टीपी भारत
TP Bharat · Data. Insight. Comparables.Explore the databaseAccessibility

Search guide

How to set up a search, choose industries and filters for the usual kinds of tested party, and read the results.

1. Set up

  • Data: start with TP Bharat Detailed. Switch to TP Bharat Abridged if the detailed set is too small, for example in a niche industry or for small companies.
  • Final financial year: the year of the transaction. If the year says “filings still arriving”, you can wait, use the previous year, or untick “must have data for the final year”.
  • Activity and business type: these set sensible default filters. You can change any of them in step 3.
  • PLI: OP/TC for service providers and manufacturers paid on cost, OP/Sales or GP/Sales for distributors, the Berry ratio for low-risk distributors and agents.
  • Range: Rule 10CA for Indian documentation, the OECD interquartile range for regional or global files.

2. Industry

There are three lists, and each item shows how many companies with data in your years carry it.

  • Industry group is the broadest, for example “Computer software”.
  • Main product or service is narrower and closer to what the company sells.
  • NIC code follows the National Industrial Classification. Type a prefix such as 620 to include every code under it.

With any, a company passes if it matches any of your selections, which is useful for gathering candidates. With every list, it must match each list where you made a selection, which gives a tighter set. Keywords search business descriptions and industry classifications, not company names.

3. Filters

The free search applies the full set of quantitative filters used in Indian transfer pricing practice, the same as our team's tool. The year filter comes first: the window is the final year you chose and the two years before it, a company needs accounts for at least two of the three (Rule 10CA), and every other filter uses only the years of the window the company has data for. Two more tests are required and cannot be switched off: operating revenue of at least ₹1 crore in every year used, and positive net worth in the latest year reported. The others (persistent losses, activity share, export earnings, employee cost, R&D, selling and marketing, royalty, related-party transactions, abnormal items, falling revenue, financial year end, super-normal profit and the data checks) each have a tick box and a value, ticked the way they are usually applied for the business type; untick or change anything that does not suit the facts.

One filter set per source. Each database combines two sources, and each source carries different lines, so every filter shows one column per source. A company is screened with the values of the source of its latest year with data, the funnel counts each source separately, and a filter a source has no data for (for example related-party transactions in Source 2) is not offered for it and never rejects a company for a missing figure. The company-by-company review in the Arm’s Length Analysis then takes the survivors of both sources together. The methodology explains what each filter is for.

Starting points

Tested partySet up asPLITypical filters
IT or software development centreServices · ITOP/TCExport 75% or more, employee cost 25% or more, R&D up to 3%, no losses in two of three years, related parties up to 25%
ITeS or BPO centreServices · ITeSOP/TCExport 75% or more, employee cost 25% or more, selling up to 3%, no losses in two of three years
Contract R&D or KPOServices · KPOOP/TCEmployee cost 25% or more, R&D filter off (R&D is the service), no persistent losses
Contract or toll manufacturerManufacturing · ContractOP/TCManufacturing 75% or more, R&D up to 3%, royalty up to 3%, selling up to 5%
DistributorTrading · DistributorOP/Sales or GP/SalesTrading 75% or more, R&D up to 3%, royalty up to 3%, no persistent losses
Limited-risk distributorTrading · Limited-riskBerry ratio or OP/SalesTrading 75% or more, selling up to 3%, no persistent losses

These are starting points. The right filters depend on what the tested party actually does.

4. Results

  • The funnel shows how many companies each filter removed. If one filter removes most of the set, look at it again.
  • The range is calculated on every company that passed. The review in an Arm’s Length Analysis will remove some of them, so the final range can move.
  • The table lists the companies by margin, highest first, with names hidden. The yearly margins show how stable each one is.
  • Your tested margin: enter it to see where it falls against the indicative range.
Too many companies? Narrow the industry, add keywords or switch to “every list”. Too few? Widen the turnover band, try TP Bharat Abridged, or loosen the related-party threshold if the facts allow.

Asking for the Arm’s Length Analysis

Click Request detailed review on the results. Your search settings are sent with the request. It helps to tell us:

  • The tested party, what it does, and the transactions to be benchmarked
  • The financial year, and the tested party’s operating revenue and margin
  • Comparables used in earlier years, and any the TPO has accepted or rejected
  • Your deadline, for example the due date of the accountant’s report, or the date of the TPO hearing
  • For an assessment or appeal, the assessment year (for example AY 2024-25, that is FY 2023-24)

If you already have a set made on another database, tick I already have a search in the request form and list the comparables. We review them, flag the ones likely to be questioned and add comparables from our database.

We will come back to you with the scope and the price.

Try a search on your own case

It is free and you do not need an account. Company names stay hidden until you order the analysis.