Business

Know the Business

eClerx is a knowledge-heavy, productised BPM/KPO franchise that sells Fortune 2000 enterprises judgment-intensive operations — trade-lifecycle, KYC and financial-crime checks, digital-shelf and marketing operations, advanced analytics, and (now) agentic-AI orchestration — out of ~14,700 offshore seats across India, the Philippines, the US, Egypt and Peru [1]. It is not a generalist Indian IT-services shop and it is not a voice contact-centre. It is a mid-cap KPO operating in the upper band of the BPM spectrum, where the revenue is billed mostly on T&M but increasingly under proprietary platforms (Compliance Manager, Market360, QA360, Roboworx Cogniflows), and where the right-to-win is domain depth + IP + AI orchestration rather than labour arbitrage alone [2]. The business that emerges from that positioning is a disciplined cash compounder: 17% net margin, 26%+ EBITDA margin, ~28% ROE, ~85% free-cash-flow conversion, and a policy to return half of cash generated within 12–18 months [3].

The right way to underwrite this stock is not as an "Indian IT services" name. It is as a Fortune-2000 services compounder whose unit economics are quietly being re-rated by the same AI wave that threatens commodity BPO — and whose terminal value depends on whether eClerx's platforms keep enough of the AI-led productivity surplus to defend (or expand) its 17% net margin as the price-book moves from T&M-by-FTE toward outcome-based BPaaS.

The hero numbers come straight from management: FY2026 operating revenue of US$469 million (up 17.9% in USD, 22% in INR to ₹4,217 crore); EBITDA up 29% to ₹1,153 crore; net profit up 30% to ₹706 crore [4]. The Q4 print — operating EBITDA of ₹2,841 million at a 25.7% margin, net profit of ₹1,894 million at a 16.7% margin, and US$46 million of new-deal ACV — is the cleanest read on the business as it enters FY2027 [4].

The verdict, up front

The five-number summary that matters: a 17.2% net margin and 27.6% ROE on net cash of roughly ₹6,968 million [5], a free-cash-flow margin of 18.3%, ~100% FCF/net-income conversion, and a top-10 client concentration that has fallen from 63% to 59% as the company has scaled [6]. Those are not commodity-BPO numbers. They are the numbers of a specialist that has earned the right to be valued like a software-services compounder rather than a labour-arbitrage shop.

The economic engine: how eClerx actually makes money

eClerx describes itself as a "productized services company" employing over 19,000 people globally and serving "many of the world's largest organizations, including numerous Fortune 2000 companies" [7]. The five steps in the value chain — and where each one earns its keep — look like this:

The math in one line. Wages are 60% of revenue, technical sub-contractors 2%, other expenses 12% — leaving an EBITDA margin of ~26% [8]. Move the wage line by 200 basis points and the EBITDA line moves the same. Every margin conversation about eClerx begins and ends with that arithmetic. What separates the company from a pure-arbitrage peer is what sits on top of the wage line: proprietary platforms that price some of the work at outcome rates and at productivity multiples that wages alone cannot deliver.

The proof that this is a scaling franchise, not a one-trick growth blip, is in the eight-year compounding profile:

Revenue has compounded at ~21% CAGR over five years; net income has compounded at ~20%. ROE moved from the mid-teens through the FY2019–FY2020 cycle bottom to 27.6% in FY2026 — a step-change that lines up neatly with the productisation push under the current CEO. What you should not extrapolate from that history is straight-line continuation: FY2020 saw a real growth pause (5 bps revenue, –8% net income), and the ROE move from FY2021 to FY2026 was a function of both genuine operating leverage and ~₹390 crore in equity returned via the FY2025 buyback at ₹2,800/share [9].

The two-axis model: industries × services

In FY2025, management formally moved the business onto a "two-axis" reporting structure — industries (BFSI, CMT, HiTech & M&D, Fashion & Luxury, Emerging/F&A) on one axis, services (Financial Markets, Customer Operations, Digital, Analytics & Automation) on the other [2]. The headline is that they sold the SAME work differently. Both views matter for an investor: industries tell you where the demand cycle is; services tell you which platforms and people deliver the margin. The vertical view is the one management discloses with precision.

