Case Study 25.1 — The Local Filter and Proximity: What "Possum" and "Hawk" Revealed

A study of how Google's local ranking system actually behaves — the proximity weighting and the "local filter" that groups similar nearby businesses — using two widely-documented, community-named local updates as the window. Evidence note: Google did not officially name these updates; "Possum" (2016) and "Hawk" (2017) are names the SEO community gave to observed, reproducible changes. The behaviors they describe — proximity as a heavy factor, and filtering of similar businesses at the same address or proximity — are consistent with Google's own documentation and with the everyday experience of local SEOs. Where this case reconstructs mechanics, it is labeled as such, and no specific ranking numbers are invented.

Background

By 2016, local SEO had matured into a discipline, but its practitioners kept hitting a wall they couldn't explain from classic-SEO principles: two near-identical businesses would behave in ways that made no sense if you assumed local ranking worked like organic ranking. A business would "disappear" from the pack for a query it used to appear for, then reappear when the searcher moved a few blocks. Two law firms sharing an office building would find that only one of them could show in the pack at a time. A business just outside a city's official boundary would rank fine for the city name one week and vanish the next.

In September 2016, a broad set of these behaviors shifted at once, and the community named the change "Possum" (the joke being that the businesses weren't gone — they were "playing possum"). A little over a year later, in August 2017, a follow-up adjustment the community called "Hawk" refined one specific piece of it. Together they gave local SEOs their clearest-ever look at two mechanics that make local ranking a genuinely different animal from organic ranking.

The SEO issue: two mechanics that organic SEO doesn't have

Mechanic one — proximity is weighted heavily, and it's dynamic. The "Possum"-era observations made vivid what Google's documentation states plainly: distance is a primary factor. Practitioners watched the same query return materially different packs as the searcher's location changed, and watched businesses just outside a city's borders gain or lose visibility for that city's queries based on how Google was drawing the lines that week. The lesson wasn't subtle: in local, where the searcher is standing is part of the query. This is the moving-target property from §25.1 — there is no single ranking, only a ranking-for-a-location.

Mechanic two — the local filter. Google's local system tries not to fill the pack with what looks like the same business several times over. So it applies a filter that, when two profiles look too similar — same category, same or very close address — will often show only one of them in a given pack, suppressing the "duplicate." "Possum" appeared to widen this filter (catching more businesses, including some sharing a building or a category in a tight area); "Hawk" appeared to tighten it back, so that businesses merely near each other were filtered less aggressively than those literally at the same address. For a genuine multi-location or multi-practitioner business, this is the difference between all your locations being visible and only one showing up — which is exactly why §25.7 insists on distinct, real locations with distinct addresses, phone numbers, and profiles.

What it shows

  • Local ranking is its own system. You cannot reason about the pack using only organic-SEO intuitions. Proximity and the local filter have no organic equivalent.
  • "We ranked, then we didn't" is often geometry, not a penalty. A business that "drops" from the pack may simply be filtered against a similar competitor, or be seeing a different pack because the searcher moved. Diagnosing local volatility starts with "for whom, standing where?" — not "what did we do wrong?"
  • Distinctness is protection. The filter rewards businesses that are genuinely, verifiably distinct. Shared addresses, cloned profiles, and near-duplicate data invite suppression; real separate locations with clean, distinct NAP do not.

Outcome

Neither change was ever officially confirmed or named by Google, and both were absorbed into the normal evolution of local search — which is itself the point. The named-update framing is a community convenience; the durable, documented reality underneath is the two mechanics above, and they remain how local ranking behaves years later. Local SEOs stopped treating pack volatility as mysterious and started treating it as a predictable consequence of proximity plus filtering.

The lesson

Local ranking is a contest scored on relevance, distance, and prominence, and distance plus the local filter are what make it feel alien to anyone arriving from classic SEO. You don't fight proximity; you cover it with real locations (§25.7). You don't beat the filter with cloned listings; you avoid it by being genuinely distinct. And you treat "we dropped from the pack" as a diagnostic question about which searcher and where, not an automatic sign of a penalty. The honest practitioner reads local volatility the way a sailor reads weather: expected, patterned, and not personal.

Discussion questions

  1. Explain, to a business owner convinced they've been "penalized," why their pack appearance changing through the day is probably proximity and filtering rather than a penalty. What would you measure to confirm?
  2. Two of Rivertown's branches are relatively close to each other near the metro's center. What does the local filter imply about how carefully their profiles' addresses, categories, and data must be kept distinct?
  3. "Possum" and "Hawk" are community names for unconfirmed updates. Using the book's evidence tiers, how should you talk about them to a client — as fact, as folklore, or as something in between?
  4. If proximity is heavily weighted and you can't change your address, what are the only honest levers left to expand the set of searchers who can see you in the pack?