Market and audience research with Perplexity

#01What market research questions suit Perplexity?

#02How do you build a market sizing source sheet?

1. Hunt for published estimates.

Find published market size estimates for accounts-payable automation software. Table: publisher, year published, market definition, region, value, CAGR if given, methodology note, URL. Exclude press-release aggregators that don't name the original publisher.

1. Compare definitions. Estimates differ mostly because definitions differ (software only or with services, global or regional).
2. Source bottom-up inputs.

Find official or reputable sources for: number of US companies with 200–2,000 employees; typical AP software price per month for mid-market; adoption rate of AP automation in mid-market (cite survey and sample size).

1. Do the maths yourself. Companies × adoption headroom × price × 12, in a sheet with every input linked.
2. Sense-check with Model Council (if you have it). "Here is our bottom-up model. Which inputs are weakest?"

Our view: a bottom-up model with linked inputs is far more persuasive to a CFO than a big top-down number from a report you can't read.

#03How do you scan for trends?

Use date windows and source rules:

What changed in AP automation in the last 90 days? Cover product launches, regulation, funding and analyst commentary. Only use sources published after [date]. Table: date, change, source, why it matters for mid-market finance teams.

Follow up with "Which of these are likely to matter in 12 months, and what evidence supports that?" Perplexity will give reasons. You decide which you believe.

For recurring scans, the API's search_recency_filter or search_after_date_filter makes this reproducible. Perplexity's cookbook includes a news-monitor example using recency and domain filters (lesson 11).

#04How do you collect voice-of-customer language?

Choose Social plus Web sources, then ask:

Search community threads and review text from the last 12 months where finance managers discuss accounts-payable problems. Extract exact phrases they use to describe pain, grouped into themes. Include the link for each quote. Do not paraphrase.

Then:

For each theme, what words do they use for the outcome they want? Which competitor names come up, and in what context?

Use the output for:

• Headlines and ad copy. Use buyers' words, not yours.
• FAQ content. Turn real questions into answer-first sections (see lesson 15).
• Prompt sets. Real phrasing makes better tracking prompts (lesson 19).

#05How do you combine public research with your own data?

Inside a Project:

1. Upload anonymised call notes, survey exports or support tags. Check your data policy first.
2. Ask "Compare themes in the attached call notes with themes in public reviews. Label each theme as internal only, public only, or both."
3. Themes found only in your internal data are differentiated insight and good content angles.
4. Public-only themes show where competitors' customers are unhappy.

Enterprise plans add data protections. On personal plans, don't upload anything you wouldn't send to a vendor.

#06How do you check market research before using it?

#07Worked example: Ledgerline

Ledgerline is a fictional accounts-payable software company. Details are illustrative.

Ledgerline's board asked for "the size of our market". Marketing used Perplexity to build a source sheet instead of quoting one number.

Walkthrough

1. Ran the estimate-hunt Research prompt and removed duplicates.
2. Built a definitions table to explain the spread.
3. Sourced three bottom-up inputs and calculated a range in a sheet.
4. Ran the voice-of-customer prompts with Social sources.
5. Uploaded anonymised call notes to the Project and ran the comparison prompt.
6. Presented a range with linked inputs, and the "approval chasing" content angle.

#08Common mistakes

• Asking for "the" market size. You'll get a number without a method.
• Citing aggregators. Always find the original publisher.
• Treating Reddit as a survey. It's language, not statistics.
• Uploading sensitive data to a personal plan. Check policy and plan protections first.

#09Try it yourself

Run the voice-of-customer prompt for your category. Collect 20 exact quotes, sort them into themes, and compare them with your homepage headline. Is your headline in your buyers' words? Rewrite it if not.

#10Quick quiz

1. Why do published market estimates differ so much?

Answer: Mostly because the market definition, region and method differ.

2. What should you do with a number Perplexity gives without a named source?

Answer: Don't use it. Find the original source or mark it unknown.

3. What is community-thread evidence good for?

Answer: Language and themes (qualitative), not representative statistics.

#11What should you check before the next lesson?

• A market source sheet with definitions and linked inputs exists.
• You have at least 20 voice-of-customer quotes in themes.
• You know which themes appear only in your internal data.

Previous: Competitor research with Perplexity. Next: SEO and content briefs. Or return to the Perplexity for Marketers track. Related: Search modes and Research, ChatGPT market and audience research.

#12Common questions

Can Perplexity size my market? It can find and compare published estimates and the inputs for a bottom-up model. It shouldn't invent a market size, so always trace each number to a named source.

Is Social search good for audience research? Yes for language and pain points, because it pulls community threads. It isn't representative, so treat it as qualitative evidence.

How do I combine Perplexity with my own customer data? Upload call notes or survey exports into a Project and ask Perplexity to label each finding as internal or public, then compare the two.