The Developer Who Sat in a Sales Call
A high-impact exercise in accepting things as they are
I forced my senior developer to sit in on a sales call. “Just listen. Don’t talk.”
Forty-five minutes later, he looked like he’d swallowed battery acid.
“They want us to make withdrawals take six hours on purpose?”
“Yes.”
“But we built instant processing. That’s the whole value prop.”
“They need time to check WhatsApp.”
“…WhatsApp?”
“That’s where their risk team lives.”
That was the day he learned: everything we built was wrong. Not technically wrong. Business wrong.
Why Sales Notes Won’t Save You
After that call he asked for the sales notes. Here’s what they said:
Client interested. Needs review workflow. Budget approved. Decision by EOQ.
Here’s what he actually heard in the room:
They called our API “too complicated” (it’s REST, not rocket science)
Trevor approves everything manually, because Trevor is the system
Risk checks happen in a WhatsApp group
They want slower processing, not faster
Their reconciliation engine is Google Sheets
They adore their current terrible setup
Sales didn’t note any of that. Why would they? Sales already knows every client has a Trevor and a spreadsheet. That’s not worth writing down.
For a developer, though, that’s the entire game.
The Exercise That Changes Everything
Stop asking sales what customers want. Start listening to what they actually say.
Five calls with customers/leads. One week. Say nothing. Write everything.
Note down:
Every time they mention a name (that person is a dependency)
Every tool they mention (that’s your real competition)
Every “we just…” (that’s the actual requirement)
What gets them excited vs what makes them glaze over
The exact words they use vs the jargon you use
By call #5, the lightbulb goes on and understanding starts to emerge:
They don’t have “users,” they have “traders”
They don’t want “authentication,” they want “login”
They don’t need “real-time,” they need “same day”
They don’t trust “automated,” they trust “Sarah checks it”
How Profiles Actually Form
It’s not that every customer falls into neat buckets. It’s messier. But patterns add up.
One client has a Trevor. Another has a Sarah. Another has a “manual risk process.” Same thing, different accent.
One has a spreadsheet, another has a custom PHP monstrosity, another has Notion tables. All roads lead to Excel.
One says they want AI, another says they want “insights,” another says “competitive edge.” All of them want a buzzword they can drop in board meetings.
Listen long enough and you stop hearing “random chaos” and start seeing profiles emerge. Not clean personas, but recognisable archetypes of dysfunction. Sales spots them instantly. Devs won’t — unless they’re in the room.
What Developers Learn When They Actually Listen
Real calls sound like this:
“Your risk system is too aggressive.”
Translation: it blocked fraud but also their biggest client.
Lesson: They’d rather risk fraud than lose a whale.“We need audit trails.”
Translation: a regulator asked once.
Lesson: Nobody reads them. They just need to exist.“Can you integrate with our existing systems?”
Translation: Steve built something in PHP in 2011 and everyone’s scared of it.
Lesson: You work around Steve, not replace him.“We want AI-powered insights.”
Translation: competitors say they have AI.
Lesson: They want to say AI, not use AI.
This is what reality looks like from the point of view of the people who buy. And it’s all in the call if you want to listen.
Enlightenment
By the tenth call, you’ll likely realise:
Nobody gives a shit about your clean architecture.
Manual processes aren’t bugs, they’re features.
Your competition isn’t other software, it’s Excel and WhatsApp.
Sales wasn’t lying or clueless. They were translating customer reality into something you could tolerate.
In other words, after those five calls you’ll either quit in disgust or become 10× more valuable. Both responses are valid. Choose wisely.
The Choice
Keep building what’s “right” and wonder why nobody buys it. Keep butting heads with sales and waste resources in pointless back and forth until you bleed out.
Or spend five hours listening to actual customers and finally understand what to build.
Sales lives in these calls. They already know. You just don’t believe them because it sounds technically wrong.
But technically wrong that sells beats technically right that doesn’t. Every time.
Now AI can code, faster and better than you. It doesn’t care if the feature makes sense, or if the idea is elegant. It just gets the outcome done.
It can’t sit in a sales call though, understand the customer and make informed decisions that may or may not spit in the face of “well-engineered”.
If you can’t either, you’re finished.
Important P.S.
Sales won’t invite you unless you promise to shut the fuck up. That’s fine. You’re there to listen, not to explain why the customer is wrong about their own business. Worse case, get a recording, it’s not quite as good a lesson but serviceable.