How to Analyse an Indian IT Services Stock

An Indian IT services firm beats profit estimates, the stock jumps, and then one sentence on the concall — "we're seeing some softness in discretionary spend" — quietly erases the gain by the next morning. The headline number was the past; the sentence was the future, hiding in the deal pipeline and the book-to-bill rather than the profit line. An IT services company is, at bottom, a labour-arbitrage business that sells billable hours at scale, and almost everything that decides the next two years shows up not in this quarter's revenue but in the forward indicators — and in whether management's reading of them holds up call after call.

We'll be blunt about our own blind spot first, because it's the same one that catches most readers. When we started tracking IT managements quarter by quarter, we under-weighted medium-term margin aspirations — the "we'll get to 15% over the medium term" lines that aren't a hard number anyone can hold you to. They sound like guidance. Across the broader corpus they behave more like mood music: across more than 15,700 management commitments tracked on Inve, barely half are delivered as stated and over 1,300 were simply never mentioned again on any later call (Inve data, as of 2026-06-12). The ones that vanish are disproportionately the soft, directional ones — and an IT firm's margin story is usually told in exactly that register.

This piece is about the forward indicators that drive an IT services business, the concall questions that separate a real recovery from a hopeful one, and the one disclosure habit that should make you nervous. One boundary up front: demand signals are noisy and lagged. You can read the direction and test management's candour about it; you cannot forecast the cycle. That is the circle this stays inside.

Why is revenue growth the wrong place to start?

Reported revenue is a lagging indicator — bills going out on work sold months ago — so the experienced reader starts with the leading indicators that become revenue: the order book and the deal wins.

There is a second trap in the revenue line itself: currency. An Indian IT firm earns most of its money in US dollars (plus euros and pounds) but reports in rupees, so a weak quarter in dollar terms can look like growth in rupees simply because the rupee depreciated. This is why the industry reads constant-currency (CC) revenue growth, which strips out the exchange-rate move to show whether the business actually grew. Always read CC, never reported rupee growth, to know how the business itself did.

What does TCV actually tell you — and what does it hide?

Total contract value (TCV) is the headline deal-win number: the full value of new contracts signed in the quarter, and a big print reliably moves the stock. It is a genuine leading indicator — but also the number most easily dressed up, and reading it naively is a common error.

Three things to interrogate behind a strong TCV:

• Duration. A ₹10,000 crore TCV over ten years converts to revenue far slower than the same number over three years, so a rising TCV driven by longer-duration deals can mean flat or falling near-term revenue. Always ask the deal tenure.
• Net new versus renewals. Mostly-renewal TCV is the company running to stand still; net-new TCV is genuine expansion. Managements love to report total TCV and stay vague on the split — the vagueness is the tell. When Cyient noted on its Q1 FY26 call that the "non-renewable portion" of order intake had risen to 21% from 18% (Cyient Q1 FY26 concall), that single line said more about the quality of the order book than the headline intake figure did.
• Book-to-bill. TCV divided by revenue (annualised) tells you whether the firm is signing faster than it is billing. Comfortably above 1 means the backlog is growing; below 1, quarter after quarter, means the pipeline is thinning even if this quarter's revenue looks fine.

In short: only net-new, shorter-duration TCV with a book-to-bill above 1 is a signal you can lean on. A record total-TCV headline sitting on renewal-heavy, long-duration deals with a sub-1 book-to-bill is engineered to reassure, not inform.

A framing note: TCV and book-to-bill are sector reporting conventions IT managements volunteer, not part of the formal NISM-XV KPI taxonomy. The curriculum's named forward KPIs for IT/BPO are average FTEs billed, revenue per FTE, bench strength and attrition, constant-currency growth, and customer concentration. Treat deal-book metrics as a useful supplement to those, not a substitute.

How do attrition and utilisation reveal margin before margin moves?

This pair lets you see the operating story a quarter ahead of the P&L.

Utilisation is the share of billable employees actually deployed on client work (versus sitting "on the bench"). Every benched engineer is a salary paid with no revenue earned, so utilisation is one of the most direct levers on margin — alongside wage costs, attrition, pricing and offshore mix — and a couple of points can swing it meaningfully. When demand softens, utilisation falls first (the bench fills before revenue drops), making a falling rate an early warning the revenue line won't show for another quarter or two.

Attrition is the rate at which employees leave, and it cuts both ways. High attrition (the 20%+ levels of the post-2021 talent war) is expensive — replacement hiring, training, and wage inflation all compress margin. But very low attrition in a soft demand environment isn't automatically good either: it can mean people aren't leaving because there's nowhere to go, often alongside a hiring freeze and shrinking headcount — a quiet sign the firm sees thin demand ahead.

Rising utilisation with cooling attrition is the textbook margin-expansion setup; falling utilisation with a frozen headcount is demand softness dressed up as "cost discipline." The question to carry into a results season: which way are these two moving, and does management's margin guidance match what they imply?

