Run Discovery as a Weekly Team Sport, Not a Phase

Discovery keeps dying the same way. A leader kicks off a big research phase, hands it to one sharp PM or researcher, and waits for a report. The report arrives, the team nods, and then everyone goes back to shipping the roadmap they already had. Six months later nobody remembers the last time they talked to a customer.

The fix is not more research. It is a different shape. Discovery works when it runs every week, when it belongs to the whole team, and when learning is fast enough that it changes what you build. Here is how to set that up without adding a heavy new process.

Talk to customers every week, not every quarter

The core habit is small and boring: a standing customer conversation every week. Teresa Torres calls this a weekly touchpoint, and she is blunt that it beats big-bang research bursts. A quarterly deep dive gives you a snapshot that goes stale. A weekly rhythm gives you a live feed.

The point of weekly is not volume. It is that you never drift too far from what people actually need before you catch it. When a talk happens every week, one bad assumption costs you days, not a quarter.

Start with one interview a week and protect it on the calendar. Torres makes the case that anyone can start this, even a team that swears it has no time. The teams that skip it usually do not lack time. They lack the habit.

Make it a team sport, not a hero's job

When discovery falls to one person, it stalls the moment that person gets busy. Jeff Gothelf argues discovery is a team sport, owned by PM, design, and engineering together, not parked with a lone researcher. Shared ownership is what keeps it running when any one person is swamped.

There is a learning cost to a single owner too. If only the PM hears the customer, only the PM understands the problem. The engineer builds from a secondhand summary. Sophia Höfling points out that discovery is messy and only pays off when a multi-skilled team works through it together, in the same room, around one objective.

So put an engineer or designer on the interviews. Not to run them, just to hear the raw thing. Cheaper than a debrief, and the understanding sticks.

Build the plumbing so learning is fast

A weekly habit dies fast if every interview means hours of scheduling and note-wrangling. Torres treats infrastructure as its own best practice for a reason. You need a steady way to recruit people to talk to, a place to keep what you learn, and a fast path from insight to test.

Set up automatic recruiting so the next interview is always booked. Keep notes somewhere the whole team can search, not in one person's docs. The goal is that talking to a customer this week takes almost no setup.

Katie Marcus, recapping a full quarter of this, found the real gains came from shipping experiments faster and running discovery and delivery as dual tracks. The plumbing is what let the team move at that speed.

Wire it into rituals you already run

You do not need a new meeting cluttering the week. The Product Owner Summit frames cross-functional discovery as a set of patterns you fold into work the team already does, cutting the handoffs where one role passes findings to the next.

Bring what you learned this week into your existing standup or planning. Look at where a handoff is slowing you down, a spot where design waits on a research summary, or eng waits on a decision, and collapse it by having the right people in the room live.

Marcus went further and put weekly user talks right into the team's OKRs. When the habit is a tracked goal, it survives a busy sprint instead of being the first thing cut.

The deep cut

The part leaders miss: shared ownership is not about being fair or inclusive. It is the mechanism that makes fast learning possible. When the whole team hears the customer, you skip the slow relay of writing a report, presenting it, and hoping people believe it. The insight lands in the head of the person who will act on it, in real time. That is why weekly cadence, team ownership, and fast infrastructure are one system, not three tips. Pull any one out and the other two stop working.

Three questions for your team

• Are weekly customer talks actually on our calendar, and are PM, design, and eng all in the room, or is one person carrying it alone?
• Where is a handoff slowing our discovery down, and could we cut it by having the right people hear the customer live instead of reading a summary?
• What is our shared discovery objective this quarter, and is the weekly talk habit tracked as a goal so it survives a busy sprint?