The dev shop will say yes
Why “can you build it?” is the wrong question to put to people who are paid to answer yes.
Ask a firm whether they can build your product and you will hear yes.
It is the correct answer to the question you asked. It is also nearly useless, and I say that as someone who spent years being the yes.
For most of my twenties I took contracts through agencies and directly with companies that had already decided what they wanted. I was competent, I was available, and I was being paid to build the thing rather than to interrogate whether the thing should exist. So I built it. I have built a from-scratch e-commerce application for a retailer that could have been trading on an off-the-shelf platform in a fortnight. I have built internal dashboards that a licence fee would have covered. Every one of those invoices was honest and the work was good. In several cases the most valuable sentence available to the client was one nobody in the room was being paid to say.
That is the shape of the problem. Capability is abundant. Almost anyone competent can build almost anything, eventually, at some price. What is scarce is the decision about what should exist, and that decision is never the thing you are shopping for when you ask for a proposal.
The questions are expensive for the person answering
Consider what you actually need to know before you commission software.
Should this be built at all? Should it be bought, or rented, or assembled from things that already exist? What in your operation breaks if it takes twice as long? Which parts of this become impossible to change later, and are you sure you want to be committed to them? What is the smallest version that would tell you whether the idea is right?
Now notice that a firm whose revenue is a function of hours sold1 cannot lead with those questions. Not because the people are dishonest. Because the honest answers frequently reduce the size of the engagement, and no organisation is built to argue itself smaller. If the true answer is “buy this one instead, it costs a licence and two weeks of configuration,” that answer costs the person giving it a quarter of their pipeline.
None of this makes builders villains. The good ones are precise about what they are: an instrument for converting a decided plan into working software, and often a very good one. The failure is upstream, and it is the client’s. It happens the moment “what should exist” gets answered by the party who will invoice for its existence.
WHO DECIDES WHAT
| OWNER | INDEPENDENT ADVISOR |
THE FIRM THAT BUILDS |
|
|---|---|---|---|
| WHAT SHOULD EXIST | Decides | Decides | Not their call |
| BUILD, BUY, OR RENT | Decides | Decides | Not their call |
| WHAT IT MUST NEVER DO | Decides | Decides | Not their call |
| WHAT BECOMES IRREVERSIBLE | Consulted | Decides | Consulted |
| SEQUENCE & MILESTONES | Decides | Consulted | Consulted |
| HOW IT GETS BUILT | Not their call | Consulted | Decides |
| WHO IS ON THE BUILD TEAM | Not their call | Consulted | Decides |
What an honest estimate looks like
When I review proposals on a client’s behalf now, I look for one thing before I look at the rate card. I want to know whether the estimate admits the conditions under which it fails.
An estimate that has been thought about reads like a contract with reality:
range: - 6 to 10 weeks
assumptions: - payment sandbox available in week 1
- one round of design revisions
- no changes to the vendor payout model
breaks_if: - scope grows to include seller payouts
- the sandbox slips past week 2A single number with no conditions attached is not an estimate. It is an opening position. You are entitled to ask what it assumes and what would break it, and the reaction to that request will tell you more about the next twelve months than any portfolio will. A firm that has thought properly about your project will already have the list and will be relieved you asked. A firm that has not will treat the question as a lack of trust.
Ask it early, before anyone is emotionally committed to a number.
Decide first, then ask
The fix is not to find a more virtuous supplier. It is to stop outsourcing the decision.
Before you ask anyone whether they can build it, write down what it must do, what it must never do, what it must be able to become, and what you will not spend. Decide which parts are load-bearing and which are opinions. Get someone whose income does not depend on the answer to tell you where the plan is wrong, and let them be rude about it. Then take that document to the people who build, and ask a much better question: not “can you build this?”, but “here is what has been decided. What would you do differently, and what does it cost?”
You will get more useful answers, because you will have made the question answerable. You will also find out something about each firm you talk to, since a good one will push back on your document in specific ways and a weak one will simply agree with it. Enthusiastic agreement with a plan is not validation. It is a sales technique with excellent manners.
The yes you want is not a yes to a question about capability. It is a yes to something you have already decided, from someone who understood it well enough to tell you where it is weak.