Home › Methodology

How WIB Verifies Data

WIB's competitive advantage is honest, manually-verified data. Most Indian comparison sites publish scraped or sponsored rankings; we don't. This page explains exactly what "verified" means on each of our datasets, how research reports are written, and where we acknowledge the limits of our coverage.

Methodology Maintained by: WIB Editorial · Last updated: May 2026 · Revision cycle: Quarterly
On this page Editorial principles Hospital data Restaurant data Schools & colleges Research reports Data freshness Reporting errors What we don't do

01Editorial principles

Four rules every dataset and every report has to pass through.

  1. Manual verification, not scraping. Every business in our hospital and restaurant datasets is added by editorial review, not a crawler. We check the source website, validate addresses, and confirm specialties/cuisines against the business's own listing — not third-party aggregators.
  2. No paid placements, ever. No hospital, restaurant, or school can pay for inclusion, higher rating, or better placement on WIB. This is a hard rule; there are zero exceptions. If a business asks to pay, we decline and document the request.
  3. Open data, transparent attribution. Every dataset is published as CSV and JSON at /data/ under CC-BY 4.0. Anyone can verify our numbers. When we publish a research report, we link to the underlying dataset so the analysis is independently reproducible.
  4. Acknowledge what we don't know. WIB is editorial, not statistical. Our samples are curated, not random. Every report includes a methodology section that explains coverage limits. We publish what we can stand behind; we don't manufacture coverage to sound bigger.

02Hospital data — verification protocol

WIB tracks 463 multi-specialty hospitals across 36 Indian cities (as of May 2026). Each hospital record carries 9 fields: name, city, type (Government / Private / Trust), bed count, rating, specialties, NABH accreditation, address, and a source URL pointing to the business's WIB page.

Every record is added or updated via this protocol:

  1. Source identification. Hospital must have a verifiable public website OR a current presence on Practo / Apollo247 / Vesta Care / official AIIMS portal. Dental clinics, single-specialty chains, and government primary health centres are excluded — only multi-specialty hospitals qualify.
  2. Name + address cross-check. Hospital name and complete street address are confirmed against the hospital's own website. Where the website is dormant, we cross-check with Google Maps listings and the local NABH registry.
  3. NABH status check. We mark a hospital as NABH-accredited only if it appears on the National Accreditation Board for Hospitals registry at the time of inclusion. JCI status is checked separately when claimed.
  4. Bed count. Sourced from the hospital's own About page or annual report where available. Where bed count is not publicly stated, we either omit it (mark as null) or use a conservatively-reported range from the latest credible secondary source.
  5. Specialty list. Capped at the top 5 specialties the hospital itself highlights on its homepage. We don't list every specialty a hospital offers; we list the ones it positions as primary.
  6. Rating normalisation. Ratings are aggregated from publicly available review platforms (Google, Practo) on a 5-point scale. We don't accept reviews submitted to WIB without verification. Where a hospital has fewer than 10 public reviews, we either omit the rating or use a conservative 4.0 default and flag it.
  7. Quarterly re-verification. Each hospital record is re-checked every quarter. If a hospital's website is dead, NABH lapses, or the entity has closed, we mark the record stale and either update or remove it.
The single biggest quality signal we publish: NABH accreditation status. It's a binary, government-verifiable fact, and the most rigorous independent quality marker available for Indian hospitals. When we say a hospital is NABH-accredited, you can verify it on the official NABH website.
What our hospital data does NOT tell you Bed availability on a given day, individual doctor quality within a hospital, specific surgical outcomes, infection rates, or pricing for procedures. These are real and important questions; they require primary data WIB doesn't yet collect.

03Restaurant data — verification protocol

WIB tracks 605 top-rated restaurants across 28 Indian cities. Each record carries 8 fields: name, city, cuisines, rating, budget tier, address, vegetarian-friendly flag, and a source URL.

