If you ran local SEO campaigns in 2020 and stopped paying attention until now, you would recognise most of what you found. The map pack still holds three results. Categories still decide eligibility. Proximity still wins arguments. But the surface above the pack has been rebuilt, the tooling has changed hands, and about a third of the tactics that were standard practice then are now either useless or actively risky.
This is a comparison, not a news round-up. Then against now, honestly, including the bits where the honest answer is "nothing much changed and people wasted six years pretending otherwise".
What has genuinely changed
Start with the page itself. In 2020 a local query on a phone showed a few ads, then the pack, then organic results. Today the same query often shows a generated summary first, sometimes several hundred words of it, and the pack sits below. On queries where that summary fires, the first local listing can be below the fold on a phone when six years ago it was the first thing visible.
The tooling changed too, and this one caught a lot of agencies out. The standalone Google My Business dashboard was retired and management moved into Search and Maps directly for single locations, with the Business Profile Manager retained for groups. A range of features came and went along the way: the follower model, the short-lived Google Posts formats, the messaging apps, the product tabs. Anyone whose process documentation still says "log into GMB" has not looked recently.
| Area | 2020 | 2026 |
|---|---|---|
| Above the pack | Ads, then map pack | Ads, often a generated summary, then map pack |
| Where you manage the profile | Google My Business dashboard | In Search and Maps, or Business Profile Manager for groups |
| Verification | Postcard by default | Video walkthrough common, postcard less so |
| Citation building | Standard monthly deliverable | Consistency maintenance, not a growth lever |
| Review volume expectations | Dozens looked respectable | Category norms are far higher and visible to everyone |
| Reporting focus | Ranking screenshots | Profile actions and query mix |
| Spam enforcement | Sporadic, manual reports effective | More automated, faster suspensions, harder appeals |
Verification deserves a mention because it changed the economics of onboarding. Postcards were slow but passive. Video verification asks someone to walk the premises on camera showing signage, interior, equipment and proof of authority in one continuous take. For an agency onboarding remotely that is a real coordination cost, and it belongs in your process rather than being discovered on day one.
Enforcement is the other substantive shift. In 2020, reporting a competitor's fake listing worked reasonably often and suspensions were comparatively rare. Detection is now more automated, suspensions arrive faster and with less explanation, and reinstatement is a documentary exercise rather than a conversation. Borderline tactics carry a far worse risk profile than they did.
What quietly stayed the same
Here is the part that gets less coverage, because "unchanged" is not a headline.
Proximity still dominates. The distance between the searcher and the business remains one of the heaviest factors in the map pack, and it is the one you cannot edit. A perfectly maintained profile four miles out still loses to a mediocre one round the corner for a large share of queries. Every year someone announces that proximity has been de-weighted, and every year the grid tests say otherwise.
The primary category is still the field that does the most work. Get it wrong and nothing else you do compensates, because it governs which searches you are eligible for before relevance is even weighed. That was true in 2020 and it is true now.
Consistent name, address and phone data still matters, though for a different reason than people used to give. It was never really about accumulating citations as a ranking currency. It was about Google having no contradictory evidence to reconcile. That logic is unchanged; only the tactic built on top of it has aged.
Reviews still function as both a visibility input and the thing that actually persuades a human being. Photos still influence whether someone walks in. Hours accuracy still determines whether a click turns into a visit or a complaint. And local SEO is still a maintenance discipline rather than a project, which remains the most common misunderstanding among clients paying for it.
Tactics that stopped working
Some of these died outright. Others still technically function but return so little for the effort that they belong in the same bin.
- Bulk citation building - Buying two hundred directory submissions was a normal monthly line item in 2020. The returns had already thinned by then and are now negligible. Accuracy on the twenty that matter beats volume on two hundred that nobody reads.
- Keyword stuffing the business name - It worked, visibly, for years. It is now one of the fastest routes to a suspension, and competitors report it more readily because the tools to do so are easier to find.
