How to Run a Continuous Research Practice in Dovetail
A continuous research practice means you run a small study every two weeks, tag into a shared taxonomy, and publish a monthly evidence digest instead of a one-off report. Dovetail stores the evidence; the cadence is what keeps it alive.
Most teams buy Dovetail, load one big study, and let the repo go stale within a quarter. This is how to avoid that.
Why do research repositories go stale?
Because the work is episodic. A study happens, a report is written, everyone moves on, and six months later the evidence is out of date but nobody removes it. Search then returns confident quotes about a product version that no longer exists.
The fix is a rolling cadence, not a bigger study. Four interviews every two weeks beats twenty interviews once a year, because the insights stay attached to the current product.
How do you keep tags consistent across researchers?
Lock the taxonomy to a small, reviewed set. Practical rules:
- One owner maintains the tag list and approves new tags.
- Cap it — over about forty tags, people stop browsing and start inventing.
- Use a fixed prefix per theme, like
onboard:orbilling:, so search groups cleanly. - Review unused tags quarterly and merge duplicates.
Naming discipline sounds fussy until you search for pricing feedback and find it split across five tags.
What does a monthly digest look like?
One page, three sections, tied to decisions:
| Section | Content |
|---|---|
| New evidence | Top three insights from this month’s sessions, each with a clip link |
| Confirmation or conflict | What supports or contradicts last month’s findings |
| Open questions | What the next study should test |
Send it to the people who make roadmap calls, not just to the design team. A digest nobody with authority reads changes nothing.
How do you connect insights to decisions?
Attach every insight to a ticket or roadmap item, and record when it was last validated. An insight older than two product releases should be re-validated before anyone quotes it.
Keep a short “stale evidence” list and revisit it monthly. It is the cheapest way to stop the repo from quietly becoming fiction.
When should you drop the cadence?
During a discovery sprint you may go weekly; in a maintenance quarter, monthly is plenty. What you should not do is pause and restart — a three-month gap means re-learning context and re-reading old transcripts, which is where the cost of the tool stops paying back.
Quick Summary
- Run four interviews every two weeks rather than one annual mega-study.
- Assign one tag owner and cap the taxonomy near forty tags.
- Publish a one-page monthly digest to decision-makers.
- Link insights to tickets and re-validate anything older than two releases.
- Adjust cadence to the quarter, but never pause for months.