What should you make of the margin and the headcount line?

Operating margin (often reported as EBIT margin) tends to sit in a band, and the number to watch is the direction and the story, not the absolute level. Margin is squeezed by wage hikes (annual, usually April), high attrition, low utilisation, and pricing pressure; it is expanded by utilisation gains, automation, offshore mix shift, and pricing power. A guided band is implicitly a forecast of how those forces net out.

Headcount is the demand barometer hiding in plain sight. IT firms hire ahead of demand and cut ahead of weakness, so net headcount turning negative for consecutive quarters — even while revenue holds — is one of the clearest forward signals in the sector: management voting with its own cost base on what it expects. A firm talking up a "strong pipeline" while quietly shrinking headcount is contradicting itself.

Read the client side alongside it. The curriculum names customer concentration — and the count of million-dollar clients — as core IT/BPO KPIs and a key constraint: a top-client or top-10-client revenue share that is creeping up, or a million-dollar-client count that has stopped growing, is a forward demand-and-risk signal sitting right beside headcount and utilisation.

A real case: when the margin band keeps walking away

Forget hypotheticals. Watch what an actual margin aspiration does when the demand doesn't show up. Here is Cyient's mid-tier engineering-services arm (the DET, or Digital, Engineering & Technology, business) talking about the same target — a 15-16% EBIT margin — across four consecutive calls, with the number the company actually delivered in the last row.

Illustration of how to read a guidance sequence — not a view on the stock.

Source: Cyient Q3 FY25, Q1 FY26, Q2 FY26 and Q4 FY26 concalls; FY26 actual confirmed in the Q4 FY26 transcript ("Normalized EBIT margin for FY26 is 12.2%") and on Inve's Promise Tracker (Inve data, Q4 FY26).

Read the sequence the way you'd read a friend's running excuse for a loan he keeps not repaying: first the date, then a vaguer date, then "soon," then silence on the original commitment altogether. The number didn't just miss — the target itself migrated from a hard year (FY26) to a soft horizon (medium term) to a new exit quarter two years out, each revision sounding reasonable in isolation. Only stacked in one table does the pattern show: a margin that was supposed to recover this year is now a 15% aspiration for Q4 FY27, while the actual print drifted the other way to 12.2%. On the Q4 FY26 call, an analyst (Madhur Rathi) pressed on recouping the lost profitability — roughly the ~Rs 1,300 cr of EBITDA the company had earned in FY24, and a return toward ~18% EBITDA margins. Management reframed the question in its own words — "when will we come back to, let's say, a 15% EBIT margin?" (Krishna Bodanapu, Vice-Chairman) — and the answer was, in effect, the same medium-term horizon the company had been offering for two years (Cyient Q4 FY26 concall).

The damage isn't abstract. As that engineering margin sagged, Cyient's consolidated quarterly net profit fell from ₹186 crore in Q4 FY25 to ₹66 crore in Q4 FY26, with operating margin slipping from 16% to 12% over the same span (Inve data, Q4 FY26). A reader who took the "15% medium term" line at face value in mid-2025 was anchoring on a number the business was moving away from. The same call had a smaller tell pointing the same way: the new semiconductor unit, guided in Q1 FY26 to a ~$10 million per quarter run rate by Q3 FY26, reported $7.2 million in Q4 FY26 — a near miss on a freshly-minted, eagerly-promoted target (Cyient Q1 FY26 and Q4 FY26 concalls). When a brand-new growth metric undershoots its first stated target, treat the next one with a discount.

The contrast: a management that guides narrow and hits it

Now the other kind. Wipro doesn't issue medium-term margin poetry; it issues a narrow, hard, sequential constant-currency revenue band every quarter and then reports against it. Here is the record, band by band.

Illustration of how a delivery record reads — not a view on the stock.

Source: Wipro Q3 FY25 through Q4 FY26 concalls; verdicts from Inve's Promise Tracker (Inve data, as of Q4 FY26).

Five hard bands, five landings inside the range — including two quarters of decline that management called in advance rather than dressing up. That is the difference that matters. Wipro's revenue didn't grow faster than Cyient's engineering business; for stretches it shrank. But the guidance was a number you could hold them to, and they held themselves to it. A band you can verify, delivered as stated through bad quarters, tells you more about a management than a glowing medium-term aspiration that keeps sliding.

One honest qualifier, because it's the steelman for distrusting all of them. Even Wipro's disciplined house misses the soft stuff: a Q3 FY25 plan to hire "10,000 to 12,000" campus freshers in FY26 was not delivered as stated (Wipro Q3 FY25 concall; Inve Promise Tracker verdict: missed). The lesson isn't "Wipro good, Cyient bad" — both face the same weak cycle. It's that the form of the guidance predicts whether you can trust it: a verifiable band beats a directional aspiration, and a hard sequential band is one you can hold management to in a way a headcount target isn't.

