Rank OS Scoring — the 5-dimension GBP health score.
Rank OS Scoring is Angryturtle's per-listing GBP health score. Five dimensions — Completeness, Consistency, Authority, Activity, Sentiment — each scored 0-100 from underlying sub-scores, then rolled up into a single portfolio number. Refreshed daily. 76 is the current average across 143+ managed Indian healthcare listings.
How it works.
Rank OS is the score under the score. Instead of a single opaque "health" number, five dimensions each run their own audit — 60+ sub-scores in total — then weight-blend into the 0-100 headline that appears on the Listing Overview screen. Each dimension is engineered to answer a specific operator question: is the listing fully populated, does its data agree with the rest of the internet, is it earning citations and links, is anyone actually maintaining it, and what do patients say about it. The reason for five instead of one is diagnostic — when a listing drops from 84 to 71 in a week, an operator needs to see immediately which dimension moved. A single blended score can never carry that.
- Completeness. Every GBP field populated, category primary and secondaries correct, services and attributes filled, photos present in every category.
- Consistency. Name, address, and phone match verbatim across GBP, website schema, and 40+ Indian healthcare directories.
- Authority. Citation depth, backlink profile of the listing's landing page, entity signals in structured data.
- Activity. Posts in the last 30 and 90 days, photo uploads, Q&A response rate, service edits, review-response velocity.
- Sentiment. Star average, recency curve, response coverage, review-attack detection, sentiment drift on new reviews.
- Roll-up. Weighted blend into a single 0-100, plus per-dimension delta vs the last 7-day and 30-day baseline.
Why it matters for
Indian healthcare.
The five dimensions carry compliance in ways a generic health score cannot. Sentiment inspects every review response for NMC Code of Ethics violations — a response that confirms treatment, names a diagnosis, or references clinical detail is auto-flagged before it publishes. Completeness reads the listing description against ASCI unsubstantiable-superlative filters, so a description carrying "best", "leading", "#1", or "most trusted" scores lower until rewritten. Consistency covers DPDP Act 2023 Section 8 — reviewer data across acquisition workflows must remain confined; a directory holding stale reviewer PII lowers the score. For IVF and reproductive-health listings, Completeness also runs an ART Act 2021 pass on the description — success-rate claims and comparative language never make it to a high Completeness score.
How ICG uses this.
Rank OS is the working screen on Angryturtle — every Monday, the ICG delivery team opens the Rank OS view, sorts by lowest-scoring dimension, and pulls the corresponding action queue into the week's task list. Across 143+ Indian healthcare listings, portfolio average sits at 76, with a working target of 80+ on every listing under active engagement. 4.76-star portfolio average and 28,137 aggregated reviews feed the Sentiment dimension; 531 risk factors feed the Authority and Consistency dimensions; 0 suspensions to date is the leading indicator that the whole model is working.
| Dimension | Sub-score inputs | Healthy threshold |
|---|---|---|
| Completeness | Fields populated · category · services · attributes · photo coverage | ≥ 85 |
| Consistency | NAP match across 40+ directories · schema parity · hours parity | ≥ 90 |
| Authority | Citation depth · backlink profile · entity signals · landing page trust | ≥ 70 |
| Activity | 30/90-day posts · photos · Q&A · edits · response velocity | ≥ 75 |
| Sentiment | Star average · recency · response coverage · attack detection | ≥ 80 |
| Rolled-up score | Weighted blend of five dimensions · refreshed every 24 hours | ≥ 80 (portfolio avg: 76) |
Frequently asked questions.
What is a good Rank OS score for a healthcare GBP?
Above 80 is a healthy managed listing. 60-80 is functional but under-optimised — usually a missing dimension like Activity or Authority. Below 60 signals a listing at meaningful suspension or ranking risk. The ICG portfolio averages 76 across 143+ managed Indian healthcare listings, with active work to push every listing above 80.
How are the five Rank OS dimensions weighted?
Weighting is not equal. Consistency and Sentiment carry the largest weight because both are the strongest local-pack ranking signals in the 2026 India-specific weighting. Completeness sits at a medium weight, Authority and Activity at slightly lower weights. Weights are re-tuned quarterly against actual rank movement in the ICG portfolio, not held static.
How often does Rank OS refresh?
Every 24 hours. GBP field state, NAP-directory scans, review corpus, activity logs, and geo-grid rank sample all recompute overnight. Any sub-score change above threshold generates an action in the next-morning task queue. Portfolio-wide roll-ups update alongside — an operator opens Angryturtle in the morning and sees yesterday-vs-today deltas.
Why is Activity a separate dimension from Completeness?
A listing can be 100% complete on day one and never touched again. Google reads dormancy as a demotion signal — no posts, no photo uploads, no Q&A responses, no service edits. Activity scores the last 30 and 90 days of cadence so that a fully-populated but abandoned listing shows up separately from a live, in-cadence listing on the same portfolio view.
Read next.
See your Rank OS score
on your listings.
ICG runs a free 30-min Brand & Growth Diagnostic that pulls your live GBP and shows exactly where Rank OS scoring sits on Angryturtle — dimension by dimension, with the next-action queue.
