Concall vs Annual Report: What Each Reveals

I used to think the annual report was the serious document and the concall was the marketing event. I had it backwards, and it cost me. The annual report is what management is willing to write down, edit at leisure, and have audited. The concall is what management is forced to say out loud when an analyst pushes on the soft spot — and then, three quarters later, what management quietly stops saying. Both are honest in their own way. They just hide their truths in completely different places, which is the whole reason a careful investor reads one against the other.

The gap is measurable. Across roughly 18,700 management commitments tracked on Inve (about 1,600 companies, as of mid-2026), most guidance never reaches a clean "delivered as stated" outcome: of the lines that have actually resolved — set aside the ones still in flight — only about two in five (roughly 39%) ended up achieved as originally framed. And more than half of all companies tracked have at least one piece of guidance that quietly went silent — given on one call and never mentioned again on any later call. (These are Inve's tracked figures and move with each quarterly snapshot; the denominator for the "achieved" rate is resolved guidance only, not in-flight targets — we unpack that base rate in do Indian companies deliver on their guidance?) You can only catch that silence by reading the calls in sequence and checking them against the annual record. Neither document, alone, shows it to you.

This is a field guide to the two: what each is for, what each cannot do, and how to read them against each other — built around a single real case so you can watch the same fact behave differently in each source.

What is the core difference between a concall and an annual report?

One looks back; one looks forward. One is audited; one is live. That single pair of differences explains almost everything.

The annual report is the audited, once-a-year record. It is legally attested, comprehensive, and backward-looking — it tells you, with accountability, what happened over the full financial year. It is also written and edited at leisure, which means the narrative sections are polished and the uncomfortable facts are placed where they are technically disclosed but easy to miss.

The concall is the quarterly, unaudited, real-time picture. It is forward-looking — management gives guidance on what is coming — and, in the Q&A, it is unscripted. It is where you hear management answer a question they did not get to rehearse.

A real case: the same fact in both documents

Here is the homely version of what follows. Imagine a contractor who tells you in March that your renovation will cost ₹14–15 lakh. By June he's quietly revised it down to ₹11–12 lakh; by August he's talking ₹13–15 lakh again; by autumn he's stopped quoting a total at all and only talks about "the demand environment for tiles." By the time the written bill arrives, it records ₹8 lakh of work done so far, "margins impacted by input costs," and not a word about any of the totals he quoted along the way. Each statement was true on the day he said it. Only the sequence tells you the estimate kept moving — and that the paperwork never reconciles to any of the totals he quoted out loud.

That is almost exactly what happened at Centum Electronics, a Bengaluru defence-and-aerospace electronics maker. This is an illustration of how to read the two documents against each other, not a view on the stock.

On the Q2 FY25 concall (October 2024), asked about the medium term, management said it plainly: "in the next 2 years, 3 years we should be able to go to 14 or 15" on consolidated EBITDA margin (Centum Q2 FY25 concall). A clean, specific, forward number — the kind you can hold management to.

Then watch the same target across the next five calls, exactly as Inve's Promise Tracker logged it:

The path was not a clean one-way descent. The 14–15% number, given as a firm two-to-three-year target, went quiet one quarter, was cut to 11–12% the next, then was raised back to 13–15% the quarter after that, and then went quiet again on the two most recent calls. Nobody ever stood up and said "we are abandoning the 14–15% guidance" — but nobody clearly re-committed to it on a fixed timeline either. That non-monotonic, in-and-out pattern is the thing to notice: a target that keeps moving and keeps disappearing from the conversation, rather than a single retraction you could point to.

Now open the audited document covering that same period. The FY25 annual report describes the consolidated result as "a stable EBITDA margin of 8.3%, underscoring our focus on margin resilience and long-term fundamentals" (Centum FY25 annual report). It is a true sentence. It is also a backward-looking sentence with every incentive to leave out the forward number that was given to analysts and then walked around — and it does leave it out. The phrase "14–15%" does not appear in the report at all. Read the annual report alone and you get a calm, plausible story: a stable 8.3% margin, resilience, fundamentals. Read the calls in sequence and you get the more useful story: a specific target that was cut, re-raised, and repeatedly left out of the conversation — a moving number that the written record simply never engages with.

Neither source is lying. The annual report selects; the call record remembers. Only together do they tell you what actually happened to the commitment.

Where does the truth hide in an annual report?

In the parts management is required to disclose but never highlights. The Chairman's letter, the photographs, the vision statement — that is the curated front. The truth is further back, and Centum's own FY25 report shows you exactly where.

The notes to accounts (the subsidiary schedule). The headline "stable 8.3%" consolidated margin is the average of a healthy Indian parent and a set of bleeding overseas units. You only see the bleeding in the AOC-1 subsidiary statement, deep in the notes: Centum's two Canadian subsidiaries posted after-tax losses of ₹113 million and ₹10 million for FY25 — the AOC-1 Part A figures are reported in INR million, with CAD shown only as the functional-currency column for the exchange-rate note — and the French entities lost more on top (Centum FY25 annual report, AOC-1). The MD&A admits the direction — "the ER&D segment continues to face profitability headwinds, particularly in its Canadian subsidiary… revenue and margin levels have remained inadequate to offset operating costs" (Centum FY25 annual report) — but the size of the drag lives in a table most readers never reach. That single subsidiary line is what the 8.3% average quietly absorbs, and it is the same Canada problem the concalls kept circling.

