A core update cut our client's traffic 21%. Thirty days later, clicks were up 136%.

The hit

Diamond Locksmith is a real locksmith in Jacksonville, Florida. Real trucks, real emergency calls at 2 a.m. In March 2026, Google rolled out a core update, and by the time the dust settled, organic traffic was down 21%.

Nobody at Google emails you when this happens. You just watch the Search Console graph bend downward and try to figure out whether it's seasonality, a tracking problem, or the algorithm deciding your site is worth less than it was last month. In this case, the timing matched the update's rollout window exactly, and the losses concentrated in one place: the blog.

That concentration was the first clue. When a core update hits a specific section of your site rather than everything evenly, Google isn't punishing your business — it's re-scoring a batch of pages it no longer believes in.

The diagnosis: the site had earned the drop

Here's the uncomfortable part of this story: the update wasn't wrong.

The site was carrying a stack of thin, scaled blog content — the kind of generic "10 tips to avoid getting locked out" posts that a hundred other locksmith sites also publish, saying nothing specific to Jacksonville, this company, or any real customer situation. That playbook worked for years. Google's 2025–2026 updates have been systematically dismantling it, and the March update caught this site squarely.

Meanwhile, the signals that actually decide local service searches were underbuilt:

• No neighborhood-level pages. Someone locked out in Mandarin or Riverside searches with their area in mind. The site had one generic service area page trying to rank for all of Jacksonville.
• No vehicle-specific pages. Car key replacement is searched by make — people type the car they own, not "automotive locksmith." The site had nothing to match that intent.
• A shallow FAQ. Eighteen questions, no structured data, so Google couldn't read them as answers.
• 45 Google reviews against competitors carrying hundreds. For a trust-sensitive trade — you're letting this person into your house — that gap shows up in the map pack every single day.

So the diagnosis wrote itself: the site was heavy on content Google had stopped rewarding and light on the signals that decide local searches. The recovery plan was just that sentence, inverted.

The plan: stop, cut, rebuild

Step one was the one most agencies won't take: we halted the old content strategy entirely. No more generic blog posts. When the thing you're doing is the thing that got you hit, doing more of it faster is not a recovery plan.

Then we rebuilt in four moves, ordered by how directly each one maps to a real search a real customer makes.

• Six ZIP-targeted location pages. One per service area, each with the neighborhoods it covers, realistic response context for that part of town, and LocalBusiness schema. Not six copies of one page with the ZIP code swapped — that's the thin-content trap wearing a different hat. Each page has to say something true and specific about serving that area, or it doesn't ship.
• Six vehicle-specific car key pages. One per major make. A page about replacing a Honda key fob — what it involves, transponder programming, what dealers charge versus a mobile locksmith — matches the query "Honda key replacement Jacksonville" the way no generic automotive page can.
• FAQ expanded from 18 to 40+ questions, with FAQPage schema. Every question sourced from things customers actually ask on calls, marked up so machines can parse each question-and-answer pair. This matters twice now: once for Google's result features, and once for the AI answer engines that increasingly answer these questions before anyone clicks anything.
• A review system. Automated, consistent review requests after completed jobs — ask every customer, at the right moment, every time. No incentives, no gating, no shortcuts. Volume came from process, not tricks.

Notice what's not in the plan: no link buying, no content blitz, no "authority hacks." Foundation first. It's less exciting to sell and much better at surviving the next update.

What moved

We measured the 30 days ending June 8, 2026 against the prior period, in Google Search Console. Three numbers tell the story:

"Locksmith near me" — the money query for this entire trade — sat at position 7 by the end of the window, on a site that a core update had marked down three months earlier.

A caveat, because we publish those: this is a local locksmith site, so the absolute traffic numbers are small — that's why we report percentages and positions here, not raw click counts. The percentages are honest, the trend is real, and small businesses should read this as a direction, not a promise of the same multiple. What matters is the shape of the recovery: cut what the update punished, build what customers search for, and the same update logic that took traffic away gives it back.

The CTR number deserves one more sentence. Clicks doubling could just mean more impressions. CTR doubling means that when people saw this business in results, they chose it at twice the old rate — that's the reviews and better-matched pages working, not just visibility. It's the difference between showing up and getting picked.

What a small business should take from this

You don't need to be our client to run this playbook. Here's the transferable version:

• A core update drop is a diagnosis, not a death sentence. Look at which pages lost. If it's concentrated in generic content, Google is telling you exactly what to fix.
• Stop before you build. If scaled thin content got you hit, the first move is turning that pipeline off. More of the same digs the hole faster.
• Build pages that match how people actually search. By neighborhood, by vehicle make, by specific problem. One specific page beats ten generic ones — but only if each page says something true and particular. Templated pages with a city name swapped in are the disease, not the cure.
• Reviews are a ranking input and a conversion input at once. The fix is boring: ask every customer, systematically, at the right moment. 45 to 165 came from process, not personality.
• Structure your answers. FAQPage schema costs almost nothing to add and makes your content legible to Google's features and to AI answer engines. In 2026, being the citable answer is the game.

The theme across all five: nothing here is clever. It's the unglamorous work of making a site genuinely useful to the people searching for it — which, not coincidentally, is exactly what core updates are trying to measure.

Questions we get about core update recovery

If your traffic dropped after a core update — or you'd rather fix the weaknesses before the next one finds them — this is the work we do. See how our SEO & AI search retainers work, or check the rest of the receipts.

Get the free 20-point audit → See all results →