- Review gating - Screening customers so only the happy ones get asked was widespread and is against policy. Detection has improved and the consequence is review removal, which takes your legitimate ones with it.
- Ranking screenshots as reporting - A screenshot from one location on one day was never good evidence, and with results varying by proximity and session it is now close to meaningless. Clients have also got better at checking from their own phones and finding something different.
- Chasing every new profile feature - Followers, product tabs, the various messaging implementations. Teams built processes around features that were withdrawn within eighteen months. Wait for a feature to survive a year before building around it.
- Thin city landing pages - Fifty near-identical pages with the town name swapped out. Cheap in 2020, ineffective now, and a liability if the pages contradict your actual service area.
- Exact-match domains and doorway sites - Still occasionally seen, still not worth the risk, and no longer producing the lift that justified it.
Tactics that got more important
The flip side. These existed in 2020 but sat low on most task lists, and they have moved up considerably.
| Tactic | Then | Now |
|---|---|---|
| Services with written descriptions | Optional extra | Core relevance work, still missing on most profiles |
| Review reply quality | Nice to have | Read by prospects and increasingly summarised elsewhere |
| Q&A seeding | Ignored | Public, unmoderated, and answered by strangers if you do not |
| Photo freshness | Upload once at launch | Monthly habit, exterior shots especially |
| Answering real questions in content | Blog filler | The content most likely to be cited in generated summaries |
| Profile action metrics | Secondary | The primary measure of whether work is landing |
Two of those deserve elaboration. Review text has become an input to more than the star rating. When a summary somewhere describes a business as being known for short waiting times or for turning up when it said it would, that phrasing came from review text. Asking a customer "what did you have done and how did it go" produces material of a completely different quality from "how did we do", and the difference compounds over a couple of hundred reviews.
And answering genuine customer questions in plain language, near the top of the page, has gone from filler to among the more valuable things you can publish. Pages that answer the exact question directly are the ones that get pulled into generated answers. You may lose the click and gain the mention, which is an uncomfortable trade but a better one than being absent from both.
What this means for how you spend your week
Take everything above and turn it into hours, because that is the only form in which it is useful. If you managed a local client in 2020, your week probably looked like citation submissions, a monthly ranking report, some link outreach and an occasional profile update. That allocation no longer matches where the return is.
A more sensible split now, per client, looks like this. Reviews and replies take the largest single share, because that is the compounding activity and the one clients drop first. Profile content - posts, photos, services, hours accuracy - takes the next block and is largely batchable. Location pages and question-answering content take a similar block but arrive in bursts rather than weekly. Measurement gets a small, fixed, non-negotiable slot. Citations get an hour or two a quarter for cleanup. And a real amount of time now goes into the thing nobody scheduled in 2020: explaining to clients why a metric moved when the layout above them changed.
The reporting change is the one worth forcing. Retire the ranking screenshot as your headline. Lead with profile actions trended month on month, alongside the ratio of actions to impressions. When a generated summary pushes the pack down a page, impressions can hold steady while actions soften, and only the ratio shows that clearly. A client taught to read that number will not panic at a screenshot taken from the wrong postcode.
At one or two locations this is calendar discipline. Past a dozen it needs a system, which is where a tool earns its place. Ampli5 Pulse is what we built for that layer: one dashboard across every profile, scheduled posts, a shared media library, AI-drafted review replies a person approves before publishing, performance insights for calls, directions, clicks and views alongside Search Console, white-label reports, role-based access, and web, iOS and Android apps. It is not a rank tracker and not a citation builder, which given everything above is arguably the point.
The closing observation is the one I would have wanted in 2020. The tactics churn; the fundamentals do not. Accurate data, the right category, reviews that keep arriving and get answered, photos that show the real place, and honest measurement have survived every layout change of the past six years. They will probably survive the next one. Everything else on your task list should be held loosely.