Algorithm updates

Google local search updates: a running log

Most update coverage is written for other marketers. This one is written for the person whose phone stopped ringing. Newest first, plain language, and what we would actually do about it.

Published September 19, 2026 · Updated September 19, 2026 · 6 min read

This page is the log we keep for our own clients. When something moves, we write down what moved, what it appears to have targeted, and what we changed in response. Then we leave it here.

We are keeping it public because most update coverage is useless to a contractor. It is written for other marketers, it is padded to hit a word count, and it tells you to “focus on quality content” as though that were an instruction. What you actually need to know is narrower: did this hit businesses like mine, how do I tell, and is there anything worth doing this week.

Current status as of September 20, 2026: no new core or spam update is in progress. The last confirmed ranking update was the August 2026 spam update, which completed on August 21.

Every date below comes from Google’s own Search Status Dashboard, which is the only authoritative record of when an update started and finished. Most coverage you will read uses the date the article was published instead, which is usually days out.

UpdateStartedCompletedDays
August 2026 spam updateAug 18, 2026Aug 21, 20263
June 2026 spam updateJun 24, 2026Jun 26, 20262
May 2026 core updateMay 21, 2026Jun 2, 202612
March 2026 core updateMar 27, 2026Apr 8, 202612
March 2026 spam updateMar 24, 2026Mar 25, 20261
December 2025 core updateDec 11, 2025Dec 29, 202518

Note the March cluster. A spam update ran March 24 to 25, and the core update started two days later on March 27. If your rankings moved in that window, you cannot tell from the timing alone which one did it, and plenty of people confidently told clients it was one when it was the other.

August 2026 spam update

Ran: August 18 to August 21, 2026. Three days, global, all languages.

Google described this as a normal update, meant to improve its ability to catch new kinds of spam and enforce existing spam policies. Short rollouts usually mean a targeted system rather than a broad re-evaluation, and that is how this one behaved.

What it meant for contractors. Almost nothing, if you are operating legitimately. The businesses that felt this one were running the things the spam policies already name: scaled content with no purpose beyond ranking, keyword-stuffed profile names, and link schemes. If your rankings moved in that window and you know which of those you were doing, you have your answer.

One thing worth saying plainly, because it comes up on nearly every call we take: a business name stuffed with keywords on your Google Business Profile is a guideline violation, not a growth hack. It works until a competitor reports it, and competitors report it constantly. We have picked up clients mid-suspension who had been sold that as a strategy.

June 2026 spam update

Ran: June 24 to June 26, 2026. Two days.

Barely registered for local. We saw no meaningful movement across our client base, and the reporting at the time did not identify a local angle either. Logged here for completeness, because “was there an update then?” is a question worth being able to answer precisely.

May 2026 core update

Ran: May 21 to June 2, 2026. Twelve days.

What it was: a broad core update adjusting how Google evaluates relevance, usefulness and content quality across search results.

Broad core updates are not penalties. Google re-evaluates what deserves to rank, and pages move because the assessment changed, not because you were caught doing something. That distinction matters because it changes the fix. There is no violation to undo. There is only content that now compares less well than it did.

What it meant for contractors. Mostly an organic-side event. Service pages and blog content moved. Map pack positions were comparatively stable, which is consistent with the pack being scored on profile signals, proximity and reviews rather than on page quality alone.

March 2026 core update

Ran: March 27 to April 8, 2026. Twelve days, and the largest of three algorithm updates inside four weeks.

This is the one that mattered, and it is still the most useful update to understand if you are trying to work out why your rankings look the way they do today.

Who it hit. Home services, legal and healthcare saw the biggest ranking shifts of any verticals. That is our entire client base, so we watched this one closely.

What it targeted. Three things, in rough order of how hard they landed.

Templated location pages. Sites built on dozens or hundreds of city pages that swap in a place name and change nothing else were hit hardest. This pattern has been a liability for years, but March 2026 is when it stopped being a slow leak and became a cliff. If an agency sold you eighty city pages and your organic traffic fell off in early April, you already know what happened.

Legacy citation volume and review count without active management. Businesses coasting on a citation profile built in 2019 and a review count that stopped growing saw measurable drops. Profiles with current reviews, consistent name, address and phone details, and genuine local content largely held or gained.

Keyword matching, in favor of entity authority. Google moved further toward understanding a business as an entity, corroborated across sources, rather than matching a string on a page. Practically, this means the connection between your site, your profile, your reviews and your citations matters more than how many times the target phrase appears.

What we changed in response. We stopped building city pages for markets clients could not genuinely speak to, and we consolidated several clients’ thin market pages into fewer, better ones. In two cases that meant deleting more than thirty pages. Both recovered. That is not a universal prescription, and we would not do it without looking at the data first, but the direction of travel is clear enough that we now build this way by default.

The update also recalibrated how review signals feed the local pack. Google’s systems now parse review content for relevance rather than counting stars, identifying mentions of specific services, location references and authentic customer experiences. A review that says “replaced our roof in Katy after the hail” does work that “Great service, 5 stars” does not.

What we actually watch

Not every update needs a response. Most do not. What matters more than reading the announcement is having a measurement setup that tells you whether anything happened to you specifically.

Fixed grid, same points, same keywords, every month. A single rank check is close to meaningless in local search, because Maps results are computed from where the searcher is standing. Your position changes block by block. If you measure from one point and it moved, you have learned nothing about whether you gained or lost coverage. If you measure from a grid of points and the green area shrank, you have learned something real.

Calls, not positions. Rankings are the input. Booked jobs are the output. We have seen a client lose two average positions and gain calls, because the positions they lost were in zip codes they did not want and the ones they gained were in the ones they did.

Profile change history. A surprising share of “the algorithm hit me” turns out to be a category edit, a service area change, or a Google-applied update to the profile that nobody noticed. Check that before you blame an update.

How to tell an update from a local problem

Run through these in order. It takes ten minutes and it resolves most cases.

  1. Did the drop line up with a confirmed rollout window? If the dates do not match a confirmed update, it is probably not the update.
  2. Did organic and map pack both move, or just one? Both usually means a core update. Map pack only usually means a profile issue, a new competitor inside your radius, or a local pack change.
  3. Did anything change on the profile? Categories, service area, hours, name, primary phone. Check the edit history.
  4. Did a competitor change? A new location opening inside your radius will move you without anything about you changing at all.
  5. Did the grid shrink evenly, or on one side? Even shrinkage points at a ranking signal. One-sided shrinkage usually points at a competitor or a service area edit.

If you get through that and still do not know, that is genuinely what a grid scan is for. It shows you the shape of the problem, which is usually enough to name it.

The part most update coverage leaves out

There is a version of this business that involves chasing every announcement, rewriting everything each quarter, and telling clients that the latest update explains whatever just happened. It is an easy story to sell and it is mostly noise.

The businesses that hold their rankings through core updates are not the ones reacting fastest. They are the ones whose profile is complete and accurate, whose reviews keep arriving, whose content is specific enough that a human would choose to read it, and whose name, address and phone details say the same thing in every place they appear. None of that is exciting. All of it survives updates.

We will keep this log current. If an update lands that changes what we do, it will show up here with what we changed.