Google's local algorithm runs on a three-pillar model: proximity, relevance, and prominence. The model is stated publicly, has stayed remarkably stable across the last decade of updates, and explains every observable behavior of the map pack — including the cases that look counterintuitive (a worse-reviewed business outranking a better-reviewed one, a business outside the city center beating one inside it). Knowing which pillar a tactic moves is the difference between investing time correctly and rotating through the same checklists for a year.

“Every local-pack tactic moves one of three signals. If you can't say whether the change you're about to make moves proximity, relevance, or prominence, you're optimizing on faith. The pillars are the framework — every chapter in this cluster is a deeper look at the signals each one rewards.”

Pillar 1 — Proximity

Proximity is the distance between the searcher's location and the business's listed address (or the centroid of its service area). It's the pillar you can't directly change — your storefront is where it is — but it's also the pillar most teams misread.

Three things to know about proximity:

• Searcher location, not business location. Google ranks against where the searcher is right now. A business in midtown can rank well for downtown searchers and badly for searchers in the suburbs, all from the same profile.
• The query modifies the proximity weighting. "Coffee near me" weights proximity heavily. "Best coffee in [city]" weights it less. "Specialty coffee roaster" weights it least, because the modifier signals willingness to travel. Long-tail and specialty queries open up reach beyond the immediate radius.
• Service-area businesses get measured by service-area centroid. Plumbers, mobile pet grooming, in-home tutoring — proximity is calculated against the center of the declared service area, not the dispatcher's office. Pick service areas with that math in mind.

Pillar 2 — Relevance

Relevance is how well the business matches the query intent. This is the pillar most teams under-invest in because it requires precision rather than volume.

The signals Google reads for relevance, ordered roughly by weight:

• Primary category. The single biggest relevance lever. The chapter on Google Business Profile covers selection in depth.
• Secondary categories. Three to five well-chosen ones extend coverage to adjacent queries.
• Services and products text. Free-text fields are keyword-matched against query intent. Specific descriptions outperform generic labels.
• Business name match. Genuine name matches help (a literal "Joe's Bakery" gets a small boost on "bakery" queries). Stuffed names trigger suspension.
• Website content alignment. Google crosses the profile against the linked site. Tight topic alignment between profile and site amplifies relevance; mismatch dilutes it.
• Schema markup on the website. LocalBusiness, Service, and Product schema reinforce the entity-to-category mapping. Covered in the local on-page and schema chapter.
• Review content. Reviews that mention service names or category-defining terms reinforce relevance ("the best deep-dish pizza in town" carries a relevance signal beyond the star rating).

Pillar 3 — Prominence

Prominence is how well-known and well-regarded the business is, on the open web and offline. It's the pillar that takes the longest to build and the hardest to fake.

The signals Google reads for prominence, ordered roughly by weight:

• Review count and recency. Both matter. A business with 200 reviews where 30 came in the last 90 days outranks one with 500 reviews where the last one came two years ago. Velocity beats archive depth.
• Average rating. Above 4.3 is competitive. Below 4.0 is a structural disadvantage. Bands matter more than decimals — a 4.5 vs 4.6 difference is small; a 4.3 vs 3.9 is large.
• Inbound citations. Listings on directories, industry sites, local press. The chapter on NAP consistency and citations covers building these correctly.
• Inbound links to the website. Local press coverage, partner links, sponsorships. Links from locally-relevant sites are weighted higher than generic links for local-pack purposes.
• Brand-name search volume. When people search "[business name]" — proxied by branded query volume — Google reads this as prominence. A business that drives offline awareness producing branded search outranks one with similar profile data but no brand recognition.
• Profile engagement. Clicks, calls, direction requests, and bookings from the profile feed back into prominence. A profile that converts at the top of the map pack reinforces its position; one that doesn't gets demoted.
• Press mentions and Wikipedia/Wikidata entries. Major-press coverage and structured-knowledge entries shift the prominence floor for established brands.

How the pillars interact

The three pillars are multiplied, not added. A business with strong relevance and prominence but weak proximity can still rank for high-intent specialty queries from across the city; a business with strong proximity but weak relevance and prominence wins only the laziest "near me" searches and falls off competitive queries.

The practical implication: shore up the weakest pillar first. A business with great reviews and a great location losing the map pack to a competitor doesn't have a review or proximity problem — it has a relevance problem (probably category, services, or website alignment). The diagnostic below isolates which pillar is the gap.

The map-pack diagnostic

When a profile is losing rankings, run the three-pillar test against the top map-pack competitor for the target query. The pillar that comes back furthest behind is the one to invest in next.

• Proximity. Open the query in incognito, set the location to a customer-realistic address (not the business's own block — that biases the result). If the competitor is closer to that address, you have a proximity gap. If you're closer and still losing, proximity isn't the problem.
• Relevance. Compare the competitor's primary category, secondary categories, services list, and website topic focus against yours. Any major mismatch is a relevance gap.
• Prominence. Compare review count, recency of last review, average rating, citation breadth (search the business name without the city; count distinct domains), and branded search volume (estimate via a tool or via direct-traffic share). A two-fold gap on any of these is a prominence problem.

What's not a ranking factor (despite the noise)

Three things commonly cited as ranking factors that are not, or are vastly less weighted than the discourse implies:

• Posting frequency on its own. Posts contribute through engagement, not as a direct signal. A daily post that nobody clicks moves nothing. A weekly post that drives clicks and direction requests moves prominence.
• Profile completeness percentage. Filling every field is good practice but the percentage isn't the signal — the actual content of the fields is. A 100%-complete profile with weak categories ranks worse than a 90%-complete profile with the right primary category.
• Time since profile creation. Profile age is mildly correlated with ranking but causally weak. Established businesses tend to have older profiles, more reviews, and more citations — the latter two are the actual signals; age tags along.

The cluster path

With the framework in place, the next chapter — NAP consistency and citations — moves into the citation graph that feeds the prominence pillar. The chapters that follow each address one or more pillars: keyword research and on-page for relevance; reviews and map-pack tactics for prominence; multi-location for the architecture that lets prominence scale across regions.