The concall questions that actually matter

The numbers tell you the what; the call tells you whether management's reading of them is honest — the same close reading you'd give any concall transcript, aimed here at the deal book and the margin band. The questions worth listening for:

• "What's the net-new versus renewal split in this quarter's TCV, and the average duration?" A management comfortable with its book answers directly; one that retreats to total TCV is steering you away from the split.
• "How is discretionary spend trending versus cost-takeout deals?" Discretionary (growth-oriented) work is high-margin and the first thing clients cut in a downturn; cost-takeout deals are defensive and lower-margin. A pipeline tilting toward cost-takeout is a demand signal dressed as a win.
• "What's the outlook on pricing and any deal ramp-downs or deferrals?" Deferrals and ramp-downs are how a soft quarter actually arrives — clients don't cancel, they slow. Vague answers, repeated, are themselves the answer.
• "What CC revenue and margin band are you guiding to, and what's it assuming about discretionary recovery?" Note whether it's a hard band like Wipro's −0.5% to +1.5%, with a stated assumption, or a directional "we expect to get to 15% in the medium term." Only the first is something you can hold them to.
• "Why is headcount moving the way it is?" If the answer to a shrinking headcount is "productivity and automation" rather than "demand," check whether revenue per employee is actually rising to back it up.

Read these across quarters, not once: a single call tells you about a quarter, a sequence tells you about management. The Cyient margin table above only became damning when four calls were laid end to end. That is the pattern Inve's Promise Tracker is built to catch — every revenue and margin band pinned to the quarter it was given, with a verdict as later calls confirm, revise, or drop it.

The one habit that should make you nervous

If you watch one thing, watch this: the disclosure that quietly disappears. When a firm that used to report net-new TCV, large-deal counts, a CC growth aspiration — or a specific margin exit-rate — stops, and nobody on the call presses to get it back, that silence is the signal. Cyient did exactly this with its FY25 DET exit-rate: a 13.5% Q4 FY25 exit margin was guided and reiterated, then the actual figure was never addressed on any later call analysed (Inve Promise Tracker, verdict: ghosted, Q4 FY26). It's the same instinct as testing the quality of earnings — whether reported profit is backed by cash: the metric a management stops volunteering is usually the one that has turned against it.

The related red flag is the serial-recovery guide: a target repeated across consecutive years, each time pushed one period further out — precisely the FY26 → "medium term" → "Q4 FY27" walk above. A single deferred recovery is a forecast that missed; a pattern of them is a management that can't read its own demand or won't say what it sees. Either way, weight its next "we'll get there" at a discount.

A hard limit worth stating plainly: a clean deal book and candid commentary do not immunise an IT firm against a client-side recession, a sharp rupee move, or a structural shift like clients building AI capability in-house instead of buying it. Reading management well lowers your odds of being surprised by deterioration; it does not predict the cycle.

A repeatable workflow

1. Read the real growth. Constant-currency, never reported rupee revenue — currency can flatter a weak quarter.
2. Interrogate the deal book. TCV split by net-new vs renewal, average duration, and book-to-bill.
3. Read margin a quarter early. Utilisation and attrition together move before the margin line does.
4. Check the demand barometer. Net headcount additions across two-to-three quarters.
5. Audit the commentary. Lay each quarter's CC and margin guidance side by side — the way the Cyient table above does — and compare the "recovery" timing against what actually happened. A migrating target is the finding.

For the metrics that don't surface on the concall — segment detail, contingent liabilities, related-party dealings — pair this with how to read an annual report like an analyst. Two curriculum-named quality-of-earnings checks matter most for a services firm: receivables (DSO) growing faster than revenue — a sign of relaxed credit terms propping up the top line — and the size of contingent liabilities, where tax disputes are common for IT exporters and large balances relative to the firm warrant caution. These complement the forward indicators the concall gives you.

Inve's KPI Screener lines up these metrics across IT firms — TCV, margin, utilisation where disclosed — with YoY/QoQ trends and a data-confidence flag per number, so step 2's peer comparison takes minutes. The concall summaries pull every forward commitment into one guidance table per quarter, with speaker and exact quote, so step 5 is a read rather than a re-listen.

Frequently asked questions

Reading an IT services company well is the discipline of looking past the bill that just went out to the pipeline that hasn't yet. Revenue is the past; TCV quality, utilisation, attrition, and headcount are the future, and they speak a quarter or two before the financials confirm them. A five-year owner has to believe one thing above all: that this management reads its own demand honestly and reports the same number whether it's good or bad — including the metric it would rather quietly retire, and the margin target it would rather keep walking forward. The forward indicators tell you what the business will do; the concall, read across years, tells you whether to trust the people reporting it.