Groundwork
All articles

Discovery

Continuous discovery: a practical weekly routine for product teams

Updated

Continuous discovery means learning from customers every week, in small steps, instead of in one big research phase before a project. The idea is popular, but many teams struggle to make it a habit. This article is a practical routine: what to do each week, what to write down, and how to keep what you learn connected to what you build.

Why weekly beats a big research phase

A research phase produces a report that is out of date the day the team starts building. A weekly rhythm keeps learning close to the decisions it should shape. Small, frequent contact also lowers the stakes: one odd interview does not swing the roadmap when there is another one next week.

Teresa Torres’ book Continuous Discovery Habits popularised the idea of a product trio (a product manager, a designer and an engineer) talking to customers every week. You do not need a perfect trio to start. You need a small, regular slot that you protect.

The weekly routine

  1. Talk to at least one customer. Keep a standing slot and recruit continuously, for example from a line in your product or your support queue, so you are never scheduling from scratch.
  2. Capture insights the same day. Write down what the person said or did, in their words, with where it came from and who they are. Do it while you still remember.
  3. Look for patterns, not quotes. Once a week, read the new insights together. What came up more than once? What surprised you?
  4. Update your hypotheses. Raise or lower your confidence in the ones the new evidence touches, and write new ones when a pattern appears. The hypothesis template helps here.
  5. Pick the next thing to test. Choose the hypothesis with the riskiest assumption you can test cheaply this week or next.
  6. Close the loop. Record what each finished experiment taught you as a finding, and link it to the decisions it changes.

What a good insight looks like

An insight is evidence from a real person, not an opinion or an idea. Keep it small: one observation per insight, so you can link it to exactly the decisions it supports.

Quote:    “I build the same noradrenaline infusion from scratch every shift.”
Source:   Interview, 12 Nov
Persona:  ICU nurse (P07)
Area:     Prescribing
Use a role and a participant reference, not a name.
  • Keep their words. “Takes forever” paraphrased as “efficiency concerns” loses the signal.
  • Record the source. An interview, a support ticket, a survey and an analytics trend carry different weight.
  • Leave out personal details. A role and a reference like P07 are enough to tell participants apart, and keep you on the right side of privacy rules.
  • Separate observation from interpretation. “She rebuilt the order three times” is an insight. “Nurses need templates” is a hypothesis.

Where routines break down

  • Insights go into a slide deck and die there. If nobody can find last month’s evidence, the team stops collecting it. Keep it in one searchable place.
  • Only the PM hears customers. Bring a designer or engineer to interviews when you can. Second-hand quotes persuade far less.
  • Discovery never reaches delivery. If requirements do not link back to evidence, discovery feels like a side project. See requirements traceability.
  • Everything becomes an A/B test. Most weeks, a prototype shown to five customers teaches you more than a two-month split test.

Start small

Do not wait for a full process. Book one customer conversation for this week, write down three insights afterwards, and turn the strongest pattern into one hypothesis. Next week, do it again. The habit matters more than the tooling.

Keep your evidence and decisions linked

Groundwork takes your team from customer insights to evidence-based requirements. Free to start, with your whole team.

Get started