Talking to 10–15 people is easy. Turning those conversations into decisions is the hard part. This guide shows a simple synthesis loop: capture a few takeaways, add 1–3 opportunities, group everything into themes (manually or with AI), and filter by attributes like role and stage to avoid recency bias. You'll end with 2–5 themes you can actually build from.
The real problem isn't interviews — it's what happens after
Most founders can do discovery the founder way: events, coffee chats, quick calls, a burst of 10–15 conversations.
Then the learning disappears.
- your notes are scattered
- the last conversation "wins"
- every idea feels equally important
- you can't tell what repeated vs what was a one-off
- you don't know who the insight applies to (user vs buyer, seed vs Series A, etc.)
You don't need a research department. You need a lightweight synthesis loop that turns conversations into patterns you can act on.
What you're aiming for after 10–15 conversations
Your output shouldn't be "a bunch of notes."
It should be:
- 2–5 themes that are actually repeating
- clarity on for who each theme is true
- 1–3 opportunities worth testing next
- a short list of follow-ups (who to talk to next / what to validate)
This is how discovery becomes direction.
The 15-minute synthesis loop (no sticky notes)
You can do this in one sitting. It's fast on purpose.
Step 1) Add Takeaways (5 minutes)
Skim your recent conversations and write short, factual bullets.
Good takeaways sound like:
- "Teams patch the workflow with spreadsheets + Slack."
- "Managers care about risk/visibility; users care about friction."
- "The workaround breaks when the team grows past ~10 people."
Bad takeaways sound like:
- "They liked the idea."
- "This seems interesting."
- "They would use it." (unless they said it unprompted + you have evidence)
Rule: takeaways should describe what is true today, not what you hope is true.
Step 2) Add Opportunities (3–5 minutes)
Opportunities are just "what this suggests we could test/build" — not a roadmap.
Good opportunities are phrased like:
- "A lightweight way to standardize X workflow in under a week."
- "A template/checklist that prevents Y failure mode."
- "A faster way to identify the right ICP based on signals."
Rule: keep opportunities small enough to test quickly.
Step 3) Group Takeaways + Opportunities into Themes (5 minutes)
Now turn your pile of items into 2–6 theme buckets.
You can do this two ways:
Option A: Manual (fast and simple)
Create a few theme labels and drag/tag items into them.
Examples:
- "ICP narrowing"
- "Workflow breaks at handoffs"
- "Tool overload / too many systems"
- "Buyer vs user mismatch"
- "Need proof to justify change"
Option B: AI grouping (even faster)
If you don't want to do it manually:
- run AI grouping
- review the proposed group names
- rename anything that feels off
- confirm
The goal isn't perfect categorization. The goal is momentum and pattern visibility.
The part most founders miss: "What's true for who?"
A theme isn't useful until you know where it applies.
This is why attributes matter.
When you look at a theme, filter by just 1–2 attributes:
- Person (role): user vs manager vs buyer
- Company: stage or size
Now you can answer questions like:
- "Is this pain mostly a user pain or buyer pain?"
- "Does this show up at 1–10 people or only after 50+?"
- "Is this specific to one context or broadly true?"
This turns "interesting" into "actionable."
A simple rule for separating signal from noise
You don't need stats. Just use a rule that keeps you honest:
A theme is "real enough to act on" when:
- it shows up 3+ times, and
- it shows up across at least 2 different contexts (different roles, stages, or industries)
If something shows up once, don't throw it away — just label it as:
- "one-off"
- "edge case"
- "future idea"
That prevents recency bias from hijacking your roadmap.
Turn themes into next steps (without over-planning)
Once themes are clear, write one sentence for each theme:
Theme: ___ True for: ___ (role/stage/size) Evidence: ___ (short phrase) Next test: ___ (prototype / landing page / follow-up question)
You now have a validation plan that fits real founder life.
What this looks like in practice (quick example)
Let's say your themes end up as:
- Theme 1: "Coverage is messy" True for: early founders doing ad-hoc discovery Next test: "event-mode capture + quick theme grouping"
- Theme 2: "Buyer vs user mismatch" True for: B2B workflows with manager approval Next test: rewrite pitch for outcomes + run 5 manager calls
- Theme 3: "Tools are scattered" True for: teams using docs + sheets + Slack Next test: prototype a single workflow view
You didn't overcomplicate anything. You just turned conversations into a buildable direction.
Intervool fit
You can do synthesis in Docs and it'll work… until it doesn't.
Intervool helps when you want this loop to stay lightweight but consistent:
- capture Takeaways and Opportunities quickly
- group into Themes (manually or with AI)
- filter themes by attributes (role/stage/size) to see what's true for who
- carry those themes into Ideas so your next build is tied to evidence
If you're doing discovery right now, Intervool offers a 30-day free trial.
