Know exactly why you rank. Fix what holds you back.

The situation before a score exists

Most owners find out something is wrong with their Google profile the hard way — a competitor two streets over starts showing up first, calls stop coming in on a Tuesday, and nobody can say why. A dentist in Pune with a single clinic near Kothrud might have a decent 4.3-star rating, twenty photos, and a description written three years ago. Nothing is broken exactly. It just isn't winning anymore. Without a way to weigh categories against reviews against how recently the profile was touched, there's no way to tell which of those three things is actually costing the clinic patients versus which ones are fine as they are.

Rank OS exists for that gap. It takes the same signals Google itself weighs — relevance, review health, freshness, entity authority, and AI-answer readiness — and turns them into one 0-100 number with five weighted parts behind it. Relevance and review health each carry 25 points, freshness carries 20, entity authority 15, and AIO readiness 15. The weights aren't fixed dogma; they're config-tunable, which matters more for an agency running fifty listings than for a single clinic, but even a solo owner benefits from seeing the math instead of trusting a black box.

Reading the number without overreacting to it

A score in the 70s is not a crisis. Most competitive categories in Indian metros cluster somewhere in the 60-80 range, so the real question isn't "is 74 good," it's "which of the five dimensions is dragging me down relative to who's actually beating me." If Review Health is the lowest of the five, replying to backlog reviews and asking recent happy customers for one will move more points than a description rewrite. If Freshness is lowest — nothing posted, no new photos, in three weeks — that's often the cheapest fix on the list, because it decays fast and recovers fast too.

A healthy read of the score looks like five dimensions sitting reasonably close together, with maybe one clear laggard and a ranked list of fixes attached to it. An unhealthy read looks different: one dimension sitting near zero while the rest sit comfortably above their own midpoint, which usually means something structural is broken rather than merely neglected — a suspended profile pulling Entity Authority to the floor, or a category mismatch quietly capping Relevance no matter how good the reviews are elsewhere. That second pattern is worth escalating past the ranked-fixes list and investigating directly, because no amount of small, incremental work closes a gap caused by something that's actually broken.

The false alarm to watch for is treating a small week-to-week dip as a signal. Rank OS recomputes on a sync cadence, not in real time, and Google's own data lags behind what's actually happening on the ground. A two-point drop after a slow week means very little. A dimension that's been low for a month and hasn't moved despite fixes is worth investigating properly — usually that's when a citation mismatch or a stuck verification is quietly capping the Entity Authority score no matter what else improves.

Working through it — the Pune clinic

Say the clinic's overall score comes back at 68. Relevance sits at 19 out of 25 — decent, but the category list hasn't been touched since opening and doesn't include "cosmetic dentist," which is where a chunk of nearby search volume actually lives. Review Health is 14 out of 25, dragged down less by the rating than by an 11% reply rate on the last thirty reviews. Freshness is sitting at 6 out of 20 because nothing has been posted in over three weeks.

The ranked list Rank OS produces doesn't tell the clinic to fix all three at once. It ranks them by expected point lift: replying to the review backlog first, because that's the single biggest jump available; adding the missing category second; posting an update third to stop the freshness decay. None of those three moves touches Entity Authority, which is already solid, so the clinic doesn't waste a week chasing citation cleanup that wouldn't move the score anyway.

A second decision shows up about a month later, once those three fixes have landed and the score has moved to 79. AIO Readiness is now the lowest dimension at 9 out of 15, and the ranked list points there instead — building out the question bank so the clinic's profile can actually be cited if someone asks an AI system which dentist near Kothrud does cosmetic work, rather than continuing to chase diminishing returns on a Relevance score that's already close to its own ceiling. That shift in where people ask the question at all is covered in more depth in how AI Overviews are changing local search in India.

Where Rank OS does not help

It has nothing to say about the clinic's actual website — no crawl, no page-speed check, no on-page SEO recommendation. That's a genuinely different discipline and Angryturtle doesn't pretend otherwise. It also won't tell you anything about backlinks; citations here mean directory listings, not link equity. And it isn't a promise: nothing in the model guarantees a map-pack position, because Google's own ranking algorithm has inputs Rank OS can't see or control, like proximity to the searcher at the moment of the query.

It's also worth being honest that the score updates on a lag. If a competitor pushes ten new reviews in one afternoon, that won't show up in your relative position instantly — the underlying GBP Insights data itself takes time to settle on Google's side before anything downstream can reflect it.

How it connects to the rest of the platform

Rank OS isn't a standalone report. The Review Health dimension is fed directly by AI-drafted review replies — the faster those get answered and published, the faster that quarter of the score responds. Freshness is fed by one-click Google posting, so a content calendar that's actually being used shows up as points, not just as activity for its own sake. Entity Authority reads from the NAP and citation work running underneath it, and AIO Readiness is scored by Ask Maps, which checks whether the profile can actually be cited by AI Overviews or ChatGPT when someone asks a relevant local question.

Geo positioning sits slightly outside the five dimensions but interacts with all of them — a geo-grid rank map will often show that a profile with a strong Rank OS score still loses in the neighbourhoods furthest from its pin, because distance is a factor Google weighs independently of profile quality. And once a fix is identified, whether it's a category, a business hour, or a phone number, it gets made through direct GBP editing that pushes straight to Google rather than sitting in a to-do list.

Agencies running a book of clients tend to use the score differently than a solo owner does — as a triage tool across ten or fifty listings rather than a single diagnosis, which is closer to what Agency OS and the pricing tiers above Solo are built for. Either way, the point of the number is the same: stop guessing which fix matters and start with the one that actually moves the needle.