Compare

Hey Claude, write me a content strategy.

We love PocketBase. Propaganda ran on it until October 2026, and we have nothing but good things to say about it. So we asked Claude for a content strategy for PocketBase's blog, on Fable 5.1 at maximum effort, with web search and no limit on time or tokens. It worked for 29 minutes. Then we graded it.

31of Claude's search groups, checked against Google
2landed inside Claude's own range
28were guessed high, some by 40 times

The homework

Here is what you'd get from Claude.

Fable's plan, cut down to the parts you would act on. Quotes are word for word; the table and the two long lists are its own, shortened. The handwriting is ours, and every number we crossed out, we looked up on Google.

Claude, Fable 5.1, effort max, web search on

PocketBase content strategy, Q4 2026

53 searches, 68 pages read, 29 minutes

C+Great research.
Shame about the numbers.

"PocketBase has no blog. pocketbase.io's navigation is FAQ · GitHub · Documentation."

✓ Good catch. True.

"npm downloads of the pocketbase JS SDK went from 151K/month (Nov 2025) to 938K/month (Oct 2026)."

✓ Real data. Nice.

"Search volumes and difficulties are modeled estimates (method in §4); validate the top 20 in Ahrefs/Semrush/Keyword Planner before committing budget."

? So who checks them? You?

SearchSearches a monthDifficulty
pocketbase mcp200 to 500 370 ✓Low
pocketbase vs supabase1,000 to 2,000 580Low
pocketbase flutter500 to 1,000 90Low
best backend for flutter app1,000 to 2,000 70Med to High
does pocketbase scale200 to 400 10Low
supabase alternatives3,000 to 6,000 10,950Med open ground

✗ Ten searches a month. Ten. And the biggest open search in your market went to next quarter.

Its priority list for the quarter: pocketbase vs supabase, pocketbase hosting, pocketbase docker, does pocketbase scale, pocketbase flutter, pocketbase sveltekit, pocketbase api rules, pocketbase vs firebase, pocketbase github.

✗ All nine already know PocketBase. Who brings the new readers?

"P4. From zero to production on one VPS: a complete PocketBase deployment."

✓ Good post. I'd read it.

Its sites to watch: supabase.com/blog, appwrite.io/blog, stack.convex.dev, trailbase.io, instantdb.com, nhost.io, turso.tech, fly.io/blog, litestream.io, pockethost.io/blog, sliplane.io/blog, railway.com, hostinger.com, elest.io, dev.to, stacknotice.com, toolquestor.com, zenn.dev, qiita.com, news.hada.io, Golang Weekly, JavaScript Weekly, and about 30 more.

✗ Fifty sites. Which one do I open on Monday?

Better than nothing. Not close enough to bet six months on.

Google figures are monthly averages, worldwide, from Google Ads through DataForSEO on 20 October 2026, because Fable gave worldwide numbers. Open ground: the pages on Google's page one average two linking sites.

Close enough, right?

Yeah... no. Not even remotely.

It's better than nothing, and it's not a bad plan. But your strategy is the cornerstone. You might spend six months delivering on it and get nowhere because of a few of these mistakes: posts aimed at searches nobody makes, readers who already know you, and the one open door left for next quarter.

Same company, same day

Here is what Propaganda hands you.

Our Strategist runs on Claude too. The difference is what it reads before it writes: every search priced, page one counted, the month each search peaks, and who already ranks. This is the plan it wrote for PocketBase, in the app, exactly as you would see it.

✓ It tells you what it thinks you sell, so you can correct it first.

✓ No searches with PocketBase in them. Every reader is a new one.

✓ Every volume is Google's. Nothing to check in Ahrefs.

✓ It asks before it guesses.

✓ Eight sites, each tied to a topic.

PocketBase never signed up for Propaganda. We filled in onboarding from pocketbase.io the way its maintainer might, on a test account, and changed nothing in what came back. US English, the market our customers sell to.

Volume

The big searches were close. The small ones were fiction.

Web search can't see how many people type a query, so Fable built its numbers from Google Trends, autocomplete and traffic estimates. It got the brand name about right. On the long tail, where a new blog wins its first readers, it guessed 2 to 40 times too high, and nothing in the plan tells you which number is which.

SearchClaude's guess, a monthGoogle, a monthVerdict
pocketbase25,000 to 35,00022,200Close
self hosted firebase alternative1,000 to 2,0001,060Right
pocketbase docker800 to 1,5006301 to 2x high
pocketbase hosting800 to 1,5002503 to 6x high
sqlite in production1,000 to 2,0001706 to 12x high
pocketbase api rules300 to 500506 to 10x high
best backend for react native app1,000 to 2,0005020 to 40x high

Worldwide Google Ads averages through DataForSEO, 20 October 2026, summed over the main variants Fable named in each group. Of the 31 groups we checked, two landed inside Fable's range, 28 came in under it and one came in over. Without web search, Fable did no better: 6 of 33.

Difficulty

"Hard" is a number you can look up.

Fable judged difficulty by looking at who ranks and guessing how strong they are. We count the sites that link to the pages on Google's page one today. Under 10 is open ground for one good post. Above 80, no blog post gets in this year.

SearchClaude saidSites linking to page oneWhat that means
pocketbase vs supabaseLow1Claude was right. Open ground.
supabase alternativesMedium. Parked for next quarter2Open ground, and the biggest search in this market Claude looked at.
firebase alternativesMedium to high. "Needs backlinks first"4Open ground. One good post can rank this quarter.
sqlite in productionMedium25Winnable with a strong post. Claude was close.

Average referring domains of the page one results, US English, DataForSEO, 20 October 2026.

Where Claude stops

Claude is the writer. It was never the search engine.

Fable searched the web for 29 minutes and still told PocketBase to go check its numbers in Ahrefs. Our Strategist checks them before the plan reaches you, then shows its working next to every number.

What happens next

Then it holds you to it.

You answer its questions and approve the plan. Every week after that, the pitches come from those topics and those searches, and the Goals page shows where each search stands. Your own Search Console queries in the same table Coming.

Ask us for your strategy.

Four answers and your website. You get the plan, the searches and the reasons behind every number.