  1. Source identification. Restaurant must be currently operational with a verifiable presence on Zomato / Google Maps / its own website. Pop-ups, ghost kitchens, and delivery-only operations are excluded.
  2. Cuisine tagging. Up to 4 cuisines per restaurant. Editorial judgement, not auto-tagged. A North Indian restaurant that also lists "Continental, Chinese, South Indian" gets tagged with its actual identity — usually 2-3 cuisines max. We don't accept the "Multi-Cuisine + every cuisine" tag pile that diluted Zomato's data.
  3. Budget tier classification. Three tiers: under Rs 500 per person, Rs 500-2,000, and Rs 2,000+. Anchored in cost-for-two estimates from Zomato + verified by editorial sampling.
  4. Rating normalisation. Aggregated from Google / Zomato on a 5-point scale. Restaurants below 4.0 rating are generally excluded — WIB's coverage is "top-rated," not exhaustive.
  5. Vegetarian flag. Set to "Yes" only for restaurants that are pure vegetarian. Restaurants that serve non-veg but offer extensive veg menus are marked "No" — the flag means strictly vegetarian, not vegetarian-friendly.
  6. Regional cuisine emphasis. Where a city has a distinctive regional cuisine (Saoji in Nagpur, Bhopali in Bhopal, Mangalorean coastal in Mangalore, Odia in Bhubaneswar, Assamese in Guwahati, sadya in Trivandrum), WIB intentionally over-samples regional-anchored restaurants. This is editorial, not statistical — see the Regional Cuisine Renaissance report for why.
What our restaurant data does NOT tell you Daily availability, kitchen hygiene scores (no Indian equivalent of UK FHRS), individual dish quality, real-time wait times, or whether the chef is still in place. We capture stable structural information, not day-to-day operational signals.

04Schools & colleges

WIB's schools and colleges datasets are smaller and Firestore-backed (not yet exported as open CSV). Coverage:

Schools and colleges go through the same editorial-verification process as hospitals (Step 1-4 above), with the addition of board / accreditation registry cross-checks for institutional credentials.

05How research reports are written

WIB publishes original India-focused research reports. Each report is built on data WIB already maintains, plus editorial analysis. Reports are not commissioned by industry, sponsored, or paid for in any form.

The standard report workflow:

  1. Question framing. A report starts with a specific question we can answer using our dataset (e.g., "What does tier-3 hospital quality actually look like?"). If the question requires data we don't have, we either gather it editorially or don't publish.
  2. Sample definition. We define the exact sample — which cities, how many businesses, how selected. This is curated, not random. We over-sample where editorial judgement says it matters (e.g., we pick the top 8 hospitals in each tier-3 city rather than randomly sampling 40 hospitals nationwide).
  3. Comparative anchor. Every report compares the focus dataset against a benchmark. The tier-3 report compares against our metro cohort; the city living index compares 8 metros side-by-side; the salary report breaks down by city × experience × role.
  4. Editorial analysis. Numbers don't write the report; editorial does. We interpret what the data means, where the gap is, what would close it. The interpretive prose is always WIB editorial opinion, distinct from the numerical findings.
  5. Methodology section. Every report includes a "Method & data" section explaining sample, source, and limitations. Datasets are linked at /data/ so any reader can verify.
  6. Publishing. Reports are published as static pages with Report + Article + FAQPage schema. Each report has a stable URL and date. We don't backdate revisions; significant updates are published as new reports.

06Data freshness commitments

07Reporting errors and corrections

WIB makes editorial mistakes. Bed counts are mis-typed. NABH expires without our knowledge. A restaurant closes and we miss it. We treat error reports as gold.

If you spot an error:

Major corrections (factual errors in a research report, dataset rollbacks) are documented in the changelog. We don't silently edit published research.

08What WIB doesn't do

Equally important to what we do is what we explicitly refuse to do:

Editorial standards in one paragraph

Every datapoint on WIB is added by editorial review against the business's own source. No paid placements, ever. Data is open and citable. Research reports are anchored in publicly-linked datasets. We acknowledge what we don't know. When we make a mistake, we correct it visibly. Methodology is documented here, on a stable URL, and updated quarterly. If a fact on WIB is wrong, tell us and we'll fix it within 5 working days.

Want to know about a specific dataset's methodology? Email hello@wibest.in. Want to cite WIB in research? See /data/ for citation format and CC-BY 4.0 terms.