Running & scheduling research

Running & scheduling research

This page walks through starting a research run, what to expect while it runs, how to read the results, and how to set up recurring research so your market view stays current.

Step 1 — Pick a phase

Open your idea’s Research section and choose a phase. Each phase shows a short description of what it investigates, so you can pick the one that answers your current question. See the phases guide for help choosing.

Some phases ask for inputs before they run:

• SEO / Hacker News / Product Hunt — keywords or a topic description.
• Market sizing — a short description of your idea and market.
• Competitor analysis — your competitors’ names if you already know them.

The run starts from your idea’s context by default, so even phases with no inputs produce relevant results.

Step 2 — Start the run

Click Run and the phase is queued. You can start several phases at once — they run independently — and keep using the app while they work.

Runs move through three states:

• Pending — queued and waiting for a worker.
• Running — actively working.
• Completed — results ready to view.

If something goes wrong, the run is marked Failed with a clear reason shown in the interface, so you know exactly what to fix and re-run.

How long does a run take?

Most phases complete in a few minutes. The exact time depends on the phase and the data sources involved:

• Fast phases (trends, Hacker News, Product Hunt) usually finish in a minute or two.
• Heavier phases (market sizing, competitor analysis, red-team) involve more analysis and can take several minutes.

There’s no need to watch the page — come back when it’s done, or run something else in the meantime.

Step 3 — Read the results

Completed runs render as structured results, not raw text: metrics, lists, charts, and summaries depending on the phase. Key things to look for:

• Headline numbers — e.g. search volumes, market size estimates, growth direction, competitor counts.
• Evidence — the data behind each conclusion, so you can trace why the phase says what it says.
• Data source marker — whether the run used live data or produced simulated results. Simulated results are always labeled.

If a phase depends on another (for example, keyword gaps build on a completed competitor analysis), the app will tell you which phase to run first.

Results flow into your PMF score, canvases, and the comparison dashboard automatically — no extra work needed.

Step 4 — Set up recurring runs

Markets change, and a single research snapshot goes stale. For topics you want to keep monitoring, set up a schedule:

1. In the research section, open Schedules for the phase you care about.
2. Choose how often it should run — the app supports standard schedule expressions like “every day at 9am” or “every Monday morning.”
3. Save it. The schedule shows you when the next run is due.

Scheduled runs re-run the phase automatically, and each run is stored alongside your manual runs, so you build up a history you can compare over time. You can enable or disable any schedule at any point without deleting it, and edit the frequency whenever your needs change.

Note

Recurring runs are a monitoring tool, not a set-and-forget habit. Use them for the few signals that genuinely drive your decision — like a keyword trend or a competitor watch — rather than scheduling every phase.

Troubleshooting

• A run stays pending for a long time — try starting it again; the queue sometimes needs a nudge.
• A run fails with a configuration message — this usually means a data source isn’t available for your workspace. Fix the configuration or use the clearly-labeled simulated results to keep exploring.
• Results look thin — some phases produce richer output when given better inputs (more specific keywords, a fuller idea description). Re-run with more context.

For custom multi-step investigations that combine several tools into one research session, see AI research agents.