Possum is the name the local search industry gave to a set of changes Google made to its local ranking systems on September 1, 2016. The update was never announced, named or confirmed by Google. It nonetheless became the reference point for how the company's local results behave, because it changed the local filter: the mechanism deciding which business listings are eligible to appear in the three-result local pack and the longer Local Finder list behind it. Listings that vanished overnight had not been removed from Google's index. They were suppressed, and the community named the update after an animal that survives by appearing dead.

The distinction matters because local search runs on a separate system from organic web ranking: a business can hold the first organic position for a query and still be absent from the map results above it. Google's guidance describes local results as driven by three inputs: relevance, or how well a Business Profile matches the query; distance from the searcher; and prominence, or how well known the business is, based partly on links and review volume. The filter sits on top of that ranking layer, removing listings judged to duplicate others already showing.

How the local filter works

The filter predates Possum. What Possum changed was how aggressively it fired and on what evidence.

Before September 2016, according to contemporaneous practitioner accounts, Google's local deduplication keyed mainly on shared telephone numbers and website domains. A dental practice with a separate listing for each dentist, all pointing at one site and one number, would see most of them suppressed.

Possum extended the test to physical address and affiliation. Listings sharing a street address, or that Google inferred shared ownership, began competing for a single slot even when phone numbers and domains differed. The effect landed hardest on professions that cluster: law firms in a single tower, medical specialists in one building, brokers and accountants in shared office suites.

Sterling Sky, the agency run by local search analyst Joy Hawkins, documents the filter's current mechanics in terms that have held since 2016. Listings are filtered on business name and URL, and on address and proximity. They do not always disappear; often the ranking is simply suppressed. The survivor is the listing with the strongest ranking authority for that specific query, so a business can be filtered for one keyword and visible for another. Practitioners test for filtering by zooming the Local Finder map in one notch, at which point suppressed listings reappear.

What changed on September 1, 2016

Hawkins published the first detailed analysis on September 21, 2016. Five patterns emerged.

Businesses outside a city's administrative boundary gained ground. Before the update, a contractor with a suburban address struggled to rank for queries naming the adjacent city, regardless of listing strength. Possum replaced a hard boundary test with a proximity test, and those businesses rose.

The searcher's physical location gained weight, so identical queries began returning different local packs depending on where they were issued, on desktop as well as mobile.

Small keyword variations began producing materially different result sets. Three phrasings of one commercial intent, differing only in word order or a state abbreviation, could return three different sets of businesses.

The local filter appeared to operate more independently of the organic one. Sites suppressed organically for duplicate content were observed ranking in local results, an outcome hard to reconcile with Google's published quality guidance. And the address-and-affiliation filtering described above became visible at scale.

Measuring the shift

The most-cited quantification came from a study Hawkins ran with the rank-tracking vendor BrightLocal. The analysis covered ranking trackers for 1,307 businesses tracking 14,242 keywords, comparing positions on September 7 against August 31, the day before the change.

Nine percent of keywords showed a business entering the Local Finder for the first time. Eleven percent gained three or more positions and 15 percent gained one or two; 15 percent lost one or two and 14 percent lost more than three. Thirty-five percent showed no change. The residual figure, 64 percent of keywords registering some change, became the number the industry repeated. A narrower cut, movement of three or more positions or entry from outside the top 60 results, accounted for 34 percent.

Two caveats are rarely repeated with that number: the sample was one agency's client base rather than a randomised set, and it was measured seven days after the change, in a surface that registers daily volatility anyway. The wider search industry noticed little, because volatility trackers such as MozCast and Algoroo cover keyword sets in which three-pack queries are a small fraction.

The name, and Google's silence

Credit for the name is generally given to Phil Rozek, a local search consultant, on the reasoning that affected listings were playing possum rather than dying. The convention followed Pigeon, the community name for a July 24, 2014 local update, and the earlier naming of Panda and Penguin.

Google confirmed none of it. SALT.agency reported an update annotation appearing in Search Console reports during the period and later being removed, an observation that remains single-sourced. Asked on Twitter whether the changes would continue, Google's Gary Illyes replied that the company always tinkers and further changes could be expected. That is as close to acknowledgement as the update received.

Hawk, and the partial rollback

The address filter proved blunt. Businesses separated by a street, or on different floors of one building, were treated as duplicates. On August 22, 2017, practitioners observed the radius tightening: the filter continued to suppress listings sharing a building while sparing many nearby businesses. The community named the correction Hawk, because hawks eat possums. It was also never confirmed.

What followed

Two later local updates are routinely measured against Possum. In November 2019, Google applied neural matching to local results, confirming the change on December 2 as the November 2019 Local Search Update; Hawkins had already nicknamed it Bedlam. That was the first local ranking update Google acknowledged directly.

The second was larger. Between November 30 and December 8, 2021, Google rebalanced local ranking factors in a change the industry named Vicinity, confirmed on December 16. Search liaison Danny Sullivan acknowledged that failing to announce it had been "a miss in this case." Analysts concluded that Vicinity raised the weight of proximity and cut the ranking value of keywords stuffed into business names, shrinking the radius across which any single location could rank.

