What actually makes a local site rank (and why "just write 50 articles" is wrong)

What you're doing and why

Your site is built and submitted for indexing. Now you wait — and while you wait, everyone will tell you to "pump out 50 blog articles." We checked that advice properly before spending a single day on it: we mined the whole rank-and-rent course corpus we learned from, pulled the independent research, and then measured the actual businesses beating us in our own two markets. All three disagreed with the advice. This lesson is what the evidence says to do instead, so you spend your time on the thing that moves the needle rather than the thing that feels productive.

The exact steps

1. Build the page matrix, not a blog. Your site should cover your services multiplied by the towns you serve. Ask Claude Code:

   "List the towns within 30 minutes of <my city> with a population over 1,000, and tell me which service + town pages I'm missing on <my domain>." Target roughly 40 pages: about 15–20 town pages, one page per service, plus your core pages. Blog posts are optional. They are not the engine.

2. Check what page 1 in YOUR market actually costs. Don't guess. Ask:

   "For '<my service> <my city>', show me the top 10 results. For each real local business (skip Yelp, Angi, BBB and the directories), tell me their referring domains, backlinks, and how many words are on the page that ranks." Write down the WEAKEST real local business on page 1. That is your target, not the biggest one.

3. Compare yourself to that floor. Usually you'll find your content is fine and your links are zero. That tells you where the next month of effort goes.
4. Fix links before writing more words. The next two lessons cover citations and local links.
5. Then finish the matrix — add the missing town pages, each genuinely researched, one at a time rather than all in one dump.

What it looked like for us

We measured this on 2026-07-24 with a live SERP pull across both our markets (~$0.40 of data):

The "50 articles" claim, checked three ways:

• The course we learned from: its own example build is 43 pages — 20 location pages plus services and core pages — with zero blog posts ("we don't recommend doing" them on these sites). The hundreds-of-articles pattern belongs to a different business model (ad-funded content sites), and it was told as a cautionary tale: 300 AI articles, collapsed after 8 months.
• Independent research: the 2026 Whitespark expert survey scores local organic as on-page 33% and links 24% — there is no content-volume factor in the taxonomy at all. Ahrefs, across 174,000 pages, found essentially no correlation between word count and visibility.
• Our actual competitors: the pages beating us run 531 to 774 words. In Ozark, the #1 local business ranks with 16 keywords and a 774-word homepage — above a competitor with 394 keywords and a 1,529-word page.

Where we actually stood (both sites, days after launch, no Search Console, no Google Business Profile):

So: already ranking, content in the right shape, and the entire gap is links — about 32 referring domains versus our zero. Writing 50 articles would not have closed a single point of it.

Gotchas

• Being on page 3 in week one is normal and good. The realistic window is 30–60 days for a small market (2–6 months by the wider industry consensus). Don't judge the site before then.
• Don't copy the franchise numbers. Servpro showed 612,882 backlinks in the Prescott results. That's a nationwide total across hundreds of locations, not a local bar. Always measure the weakest real local business on page 1 — that's the number you have to beat.
• More words can hurt. Past roughly 1,000 words the correlation with traffic goes negative in the data. Add pages that cover new ground, not padding on existing ones.
• Pick markets where the map pack is weak. You won't have a Google Business Profile, and the map pack takes roughly 44–55% of local clicks. Organic still captures about 25–35%, which is plenty — but only if you chose a market where the map pack isn't overwhelming. That choice happens back in Module 1, and it's the single biggest lever on whether any of this works.