Where To Find Financial Data For Interviews: Terminals, Databases, Screeners And Filings

At 11:17 a.m., your exemplary says gross grew 18 percent; the institution filing says thing different. In finance, the look seldom kills the reply - the incorrect information root does.

Filings are the root of truth: yearly reports, quarterly results, investor presentations, DRHP/RHP documents and speech announcements.
Terminals are the fastest master layer: Bloomberg, Refinitiv Workspace, FactSet and Capital IQ aggregate prices, estimates, news and analytics.
Databases are champion for system history: CMIE Prowess, Capitaline, Ace Equity, RBI DBIE and MCA records thief pinch clip bid and adjacent comparison.
Screeners are find tools, not last evidence: usage Screener.in, TIKR aliases TradingView to shortlist, past verify against filings.
Use the root ladder: filing first for audited facts, terminal/database for speed, screener for scanning, news only for context.
In a modelling test, mention each hardcoded number: root name, date, play and page aliases filing link.
Best reply habit: triangulate 1 captious number from astatine slightest 2 sources earlier building the story.

The large image is simple: information sources beryllium connected a ladder of authority versus speed. A filing is slower to publication but hardest to challenge; a screener is accelerated but must beryllium verified. Strong candidates cognize which rung to usage for which question.

The information root ladder A layered pyramid showing that filings person the highest authority while screeners and news are faster but little final. Use the highest reliable root you tin defend Authority and auditability News, blogs, marketplace chatter Screeners Databases Terminals Filings Slowest to read Fastest to scan The person you are to a conclusion, the higher up the ladder your grounds should travel from.

The Core Idea: Match the Question to the Source

“Where do I find data?” is the incorrect first question. Ask: what determination will this information support? A adjacent valuation, a in installments memo, a market-sizing estimate and a merger exemplary request different information depth.

There are 4 awesome root families you should cognize cold.

A cleanable investigation workflow moves from accelerated find to defensible evidence. Start broad, constrictive the universe, verify the important numbers, past mention your assumptions.

Research workflow from mobility to cited model A five-step process showing really to move from a business mobility to a cited, defensible exemplary input. From messy hunt to defensible input Define the input Scan screeners Extract from filing Reconcile 2 sources Cite page Do not jump from a screener number consecutive into a conclusion. A exemplary input becomes interview-safe only aft it is verified and cited.

Where to Find Which Data

Use this arsenic your placement-day map. It is deliberately practical: the interviewer is checking whether you tin move from mobility to root without wandering.

How to Judge Data Quality

Good investigation is not “I recovered a number.” Good investigation is “I recovered the correct number, from the correct source, for the correct period, and I tin take sides it.” Track these measures while building a model.

Suppose a screener shows FY24 gross of ₹10,050 crore, while the yearly study shows consolidated gross from operations of ₹10,000 crore. Reconciliation quality = |10,050 - 10,000| / 10,000 = 0.5 percent. In a model, usage the yearly study number, adhd a statement that the screener differs by 0.5 percent, and cheque whether the screener includes different income aliases a later restatement.

Speed and authority matrix for information sources A two-by-two matrix comparing sources by velocity and authority. Choose root by velocity and authority Speed Authority Filings Highest proof Terminals Fast and rich Databases Structured history Screeners Scan, past verify The champion root is not ever the fastest source; it is the 1 that fits the determination risk.

Definitions

Filing: A general disclosure a institution submits to an speech aliases regulator for nationalist record.
Terminal: A paid investigation workstation that aggregates marketplace data, institution data, news and analytics.
Database: A system repository built to store, query and comparison standardized data.
Screener: A instrumentality that filters companies aliases securities utilizing selected financial, marketplace aliases operating criteria.
Triangulation: Verifying a information constituent against different reliable root earlier utilizing it successful a decision.

Case Study - Trent: Building the Story from Public Sources

Trent shows why a beardown expert does not extremity astatine 1 website: knowing Westside and Zudio requires filings, presentations, speech disclosures and adjacent context.

Good investigation feels for illustration detective activity - 1 root gives the number, different explains the business driver.

Situation: Trent, portion of the Tata Group, has drawn attraction because of its unit formats, particularly Westside and Zudio. A speedy screener tin show financial ratios, but it will not afloat explicate shop description , format economics, guidance commentary aliases really the business operation is evolving.

The move: A disciplined expert would build the position root by source. The yearly study gives audited financial statements and guidance discussion. Quarterly consequence filings and investor presentations thief way caller performance. NSE and BSE announcements corroborate charismatic disclosures. Databases aliases terminals thief comparison Trent pinch different listed retailers.

Outcome aliases lesson: The amended reply is not “Trent is simply a bully retailer because maturation is high.” The defensible reply is: maturation is the visible result, while the superior analytical driver is format description and execution quality, supported by marque positioning, shop economics, supply-chain subject and user demand. Each portion must beryllium backed by the correct source.

How AI Changes Where to Find Data

AI does not region the request for root judgement. It changes the velocity astatine which you tin search, summarize and reconcile.

Filing hunt becomes faster: LLM devices tin summarize yearly reports, extract consequence factors and find applicable pages, but you must still verify the page successful the original PDF.
Terminal workflows go much conversational: Research platforms are adding natural-language hunt complete transcripts, filings and estimates, making it easier to inquire “what changed successful separator guidance?”
Screeners go thought engines: AI-assisted screeners tin propose adjacent sets aliases emblem anomalies, but last numbers still request filing-level validation.

Load a institution yearly report, latest quarterly consequence and investor position into NotebookLM. Ask it to create a source-linked array of revenue, EBITDA, debt, capex, cardinal risks and guidance commentary. Then unfastened each cited page yourself and paste only verified numbers into your model.

Interview Relevance

“You person 90 minutes to build a speedy valuation exemplary for an Indian listed company. Where will you get the data, and really will you make judge it is reliable?”

Use this condemnation successful interviews: “For facts, I spot filings; for speed, I usage terminals aliases databases; for discovery, I usage screeners; for judgement, I reconcile and cite.”

Common Mistake

The correction that costs candidates is treating a screener number arsenic last evidence. It costs you because screeners whitethorn usage different definitions, old periods aliases consolidated versus standalone data. The fix: usage screeners to find, filings to prove, and root notes to defend.

What to Revise Next

Now that you cognize wherever the information lives, revise the adjacent applicable step: Case Study: A Timed Modelling Test Like the Ones Firms Set. That will thief you person originated information into assumptions, statement items, checks and a cleanable proposal nether clip pressure.