This article is part of our Journal archive. Any prior offers reflect its publication date. Read our current services and approach.
Most of what a practice publishes argues for hiring it. This one runs the other way: six situations where we are the wrong call, and how to tell before anyone has spent money.
Most of what a software practice publishes is an argument for hiring it. This is the other one.
Below are the situations where we are the wrong call. We would rather say so here than discover it together in week three, when you have spent money and we have spent goodwill. If you read this and recognize your own project, the article is working.
You need someone standing in the building
Some problems only become visible when you are physically there. A counter process that falls apart in the Saturday rush. A warehouse routine nobody has written down, which three people run three different ways. A front desk where the real workflow lives in one person's head and changes depending on who is on shift.
We work remotely most of the time. We can visit, we can watch, we can interview. What we cannot do is absorb a month of floor time by osmosis. If the honest scope of the job starts with someone shadowing your staff for weeks, hire someone who can shadow your staff for weeks. The software question comes after that, and it will be a much better question.
You want the lowest number
We scope the real work as a fixed price against an agreed plan. That number will not be the cheapest number you are quoted.
There is a whole market of people who will put a website in front of you for a few hundred dollars, and for some businesses that is a rational purchase. If what you need is a page that exists, has your hours on it, and never changes again, buy the cheap one and put the difference into something that moves money. We are worth hiring when the thing has to do a job: convert a visitor, remove a recurring hour of admin, hold up under a procurement review. Not when it only has to exist.
The disagreement is internal and unresolved
Structured buying is fine. Public agencies buy through committees and staged approvals, and that process is designed to work. We build for it deliberately.
What does not work is a project where two people inside your organization want different things and neither has said so out loud. Software forces that argument to happen, usually at the worst moment, usually in a review meeting about a button. If you cannot yet write one sentence describing what success looks like that everyone would sign, the next thing to buy is not software. Settle it first. We will still be here.
What you have is a process problem in a software costume
Software makes a good process faster. It also makes a bad process faster, and more expensive to change.
If orders get missed because nobody owns the step where they get confirmed, an automated confirmation will not fix that. It will hide it for a quarter. We would rather point at the step than sell you a system that papers over it. Sometimes the correct deliverable is a conversation and a one page document, and we are not the best value for money on those.
You want AI, and the sentence stops there
A model earns its place when it removes a specific repeated task and stays honest when the answer is uncertain. That is a narrow, checkable claim, and it is the only one we make.
If you cannot name the task, there is nothing to build yet. We will say so on the first call rather than three invoices later. The useful version of this conversation starts with what your team does every week that is dull, repetitive, and mostly reading or typing. Start there and the technology choice largely makes itself.
It has to be live on Friday
We build small on purpose, so we are faster than a lot of the alternatives. We are not faster than every deadline. If the date is fixed and immovable and close, tell us the date first. If we cannot hit it we will say so in the same conversation, because a plan that quietly assumes a miracle is worse than a no.
So when is it a fit
When you can describe a problem in your own words. When one person can decide. When you would recognize a good answer if you saw one running.
That last part is the whole of how we work. Whatever you need built, a website, a mobile app, enterprise software, an AI workflow, you see a working version of it before any money changes hands. No obligation, no money down. Click it, share it, sit with it. If it is not right, you owe us nothing and you are out only the time it took to look.
Which is also why we can afford to publish an article like this one. We are not trying to win an argument on a page. We are trying to get to the part where you are looking at the real thing and deciding for yourself.