Skip to content

Methodology

How CapitalScope sources, checks, calculates and scores the data on its pages.

Sources and provenance

Every important value is stored with its provenance: the source, the specific document or page, the date of the data, when we obtained it, where the value comes from (an official record, an estimate, an analyst's work, a market quote or a user submission), whether an analyst has verified it, and how confident we are. We keep the value exactly as the source stated it, so later corrections can be traced. The data sources page lists every source and the terms we use it under. We do not copy data from other websites unless their licence allows it.

Verification workflow

Records move through draft, pending review, verified and published. Only published records appear on the site. A record cannot be verified or published without a source, and data imported from files is checked row by row and approved by a second person before anything is written.

Financial statements and ratios

Financial statements are taken from audited annual reports, regulatory filings and other primary sources, in the unit the source uses (lakhs, crores and so on), and converted to rupees. Fiscal years are named by the year they end in: the year ending 31 March 2025 is FY2025.

Margins, growth rates, ROE, ROCE, debt to equity, current ratio and asset turnover are calculated by CapitalScope from the reported figures and are left blank whenever a figure they need is missing or the result would be meaningless — for example growth from a loss, or a return on negative equity. EBITDA is the reported figure where a statement gives one; otherwise, and only when all three figures exist, it is profit before tax plus finance costs plus depreciation, and is marked as calculated. The exact formulas are in our data dictionary.

Data confidence and quality score

Each company has a quality score from 0 to 100, shown as high (75 and above), medium (50–74) or low confidence, with the reasons. It adds up:

  • source reliability (up to 25): registries, regulators and audited reports score highest;
  • recency (up to 20): how recent the newest source is;
  • completeness (up to 25): how many key details are known;
  • verification (up to 15): how much has been checked by an analyst;
  • consistency (up to 15): reduced for each open data-quality issue, such as a balance sheet that does not add up or a year-on-year jump that suggests a unit error.

Unlisted share prices

Prices come only from dealers and platforms that have given CapitalScope permission to show them, and from transactions CapitalScope has verified. The indicative price is the median of each dealer's latest buy-sell midpoint, each platform's latest price and the latest verified transaction from the last 30 days. Confidence is high with three or more prices from the last week that agree within 5%, medium with two prices (or one recent verified transaction), and low otherwise. One snapshot is kept per day for the price history. See the unlisted securities disclosure.

IPO stages, deals and research

A company's IPO stage is the stage of its latest published IPO event, and each event — DRHP or RHP filed, IPO open, listed, withdrawn — needs its date and a source, usually a SEBI filing or a stock-exchange notice. Reported IPO timing is quoted as published; CapitalScope never predicts whether or when a company will list. Deals are recorded from disclosures and announcements, and a party is linked to a CapitalScope company only when its CIN or exact name matches. Research reports list their sources and draw their figures from CapitalScope's records, shown live in the text; a report that still contains an unchecked placeholder cannot be published.

Industries

Industries follow CapitalScope's classification of sectors and industries. A company is first classified from the National Industrial Classification code in its Corporate Identification Number, using the NIC edition in force when it registered; these classifications are marked as derived, with medium confidence, and analysts correct them. Companies registered while the NIC edition in CINs was changing are classified by analysts only.

Places

The state comes from the company register. The city or district is read from the registered office address, so a company may appear under a nearby town or its district.

Duplicates and identity

A company is identified by its CIN. Different spellings of the same name (“Pvt Ltd”, “Private Limited”) are matched, but two records are merged only by a person, never automatically. When a CIN changes — for example when a company lists, or the change of Telangana companies' codes in 2014 — the old CIN is kept so both lead to the same company.

Search engines

Company pages are offered to search engines only when they contain more than a registry entry (for example published financial statements or IPO information). Lists of companies by industry, state and city are offered when they contain enough companies to be useful.

Last updated 5 Oct 2026.