Contingent liabilities. Also in the notes, disclosed without fanfare: income taxes under dispute of ₹92.06 million (nearly double the ₹46.44 million a year earlier) and indirect taxes under dispute of ₹116.53 million, all flagged as "possible, but not probable" (Centum FY25 annual report). Whether you agree with "not probable" is your call — but you can only make that call if you read past the narrative to the number.

The cash flow statement. Reported profit is an accrual figure; cash flow shows whether the profit actually arrived as cash. A profitable company with weak operating cash flow has a working-capital story the headline never tells. (We don't have Centum's cash-flow statement parsed in our database yet, so we're flagging that as a gap rather than inventing a figure — read it straight from the report for any name you own.)

The annual report rarely lies. It selects. The skill is reading the parts it does not invite you to read — for the full method, our guide on how to read an annual report like an analyst walks through the order.

Where does the truth hide in a concall?

In what management stops saying.

The annual report hides truth in placement; the concall hides it in omission. A target that was front and centre last quarter and is simply not mentioned this quarter has, in effect, gone quiet — but no one announces it. Centum's 14–15% line shows the messier reality: it wasn't retracted, and it wasn't continuously ghosted either — across the five calls that followed it went silent, was cut to 11–12%, was re-raised to 13–15%, then went silent again on the two most recent calls. The signal isn't a single dropped quarter; it's a number that keeps drifting in and out of the conversation. A target an analyst has to keep re-extracting, and that management frames a little differently each time, is a topic under stress.

The Q&A is where this lives, because it is the only adversarial test a company faces in public. Prepared remarks are written by the investor-relations team; the Q&A is live — and if you want the section-by-section method, our guide on how to read a concall transcript covers where to look first. The 14–15% number itself came out under questioning, not in the scripted opening. Analysts probe the exact soft spots — the margin walk, the loss-making subsidiary, the one-off that keeps recurring — and whether management answers directly or changes the subject is information the audited report can never carry, because it is about how they communicate, not what they reported.

How do you read the two against each other?

Use each to verify the other. That cross-check is the whole technique, and the Centum case gives you the three questions to ask of any company.

Does the concall guidance show up — delivered or missed — in the next annual report? The 14–15% margin target given on the Q2 FY25 call should be traceable to an outcome in the FY25 report. Instead the report posts 8.3% and never mentions the target. That silence — a specific forward number that simply evaporates from the backward record — is the narrative drift that matters most.

Does the annual report's average hide a divergence the calls were flagging? The calls kept returning to the Canadian subsidiary; the report's 8.3% consolidated margin is precisely the number that blends that drag away. When a concall keeps naming one problem and the annual report reports a smooth average, go find the segment or subsidiary note that reconciles them.

Do the audited annual numbers match the quarterly framing? Management's adjusted, "one-off-excluded" quarterly numbers should reconcile to the audited full-year figures. Persistent "one-offs" — like a subsidiary loss that recurs every quarter — are just costs, and the annual report is where they finally show up consolidated.

Doing this manually means holding several quarters of transcripts and the annual report per company in your head, across a whole portfolio, quarter after quarter. That memory layer — every commitment, what happened to it, what went silent — is what Inve's Promise Tracker keeps as a ledger, while the concall summaries flag each guidance line as delivered, revised, or quietly dropped. The tools don't replace reading the annual report for your top positions; they make the cross-check tractable across the rest.

Where this read could be wrong

The strongest case against treating a ghosted target as a red flag is this: a two-to-three-year guidance number given off-the-cuff under analyst questioning is not a contract, and a management that re-cuts 14–15% to 11–12% when a subsidiary deteriorates is arguably being honest, not evasive — it is updating to reality rather than clinging to a stale number to save face. And Centum's own record cuts in management's favour here, which is exactly why the full table matters: the number didn't only get walked down. On the Q1 FY26 call (August 2025), management re-raised the medium-term objective back to 13–15% — Inve logged that quarter as on-track, not ghosted (Inve Promise Tracker, CENTUM). A team that only ever lowered the bar would look different from one that cut it under pressure and then talked it back up as the subsidiary fixes took hold. Management also named the Canadian problem repeatedly and is "actively evaluating strategic alternatives" for it (Centum FY25 annual report). A team that hid the subsidiary entirely would be worse. The honest read, then, is that this is a non-monotonic path — down, then up, then quiet — and that ambiguity is the point: it's why a single ghost, or even a single cut, is a question, not a verdict, especially on a transcript record only about two years deep. The signal to watch isn't "they moved one target." It's the longer-run pattern: does management eventually own the gap on the record and reconcile guidance to outcomes, or does the written report only ever post the smoothed average while the forward number quietly drifts? One moving number is noise. A management whose hard targets keep slipping out of the conversation while the annual report launders the result is telling you something about how it will communicate the next bad year — which is the year you actually need it to be straight.

So which one should you read first?

Read the annual report to understand the business and its governance — the audited, full-year truth, with the notes and cash flow read before the narrative. Read the concalls to track how management communicates now — forward guidance, and whether they deliver it. The annual report tells you what kind of company you own; the sequence of concalls tells you whether the management is still the management you thought you were backing.

The owner's question at the end of both is the same: three to five years out, what must be true for this to work — and when it doesn't go to plan, will this management say so on the call, or will you have to find it yourself in a subsidiary footnote a year later?

Frequently asked questions