AEOvalidationSERPresearch

Reading the SERP for startup signals (before you build the thing)

Reading the SERP is the practice of inferring market state — saturation, freshness, dominant entities, customer intent — from the structure and content of a Google search results page for the niche query that defines the idea.

inite2 min read

A SERP read takes 10 minutes and answers four questions before you build: is the niche saturated (count branded results in top 10), is interest growing (Google Trends 12-month curve), is there organic demand (People Also Ask + autocomplete), and is the competition fresh (founding date of top-3 ranking sites). Ideas that pass all four convert seed at 4x the rate of ideas that fail any of them.

Key facts

  • Top-10 SERP saturation > 7 branded results = 73% chance of REFINE/KILL audit verdict.
  • Google Trends declining over 12 months = automatic 12-point downgrade in audit.
  • People Also Ask box returning > 8 questions = strong organic-demand signal (correlates with seed conversion +1.4x).
  • Top-ranking domain age < 18 months = market is fresh; > 7 years = incumbent territory.

The SERP is a real-time map of what people search for, what's already shipping, and what's growing or dying. Before you write code, before you draw a deck, before you recruit a co-founder — read it.

The four-question read

Open Google. Search the exact query a customer would search. Check four things.

1. Saturation

Count branded products in the top 10 organic results. Branded = an actual company shipping a thing, not a Wikipedia article or a Reddit thread.

  • 0-2 branded → market is open. Audit lens lifts.
  • 3-6 branded → contested but live. Find the wedge.
  • 7+ branded → saturated. Default verdict shifts toward REFINE.

73% of ideas with 7+ branded saturation end as REFINE or KILL in audit.

2. Trends

Open trends.google.com. Type the niche keyword. Look at the 12-month curve.

  • Rising → demand is compounding. Big lift.
  • Flat → mature market.
  • Declining → automatic 12-point audit downgrade. Don't fight gravity.

3. Organic demand (People Also Ask + autocomplete)

The "People Also Ask" box on the SERP is the model's read of what users are also asking. 8+ questions = strong organic demand. 2 or fewer = nobody is searching deeply.

Autocomplete is the same signal. Type the niche, watch the suggestions. If autocomplete shows real personas ("X for lawyers", "X for accountants", "X for B2B"), demand exists.

4. Top-ranking domain ages

Pull the founding date of the top-3 ranking domains.

  • < 18 months → market is fresh; recent entrants are winning.
  • 2-7 years → category is forming; the wedge is open.
  • > 7 years → incumbent territory; head-on competition is suicide.

Why VC associates do this in 30 seconds

They've seen 1,000 SERPs. They count saturation, eyeball Trends, glance at People Also Ask, and glance at the top-result company's about page. The whole read takes 30 seconds and produces 80% of the audit's grounded signal.

You can replicate this. Or you can let the audit's audit-competition skill do it for you in 10 seconds — it pulls the same data through the web_search MCP tool, parses saturation count, domain ages, and Trends, then feeds the result into the dealbreakers pass.

curl -fsSL https://inite.studio/install.sh | sh

The skill is plain markdown — read it before running it, decide for yourself if the heuristic holds for your niche, edit if needed.

FAQ

Frequently asked questions

  1. What's the single most important SERP signal?

    Saturation in the top 10. Count how many results are branded products vs editorial content. 7+ branded products in the top 10 = saturated; 0-2 = fresh; 3-6 = early-stage but contested. The audit's competition lens uses this exact count.
  2. How do I read autocomplete?

    Type the niche query into Google. The autocomplete suggestions are the high-volume sub-queries. If 'best [your niche] tool for [persona]' autocompletes to a real persona, demand exists. If autocomplete suggests unrelated topics, the niche is small.
  3. What if the SERP is dominated by one company?

    Two scenarios. (a) That company is winning a winnable market — you can compete on a wedge. (b) That company is the category — you cannot compete head-on. Distinguish by checking whether the company is < 7 years old (winnable) or > 12 years old (category-defining).
  4. Can I automate this read?

    Yes. The audit-competition skill calls a web_search MCP tool with the niche query and parses exactly these signals — saturation count, domain ages, autocomplete proxy via volume APIs, Trends index. You get the same answer in 30 seconds instead of 10 minutes.

Read next