See which demand is rising — before your competitors do.

Four words: Scaling, Broadening, Concentrating, Decaying

Those are the only four states a demand cluster can be in, and most of the value of this feature is just learning to trust what they mean before the raw keyword numbers confirm it weeks later. Scaling means more people and more different search terms are both rising together — real, broadening demand. Broadening means the keyword list is growing while impressions stay flat, which sounds unremarkable but is actually healthy: it means new ways of searching for you are appearing before the volume behind them has caught up. Concentrating means demand is narrowing onto fewer terms, which isn't automatically bad but deserves a look. Decaying means both the terms and the impressions are shrinking, and that's the one that should interrupt whatever else is on the to-do list that week.

The mistake this catches

Most small businesses pick their content and post topics based on what used to work, updated occasionally by gut feeling. A salon owner in Koramangala kept posting about hair spa packages because that's what built the client base three years ago, while the demand cluster around "keratin treatment Koramangala" had been quietly scaling for two quarters. Nobody was lying to her — the hair spa cluster still had decent volume — but she was pouring content into a cluster that was concentrating while ignoring one that was scaling, and Cluster Potential (sized from Google Ads Keyword Planner, scoped to her city) would have told her the second cluster was worth roughly four times the search volume of the first.

How to read the screen without overreacting

A cluster tagged Decaying isn't automatically an emergency. Seasonal businesses — wedding venues after wedding season, tax consultants after filing deadlines, school-admission consultants outside the admission window — will show entire clusters decaying on a predictable annual clock, and chasing that with new content is wasted effort. The state only matters against the calendar and against Cluster Potential together: a small, decaying, low-potential cluster is background noise; a large, decaying, high-potential cluster in a business's actual off-season is worth investigating, because it might mean a competitor took the demand rather than the demand disappearing.

The cluster-wise traffic compare is the tool for telling those two apart. Pull the same cluster across the same period last year. If the decay lines up with last year's calendar dip, it's seasonal. If this year's dip is steeper or starts earlier than last year's, something outside the calendar moved — usually a competitor, sometimes a category change on the profile itself, occasionally a citation or NAP problem quietly pulling the listing out of consideration.

A second decision: which cluster gets this month's budget

Every business with more than one service line eventually has to decide where the next round of content, posts and paid attention goes, and that decision is usually made by whoever's most vocal in the room rather than by what customers are actually searching for. A modular kitchen showroom in Jaipur with three product lines — kitchens, wardrobes, and office interiors — had always assumed kitchens deserved the lion's share of attention because that's what the brand was known for. The cluster data told a different story: the kitchens cluster was Concentrating, meaning demand was narrowing to a smaller set of repeat searchers, while the wardrobes cluster was Broadening, picking up new search phrasing every few weeks. Reallocating even a third of the content calendar toward wardrobes, based on that read rather than on brand history, is the kind of decision this feature is built to settle.

A worked example: a Jaipur furniture workshop

A custom furniture workshop in Jaipur sold mainly through referral and had never taken keyword clusters seriously. Its "custom wardrobes Jaipur" cluster was Scaling steadily, but Cluster Potential sized it as modest next to a smaller, newer cluster — "modular kitchen Jaipur" — that was Broadening fast, adding new search terms every few weeks without a matching jump in impressions yet. Broadening-but-not-yet-scaled is the window to move before the crowd notices, and it's exactly the situation where posting into a Content Studio calendar early, rather than waiting for the volume to prove itself, pays off the most. Waiting for Scaling to show up before acting means arriving after several competitors already have.

Where this doesn't help

Demand Clusters reads your own first-party GBP data plus Keyword Planner sizing — it does not see what your competitors' clusters are doing, only what's happening around your own listing. Two businesses two streets apart, one strong and one weak, will show the same city-level Cluster Potential number even though one of them is capturing most of it. Closing that particular gap needs Competitor Intelligence sitting next to this feature, not instead of it. The clustering also depends on the industry taxonomy being a reasonable match for the business — a hybrid business that spans two categories (a café that also does event catering, say) will sometimes see its demand split unevenly across clusters that don't map cleanly onto how the owner actually thinks about the two halves of the business, and that split is worth sanity-checking rather than trusting blind. It also won't tell you whether a scaling cluster is scaling because of genuine new demand in the city or because of a seasonal event calendar quirk specific to India's festival and school-admission cycles — the AEO India playbook covers that seasonal pattern in more detail than a per-listing dashboard reasonably can.

What it connects to

Relevance is a quarter of the Rank OS score, and cluster coverage is a direct input into it — a profile that's strong in a Scaling cluster and thin in everything else will show that unevenness in the relevance dimension specifically, and in the broader set of local ranking factors Google weighs. The momentum read also feeds directly into what Ask Maps should be watching for AI-answer readiness, since a cluster that's scaling in ordinary search tends to show up in AI-generated answers not long after. Because Cluster Potential is scoped by city, it plugs into how Insights & Performance should be read too — a rising discovery number only makes sense once it's clear which cluster is actually driving it. Multi-location brands running the same cluster taxonomy across branches will find it interacts with multi-location SEO planning generally, and any change worth making usually gets pushed live through Edit Location once the priority cluster is chosen.