We have sat through more discovery calls than we can count, almost all of them from the buying side, and we can tell you what happens in every single one. The vendor asks questions designed to sound insightful. The buyer answers with the org-chart version of the company — how the process is supposed to work, how the manual says it works, how it worked the week it was designed. Everyone is being honest. Nobody is telling the truth.
That’s not a character flaw on either side of the table. It’s the format. Ask an owner to walk you through their quoting process and you’ll get the process, the official one. What you will not get is the spreadsheet the office manager keeps because the official process loses jobs between steps four and five. You won’t get it because she doesn’t think of it as a system. It’s just the thing she does. It is also, very often, the load-bearing wall of the entire operation.
A spec written from discovery calls is therefore a description of a fictional company. A polite, complete, internally consistent fiction. The project then gets built against the fiction, and everything the industry calls scope creep is mostly just reality leaking back in, one change order at a time. The client didn’t change their mind. The truth arrived late, and it billed hourly.
We learned this the expensive way. We spent two years inside a large ERP implementation as the client — three of us, full time, on scoping and QA and acceptance testing, in a room that never had fewer than eleven people in it. There were workshops. There were process maps. There was more discovery than any project could ever want. And still, the distance between what got specified across that boardroom table and what the business actually did on a Tuesday was wide enough that we spent months of the project walking it back. You cannot transmit a business through a conference room. We tried, with more resources than most companies will ever throw at the problem.
Then we watched it happen to ourselves from the other side. When we built the estimating system for LawnJunkie — the landscaping company one of us co-owns — the build didn’t stall on anything technical. It stalled on eleven numbers that only the business could supply, and half of them we didn’t know we didn’t know until the software asked. If we couldn’t spec that company from memory, with a partner in the room, the discovery-call version of knowing a client’s business is not knowledge. It’s a transcript of a guess.
So we stopped asking and started showing up. The first engagement we do now is a day inside a client’s systems — actually inside: the quoting tool, the job board, the books, the inbox where the orders land. We fix small things as we find them, because you learn more from fixing one real thing than from an hour of being told about it. At the end the client keeps the fixes and gets the diagnosis in writing, and there’s no invoice, which means there’s also no incentive for us to make the day sound more impressive than it was.
A day like that surfaces the things no call ever will. The workaround spreadsheet. The re-keying that everyone forgot was re-keying because it stopped feeling like a decision years ago. The one person who is, functionally, the API between two departments. These are the details that decide whether a project succeeds, and they are invisible from across a table, because nobody thinks they’re worth saying out loud.
It costs us real money to work this way — a full day of three senior people, gratis, is not a rounding error for a firm our size. Most companies put that budget into salespeople. We’d rather put it into the one activity that tells us whether the project should exist at all. Because that’s the other thing an honest first day produces, sometimes: the discovery that there’s nothing worth building yet, or that the fix is a $200 setting and not a system. We’d rather say that on day one than learn it on invoice nine.
If you’re being pitched something right now, here’s the only diagnostic we’d offer: ask what you’ll have in hand before anything is signed. If the answer is a proposal, you’re in a sales process. If the answer is work, you’ve found a different kind of firm. Ours works like this.