If you’re a founder, or you lead engineering, at a small company building something technically hard, the idea of bringing someone in to look after product probably makes you wary. That wariness is usually earned. Most technical teams have seen what happens when it goes wrong: someone arrives who doesn’t understand the technology, brings in tools nobody asked for, fills the calendar with meetings, and starts deciding what gets built without understanding what it takes to build it.
If that’s the picture, pushing back is the right instinct. Nobody building something hard wants a babysitter. But whatever the role ends up being called, it’s worth separating the work itself from the bad version of it.
The work is already happening
In a young technical company, someone is already doing product work. Someone decides which customer request gets built and which gets parked. Someone explains to an investor or a prospect why the technology matters. Someone keeps track of what’s been promised and whether it’s realistic.
Usually that someone is the CEO or the CTO, squeezed in around everything else. The result is that a lot of what reaches development hasn’t really been filtered. What comes out the other side is often good work, but it isn’t always what was needed, and engineering absorbs the difference: requirements that shift halfway through, features scoped for one customer that nobody else will use, and long debates about what something is for that should have been settled before anyone opened an editor.
The question isn’t really whether the company needs product work. It’s who does it, and how much of it your engineers end up absorbing.
In practice the job is putting the pieces together. That means making sense of what customers are asking for, keeping the story consistent between what sales says, what marketing writes and what engineering builds, and making sure everyone is working from the same picture of where the product is going. It’s about deciding what’s worth building, not how to build it.
None of this should change who runs the team. Good engineering teams organise their own work, and they should keep doing so. Technical decisions stay with the engineers too. The role is there to guide, not to manage. It brings context about customers and the market, checks that what’s being built matches what’s actually needed, and asks whether something is being built bigger than the problem requires. The engineers still decide how.
What it looked like at Subterranean
Subterranean builds software for drones and robots working in hazardous, communications-constrained environments like sewers and tunnels. When I joined them two days a week, they were seven people with a live customer and technology that worked in the field. Engineering decisions were being shaped by what that one customer needed, because there was no clearer direction to follow.
I didn’t start by changing anything. I spoke with every member of the team individually, spent time with the founders and in the environments where the technology was used, and kept asking why until I understood not just what the technology did, but the reasoning behind the decisions that had shaped it.
When structure did come in, it came into the tools the team already used. They worked in GitHub, so I built the workflow in GitHub Projects instead of introducing Jira. Customer feedback and feature requests flowed in through a GitHub and Slack integration, landing in a backlog to be looked at properly rather than in an engineer’s messages halfway through a sprint.
What came off engineering’s plate
Three things changed for the engineers, and none of them involved telling them how to build.
First, the unresolved thinking moved earlier. I separated discovery from delivery, so questions about what something was for and who needed it got answered before a ticket was ready to build. Engineers stopped carrying that uncertainty into the implementation.
Second, the work started to make sense to people outside engineering. Grouping work around the value it delivered, rather than as a list of disconnected tasks, made prioritising easier inside the team and made progress easier to explain outside the team. It also gave the founders a clear way to explain the roadmap to funders, which was part of what helped secure an R&D grant.
Third, customer requests stopped arriving as instructions. They went into one place where patterns across customers could be spotted, so the team could build for the problems several customers shared rather than for whoever asked loudest.
The most telling change came from the engineers themselves. They started pushing back on release cadence discussions until real users were involved. Product thinking stopped being something I brought into the room and became part of how the team argued for its own work. Decisions no longer depended on me being there.
“We weren’t sure whether what we’d built could work beyond our first customer. Working with Chris helped us understand who else could benefit, what it would take to reach them, and how to explain it clearly. He brought enough structure to give the team clarity on what to build and why, without getting in the way of how we worked. We still move quickly, but with much more confidence that we are building the right things.”
Aidan Gallagher, Co-Founder & CTO, Subterranean
When it starts to matter
The point where this gets urgent is often the first sales hire, especially someone doing outbound. They need a message that holds up in a cold conversation, and they need it from their first week.
If product, messaging and engineering aren’t telling the same story, every new conversation brings back a different version of what the product is. The deals that follow come back shaped like whatever each prospect asked for, and engineering feels that first, as a growing pile of one-off work with a deadline attached.
Having someone who sits between customer conversations and the build is how a team keeps saying yes to customers without saying yes to everything.
Why a couple of days a week works
It’s tempting to think this needs a full-time hire. For a young company it often doesn’t, and for the engineers it can be the better arrangement.
At Subterranean I worked two days a week, but I was inside the work rather than reporting on it from a distance. I had my own company email address and access to the same tools as everyone else. I was based in the same shared building, but on a different floor, so most of my day-to-day contact with the engineers happened over Slack and Teams, much as it would for someone working remotely. When it helped to see the technology being used, I went onsite.
Working that way was possible because the work lived where the team already worked. Requests, priorities and decisions sat in GitHub and Slack, so nobody had to wait for me to be in the room to see what had been agreed. And because I wasn’t there every day, I couldn’t hover over tickets or sit in on every conversation, so whatever I put in place had to be simple enough for the team to run themselves.
The team kept organising its own work, and the days I wasn’t working with them were spent working through what we’d agreed. It also costs a fraction of a full-time salary, which matters when every hire is being weighed against runway. And because the job is putting foundations in place rather than filling a seat, it can grow or shrink as the company does.
Part-time doesn’t have to mean a slow start either. At Subterranean the foundations were in place within three months. The work then carried on into the company’s next phase, because enough trust had been built that I became a natural fit for helping lead it. The role grew with the company without ever needing to become full-time.
For a company that’s raising, it also fits the timing. Two days a week covers the gap now, and leaves room to scale up after the round if it’s earning its place.
“Having Chris two days a week worked perfectly for us, both financially and in terms of focus. He didn’t come in and tell us what we were doing was wrong. He worked within the way the team already operated and spent time understanding the thinking behind our decisions first. When he wasn’t in, we were working through what he had already put in place. That rhythm built trust across the team and made it much easier to introduce new ways of working.”
Anthony Hutton, Co-Founder & CEO, Subterranean
A useful exercise
If you’re not sure whether any of this applies to your team, you don’t need to take my word for it. For one month, keep a rough count of:
- work that was rebuilt because the brief changed partway through
- features built for a single customer
- tickets that stalled waiting on a decision from someone outside engineering
- hours technical leadership spent on customer calls, scoping, and explaining the technology to non-engineers
If the numbers are small, you probably have this covered. If they aren’t, that’s time your engineers are spending on product work they never signed up for.