Most of this Tech Team Talk series has looked at what happens once the work is already in motion: how teams read each other, carry uncertainty, delegate, and keep alignment when plans shift.
This instalment starts earlier: who is the work actually for?
Introducing Skipper
Skipper Chong Warson works with teams in the stretch before a big commitment, when there's plenty of conviction and not much evidence. His focus is on making teams state what they believe about their customer clearly enough to test it.
For engineers, this work usually happens out of view. It reaches them later as a priority call or a feature that's hard to scope. Skipper's piece shows where that starts.
Over to Skipper!
What a customer hypothesis looks like in practice
I've spent over 15 years in product and design, and I now run How This Works co, helping early-stage teams figure out which customer and problem deserves their next bet before a big build or go-to-market decision burns through their runway.
One of my main methods is a one-sentence hypothesis about a team's Bullseye Customer (their best customer). The parts are who that person is, what's pushing them to act, and the outcome they need — written down so a team can work out whether they all agree.
The Bullseye Customer
A devtools team started in a Bullseye Customer Sprint with a specific hypothesis: operations managers were dealing with fragmented tools and repetitive manual workflows. The starting hypothesis was that AI-assisted app building could help create simple create, read, update, and delete (CRUD) apps for work they were still handling manually.
But “operations manager” by itself was still far too broad to recruit against and have meaningful conversations. So, we wrote several candidate one-liners, chose one direction, built it into a fuller Bullseye Customer hypothesis, and recruited five people against it.
Those first five interviews confirmed much of the underlying friction:
Manual workflows,
fragmented tools,
and repetitive work.
They also exposed a failure in the hypothesis: the pain and discomfort wasn’t enough to produce pressure to change.
An ops manager from Toronto put it plainly in their interview: “I’d rather deal with the inefficiency I know than risk configuring something wrong.”
That sentence along with other context from the other four changed the team’s thinking. The problem wasn’t simply that the current workflow was inefficient, yes it was, but doing something different introduced a different kind of risk. Misconfiguration felt riskier than continuing with a known inefficiency.
A broad customer definition doesn't just miss that. It buries it altogether.
The one-liner makes the hypothesis explicit
A Bullseye Customer one-liner forces five things into the open:

This articulates the specific role, the specific company type or size, the trigger or motivation, the current inadequate solution, and the outcome they need. The purpose here is to give the team a specific first one-liner they can ultimately recruit against.
Different functions often reach for different customer models or types without realizing it. An engineer sees the technical constraints of an integration while a product person thinks about the workflow and then a founder zeroes in on the buyer: three people, three different customers. Each one of these answers on its own makes sense. The trouble starts when those people make roadmap decisions from different starting points.
Writing several first draft one-liners shows the team exactly where they disagree. Who owns the need/pain? Where do they hang out? What triggers action? What are they doing today? What outcome are they trying to reach?
The one-liner should point at a person you can recruit
Generate a number of candidate one-liners; I recommend 20–30. Maybe it's based on existing customer data, maybe it's a hunch, maybe it’s stealing from your competitors. You’ll know you’re working at the right level if you can turn the one-liner into a recruiting screener pretty easily.
"Operations leaders who need better automation" fails that test, for instance. It gives you no reliable inclusion or exclusion criteria.
A sharper version gives you screening rules, recruiting criteria, and a consistent basis for comparing interview evidence. If you can't derive exclusion criteria from the one-liner, it's still too broad. Go one level deeper.
That first cycle changed the recruiting criteria for the next one with my devtools team. Then, the next cycle changed them again. Across five later Listening Cycles and nearly 50 interviews, the customer hypothesis narrowed toward people who had already shipped something real with financial stakes and got stuck when it broke.
Discovery is not the same thing as a demo
Technical and sales teams typically lead conversations with the show ’n’ tell, with the thing they have already built: look at this workflow, isn’t it neat, explain the model, walk through the integration, and then hand out credits/free trial/low cost deal. Hopefully, they don’t churn out.
And starting with a demo changes what you're measuring. You’re collecting reactions to your implementation. Stronger evidence comes from their current behavior, constraints, switching risk, and asking, What is the job to be done here?
So, start with the existing system:
What happened the last time this problem occurred?
What did they do?
What tools were involved?
Where did it get more difficult?
Why did they keep doing it that way?
Treat compliments and hypothetical intent as low-signal data. “I love this” and “I’d definitely use it” and “I’ll probably sign up later” tell you less than a past story about what someone did the last time the problem occurred.
A Bullseye is a working hypothesis
The team needs one hypothesis specific enough to recruit against, test with evidence, and revise. And you can have more than one, keep the other plausible customers in the queue. The Bullseye determines which hypothesis gets the next test. For the devtools team, the interesting signal was the reason a known-bad workflow survives. That reason may matter more than the missing feature.
So: pick the one-liner, build it into a hypothesis, decide what evidence would weaken it, and then go look for that evidence.
Thank you
A big thank you to Skipper for taking the time to share his insights with TIE. If you’re interested in learning more, check out his LinkedIn and website.
Skipper has shared a one-liner activity PDF so you can try it out for yourself.
Enjoyed This?
📩 Subscribe
❤️ Hit the heart
💬 Comment
🔁 Share
🌱 And help this newsletter grow!




Well - this already sound more like a Product Manager job then a classic Engineer job. I mean, I don't disagree with this idea, after all this is what "Product Engineer" is about. Nevertheless, I still have doubts in my mind that how many Engineers could actually pivot to this mindset.
A few things that occur to me as I re-read this: the act of discovery should measure behavior, not just reactions to your product/experience. Starting with a demo can tell you whether someone likes what you built but starting with their last real experience tells you what they actually did, what got in the way, and how important the need or problem really was.