Before you commit 6 months of writing, you need a defensible reason to believe people are searching for what you plan to write about. Here are the cheap, reliable signals — and a small script that turns autocomplete and SERP into a CSV scorecard.
Background
Volume tools lie. They show round numbers for highly competitive head terms and zero for everything else, even though plenty of low-volume queries collectively drive significant traffic. For indies, the signals that matter are qualitative: do real humans phrase questions this way, does the SERP have results aimed at people like you, and does the search land on content that satisfies the question. A demand scorecard captures all of that.
How to tell
- Google autocomplete suggests at least 5 related continuations when you type the head term.
- The “People also ask” box appears for related queries and refreshes when you click into it.
- There are Reddit, Quora, or Stack Exchange threads with multiple recent answers on related questions.
- YouTube has tutorial videos with at least 4-5 figure views on the topic.
- A free Keyword Planner range shows even a small monthly range (10-100) for several related terms — non-zero matters more than the exact number.
Quick verdict
If you see autocomplete + PAA + active community threads on the same topic, demand is real. If two of three are missing, the topic is too small or too dead.
Before you start
- Have a list of 30 candidate query phrasings ready — don’t validate in a vacuum.
- Open Google in an incognito window so personalization does not bias suggestions.
- Use a fresh Google account to access free Keyword Planner.
Step by step
- Pull Google autocomplete programmatically for every candidate. The unofficial endpoint returns JSON:
for q in "astro deploy firebase" "firebase hosting cache" "vercel custom domain"; do
echo "=== $q ==="
curl -s "https://suggestqueries.google.com/complete/search?client=firefox&q=$(printf '%s' "$q" | jq -sRr @uri)" \
| jq -r '.[1][]'
doneA topic with fewer than 5 suggestions is suspicious; fewer than 2 is dead.
- Build a SERP signal scorecard for 20 of your candidates. CSV columns:
query,autocomplete_count,paa_present,reddit_recent,youtube_4_digit_views,ai_overview,big_pubs_in_top10,verdict
astro deploy firebase,7,yes,yes,yes,no,2,green
firebase hosting cache,5,yes,yes,no,no,1,green
vercel build failed,9,yes,yes,yes,yes,4,yellow
nextjs middleware example,4,no,no,no,no,8,redGreen = pursue, yellow = pursue carefully, red = drop.
- Search each query in incognito, screenshot the SERP layout. What you are looking for:
PAA box → demand signal
AI Overview → click-through risk
Reddit in top 10 → indie-friendly intent
YouTube in top 5 → consider repurposing as video too
Wikipedia in top 3 → likely informational only; harder to monetize- Search Reddit and Quora for the same terms: 30 minutes. Use the site filter:
https://www.google.com/search?q=site%3Areddit.com+%22astro+deploy+firebase%22A topic with no Reddit/Quora discussion in the last 12 months is probably too niche to scale.
-
Use the free Google Keyword Planner for relative competition. Sign in → Tools → Keyword Planner → Get search volume and forecasts. Paste 20 candidates. Focus on the “Competition” and “Low/High range” columns more than the volume estimate.
-
Score and tag each query. Aim for at least 15 in the “green” bucket before committing to the niche. The simple scoring rule:
score = (autocomplete >= 5 ? 1 : 0)
+ (paa_present ? 1 : 0)
+ (reddit_recent ? 1 : 0)
+ (youtube_views ? 1 : 0)
+ (ai_overview ? -1 : 0)
+ (big_pubs_top10 < 4 ? 1 : 0)
# score >= 3 → green
# score 1-2 → yellow
# score <= 0 → red- Ship 2 sample articles. Submit to Search Console URL Inspection → Request indexing. Wait 4 weeks. Real Search Console impressions are the ground truth. Pull them:
# Search Console API quick check (after OAuth):
curl -X POST "https://www.googleapis.com/webmasters/v3/sites/$SITE_URL/searchAnalytics/query" \
-H "Authorization: Bearer $TOKEN" \
-H "Content-Type: application/json" \
--data '{
"startDate": "2026-04-22",
"endDate": "2026-05-22",
"dimensions": ["query"],
"rowLimit": 100
}'If your 2 sample articles get >50 impressions in 4 weeks, the topic has real demand.
Implementation checklist
- 30-query candidate list captured and scored in a CSV.
- Autocomplete scraped for each (not just the head term).
- Reddit / Quora freshness checked manually.
- Free Keyword Planner relative competition logged.
- 2 sample articles published and submitted for indexing.
After-launch verification
- After 4 weeks, Search Console shows non-zero impressions for the sample articles.
- Average position is < 50 for at least one target query — there is room to climb.
- Some queries discovered in Search Console that you did not anticipate — proof the topic has long-tail width.
Common pitfalls
- Trusting one tool — Ahrefs, Semrush, and Keyword Planner all disagree, sometimes wildly.
- Confusing trending topics on Twitter with search demand; viral does not equal searched.
- Assuming silence in autocomplete means low volume; sometimes Google just suppresses suggestions.
- Ignoring AI Overview presence — if the AI snippet fully answers the question, click-through will be tiny. Detection: SERP screenshot shows a generated answer above organic results.
- Skipping qualitative signals because you found a number in a tool.
- Validating in a non-incognito Chrome session — personalization warps results.
FAQ
- How many monthly searches is “enough”?: For an indie content site, even 30-100 monthly searches per article is enough if you have hundreds of articles. Volume per article matters less than coverage breadth.
- Should I pay for Ahrefs or Semrush?: Not at the validation stage. Free Keyword Planner plus autocomplete is good enough. Pay for a tool only once you have traffic and need to scale.
- How do I judge competition?: Search 10 of your target queries and count how many of the top 10 are big publishers, Reddit, YouTube, or AI Overviews. If big publishers + YouTube dominate, you cannot win that query.
- Does AI Overview kill SEO for my topic?: It hurts top-of-funnel “what is X” queries the most. Specific, action-oriented questions (“how to fix Y in version 2026”) still get clicks.
- What if my candidates are all yellow, not green?: Pick the 5 yellows with the best Reddit signal and the worst big-publisher coverage. Ship samples. Yellow with proven impressions is fine.
Related
- How to pick a niche that has search demand
- Planning a long-tail keyword site from day one
- Submit a website to Google
- Content volume vs quality
- URL inspection tool: how to use
Tags: #Indie dev #Website planning #SEO #Long tail #Search Console