- 01
Separate the evidence types
Put practice records, public observations, and hypotheses in different sections. An aggregate appointment trend is different from an impression formed while browsing search results. Labeling the difference prevents an attractive assumption from quietly becoming a budget decision.
- 02
Record the search context
For each observation, save the query, date, approximate search location, and device context when available. Search results can vary. A single screenshot is a record of one observation, not a permanent statement about an entire market.
- 03
Identify a decision the research can change
A good brief ends with a choice: revise a service page, clarify access, adjust a campaign area, or investigate a booking problem. Research that cannot affect a decision is likely to become a report that nobody uses.
- 04
Close the learning loop
Choose a review point and the evidence needed to decide what happens next. Include unfavorable findings. A test that disproves an assumption can save the practice from expanding an ineffective approach.
A planning exercise, not a client result
Put the idea to work
Planning exercise: a team suspects its message is too broad. Compare the questions in suitable inquiries with the promise on the landing page. Rewrite one section around the strongest supported need and review subsequent inquiries. Do not call the experiment a market-wide finding.
What to watch for
- Research without a decision attached
- Comparing search screenshots taken in different contexts
- Confusing neighborhood identity with individual patient behavior
Research reference: City of Boston neighborhoods. The planning recommendations on this site are Identity’s analysis; the source does not endorse Identity.