How HewnTree helps me ask the right questions as a Product Owner

Most people think a product owner’s job is writing things down. Backlogs, user stories, acceptance criteria, the documentation that keeps a delivery team fed. That’s the visible part. The actual job, the part that decides whether any of the documentation is worth anything, is asking the right questions first.

I’ve been a BA and then a PO for about eight years. The most common mistake I see, across companies and teams, is starting to build before the idea is defined enough. Someone has a vision, everyone’s keen, a backlog appears, and six sprints later you hit a question nobody asked in week one that would have saved a month. The backlog was premature. It documented an idea that hadn’t been interrogated yet.

The interrogation is the craft. And good questions are not obvious.

What a good question actually does

A weak question collects information you already expected to need. A good one surfaces the thing you didn’t know was missing. “What’s the login flow” is fine, but you’d always ask it. “What happens to a half-finished item when the user’s session dies” is the kind that exposes a hole nobody looked at, because it lives at the seam between two features, where no single person feels ownership.

Over the years I’ve built a rough sense of where the useful questions hide. At the edges, where one feature hands off to another. In the failure paths, while the happy path soaks up all the attention. In the negative space, the things you’re deliberately choosing not to build, which usually describe the product more honestly than the feature list does. And in the gap between what someone says they want and what they’d actually use, which are rarely the same thing.

None of that is mystical. It’s pattern recognition built from being burned, repeatedly, by the questions I didn’t ask.

The hard part is doing it consistently

Knowing where the questions hide doesn’t mean I ask them every time. On a good day, with energy and focus, I run the full battery. On a normal day, on my own idea, at nine in the evening, I cut corners like everyone else. The method lives in my head, which makes it only as reliable as my attention that night.

That’s the part HewnTree systematizes. You give it a rough idea and it asks the targeted questions, batch by batch, aimed at the weakest parts of what you’ve described. It doesn’t get tired at nine. It doesn’t skip the awkward question about the failure path because it’s keen to get to the fun feature. The readiness score then shows you, per area, where you’re still thin, so the gaps surface now instead of in sprint six.

For a PO, it’s the questioning discipline I try to bring, made consistent. For a BA eliciting requirements, it’s the same thing, which is why this works just as well if your title says analyst instead of owner. The two roles share one core skill: turning fuzzy intent into something precise enough to act on. The tool is built around that skill.

A quick example

Say someone brings me “an app where freelancers can send invoices.” Clear enough to nod along to. Useless to build from.

The questions that earn their keep aren’t about invoices. What happens when a client disputes one. Does the freelancer chase payment, or is that explicitly not our problem. What state is an invoice in between sent and paid, and who can see it. What happens to the data when someone closes their account mid-dispute. None of those were in the sentence. All of them have to be answered before a developer writes a line, or they get answered by accident, in code, badly.

That’s what I mean by the defining being the job. The sentence took two seconds. The product that survives contact with reality is hiding in the twenty questions the sentence didn’t answer.

Where the human still does the work

What it won’t do is decide for you. It can ask whether you’ve thought about the offline case. It can’t tell you whether the offline case matters for your users, because it doesn’t know your users, your market, or the dozen constraints living in your head that never made it onto the page. The questions are systematized. The answers, and the judgment about which answers matter, stay yours.

That’s the right division of labor, honestly. The part I most wanted off my plate was never the thinking. It was remembering to ask everything, every time, even when I’m tired.