written from the outside, off what you can tell us in one box and what's already public. It costs you nothing, it asks for nothing, and it always will — that isn't a trial version of this page and it doesn't lead anywhere you have to pay to get to.
The hard look is a different thing. It needs your real numbers, a look inside the systems you actually run, and about a week of proper work at our end. Charging for it is the straightforward reason, and we'd rather say the other reason out loud too: a plan nobody has paid to question tends not to get questioned. Free advice gets read once and filed. A document somebody has paid for gets argued with, which is the entire point of it.
It also means we can afford to say the uncomfortable thing. Nobody who is hoping to be given a build is well placed to tell you not to have one.
Our website page has carried a guarantee for months: live within 48 hours of your photos, or your first month is free. It had been read, approved and left alone by everyone who looked at it, including us.
The same page also says, in bigger type, nothing to pay until it's live. So if we missed the 48 hours, the penalty was refunding a month nobody had been charged for yet. A promise that reads like a promise and costs the person making it precisely nothing.
Nobody had complained. Nobody had noticed. It took one reader whose only job was to attack the offer about ninety seconds. The line now reads the first month after it goes live is on us, which is a thing we can actually be held to — and you can go and check that it says so.
What we're not claiming
This is a new thing and we've run it on our own work, not for a paying customer. There are no case studies on this page because there are none to tell, and we'd rather the first thing you read about us was true than impressive. When there is one, and when the client is happy to be named, it'll be here with their name on it.
What you can check today is the rest of it: six pieces of software we built and still run, and a free read you can have in two minutes without giving us an email address.
A paragraph is enough: what you're about to build or buy, who has to use it, and roughly when. If you've already had a read from us, just reply to it and say you want the hard look. We'll come back with whether there's one worth doing here — and if there isn't, we'll say so and there's nothing to pay.
$750, one fixed fee, agreed before anything starts. Credited in full against the build if you go ahead. Nothing is invoiced until we've both agreed it's worth doing.
Not ready for that? Have the free read instead — one box, your own words, no email address, and it'll tell you if you don't need us at all.