Why it matters

Possum established the working assumption that local visibility is a filtering problem before it is a ranking problem. That framing governs how multi-location brands, franchise networks and professional practices structure listings, and it has survived every subsequent change.

The surface itself has kept moving. The local pack is now substantially monetised: sponsored listings rose from under 3 percent to roughly 22 percent of tracked mobile keywords between November 2025 and January 2026, a 733 percent increase across 1,200 mobile ranking reports. Google Maps became a selectable Demand Gen placement in May 2026, and Local Services Ads began migrating into a pay-per-lead Performance Max campaign type from August 2026. The organic slots the filter governs are fewer and more contested than in 2016.

Disambiguation

Possum and the local filter. The filter is the mechanism; Possum names one change to it. It operated before 2016 and operates now, so attributing a current suppression to Possum specifically is imprecise.

Possum and Pigeon. Pigeon, in July 2014, tied local results more closely to organic ranking signals and reshaped how directory sites performed. Possum moved the other way, with practitioners observing the local filter behaving more independently of the organic one.

Possum and Hawk. Hawk is not a separate system, but the community name for the August 2017 tightening of the proximity radius Possum had widened.

Possum and Vicinity. Vicinity, in late 2021, was a ranking rebalance that raised the weight of distance and devalued keyword-stuffed business names. Possum was principally about eligibility and deduplication. Google confirmed Vicinity and never confirmed Possum.

Limitations and disputes

Everything published about Possum is inference from observed result sets. There is no Google documentation, no product name, no rollout window and no completion notice. The September 1 date is the day trackers moved, not a date Google supplied.

The boundary of the update is contested. General ranking tools recorded organic movement around September 1 and September 13, 2016, and Penguin 4.0 was announced later that month, which makes clean attribution difficult. Practitioners in different markets reported different effects, and some argued that several concurrent changes were being bundled under one name.

The evidence base is narrow: the dominant public account of Possum, Hawk, Bedlam and Vicinity traces to one analyst and one agency's client data.

Recent developments

The infrastructure Possum acted on has been repurposed. Google Business Profile now functions as the retrieval layer feeding Gemini, Search and Maps, raising the stakes on the same record fields the filter reads. Google product leadership has described AI Mode generating local recommendations through query fan-out across a places database, a retrieval process with no pack position at all.

Entity confusion of the kind the filter was built to prevent has resurfaced in generated answers: an AI overview merged two businesses sharing a name, forcing a bakery to publicly deny it was closing.

Discovery is fragmenting. Yelp recorded 3.4 times more citations than any rival platform across 28 million AI responses to local queries in the fourth quarter of 2025, though survey research still found 93 percent of consumers trusting conventional search over AI assistants for local businesses, and only 15 percent saying they trust AI search platforms "a lot".

Timeline

  • July 24, 2014: Pigeon, the community name for a local update tying local results more closely to organic signals, rolls out.
  • September 1, 2016: Local rank trackers register large movement in the local pack and Local Finder. Google issues no announcement.
  • September 21, 2016: Joy Hawkins publishes the first detailed analysis of the change in Search Engine Land. The community name Possum, credited to Phil Rozek, takes hold.
  • November 3, 2016: A study with BrightLocal covering 1,307 businesses and 14,242 keywords reports 64 percent of tracked keywords showing some change.
  • August 22, 2017: The proximity radius tightens, sparing many nearby businesses while continuing to filter listings in the same building. The community names the change Hawk.
  • November 2019: Google applies neural matching to local results. Hawkins nicknames it Bedlam.
  • December 2, 2019: Google confirms the change as the November 2019 Local Search Update, its first acknowledged local ranking update.
  • November 30 to December 8, 2021: A local ranking rebalance runs, later named Vicinity by Sterling Sky.
  • December 16, 2021: Google confirms the November 2021 local search update and links to its existing local ranking guidance.
  • November 17, 2024: Google publishes a consolidated guide to local search ranking factors.
  • November 2025 to January 2026: Sponsored listings in the local pack rise from under 3 percent to roughly 22 percent of tracked mobile keywords.
  • August 2026: The first phase of the Local Services Ads migration into Google Ads begins for selected United States service advertisers.

Summary

Who. The change was made by Google, which never confirmed it. It was identified, documented and named by the local search community, principally by analyst Joy Hawkins of Sterling Sky, with the name credited to consultant Phil Rozek. Those affected were local businesses, multi-location brands, professional practices sharing addresses, and the agencies managing their listings.

What. A revision to Google's local filter, the layer that suppresses listings judged to duplicate others already showing. Filtering expanded from shared phone numbers and domains to shared physical addresses and inferred affiliation. Proximity replaced hard city-boundary tests, the searcher's location gained weight, and small query variations began producing materially different result sets.

When. Rank trackers registered the change on September 1, 2016. The proximity radius was tightened on August 22, 2017 in the correction named Hawk. Neither event was announced by Google.

Where. In Google's local surfaces specifically: the three-result local pack, the Local Finder list behind it, and Google Maps. Organic web results were not directly affected.

Why. The stated purpose, inferred from behaviour rather than documentation, was result diversity: preventing one operator from occupying multiple slots for a single query and reducing the value of listing-based spam. The side effect was that legitimate businesses in dense commercial buildings were suppressed against each other, which is what Hawk partially corrected a year later.