Industries — the demand-side view

Industry mix from the Q4 FY26 investor deck [10]. Two things are worth noticing. First, BFSI is shrinking as a share even as it grows absolutely — because Emerging (the Personiv F&A book selling to US SMBs) is growing faster (5.2% → 8.4% of revenue in a single year) [11]. That's a constructive mix-shift if you believe the F&A automation market has legs; it is a negative mix shift if Personiv F&A turns out to carry a structurally lower margin than the BFSI knowledge work (we cannot tell from disclosure). Second, Fashion & Luxury is the soft spot. Management has been candid that CLX Europe was sluggish in FY2026 and is expected to "return to growth in H1 FY '27, supported by new GenAI wins at a key client" [11] — a polite way of saying their luxury-brand clients are in their own cycle and that GenAI is the wedge being used to bring them back.

Services — the supply-side view

Service-line scope reproduces management's own framing of how each service line earns its keep — Financial Markets builds out of decades of BFSI ops experience, with the Fayetteville Center of Excellence for Financial Crime and Compliance now in its first full year and "emerging as a key differentiator" [12]; Customer Operations is wrapping QA360 around the existing CX delivery as an AI-led upsell [13]; Digital runs on Market360 plus the CLX Europe luxury content franchise; and Analytics & Automation has now compounded into its own US$90 million annualised book [14]. The single most important service-line data point in FY2026 is exactly that: Analytics & Automation is now a US$90 million stand-alone book — almost a fifth of consolidated revenue — and is the engine into which the company is concentrating investment.

Industry recognition — third-party validation of where eClerx sits

Everest's PEAK Matrix and Globee/BIG Innovation awards across CXM, RPA, BFS Operations, Marketing Services and F&A Outsourcing, as listed in the FY2025 annual report [15]. The honest read on these labels is that eClerx is a credible mid-tier player — "Major Contender" / "Aspirant" — in markets where Leaders/Stars are typically Genpact, TCS, Accenture and Cognizant. The company is fighting up the value chain, not down.

Customer book: deep, sticky, but concentrated

Three customer-base data points illustrate the structure investors should hold in mind:

Source: FY2025 annual report, Directors' Report, client-base disclosure [9]. Two patterns matter. The big-client base has roughly doubled in five years — clients above US$5 million went from 7 in FY2021 to 14 in FY2025 — which is the cleanest single indicator that eClerx has been mining the existing book rather than churning logos. And the US$1–5M band has been a steady 26–31 clients for five years, which suggests a stable pipeline of mid-relationship accounts that periodically graduate into the >US$5M tier.

But concentration is real and worth pricing. Top-10 clients have ranged 59%–63% of revenue [16], with the top-10 ratio coming down to 59% in Q4 FY2026 from the 63%–64% range of recent years [6]. That decline is a genuine positive — it means newer accounts (the Personiv SMB book and the Q4 FY26 agentic AI deal) are growing faster than the legacy top-10. It does not eliminate the risk; any single top-10 client roll-off or insourcing can move quarterly revenue by 200–300 basis points in a business this concentrated. The trajectory is right; the absolute level is still elevated.

The right-to-win: what is the moat, and is it real?

The honest call: eClerx has a real but narrow moat in regulated BFSI workflows where domain expertise + an outcome-priced AI-enabled platform combine to deliver tangible client savings. The single best example is Compliance Manager in KYC refresh, where management said they "have reduced the cost of refresh, let's say, by 50% because we have brought in Agentic AI" — and the work is "completely outcome basis in terms of how we are able to price it" [17]. That is the precise economic shape of a moat: a productised stack that delivers measurable client outcomes and that is priced on the outcome rather than on FTE hours. Around the BFSI/FCC core, the moat narrows quickly — the Customer Operations and Digital books look more like specialised BPM with productised tools sitting on top, and the wage line is the wage line.

Three rough proxies for the moat working: (1) top-10 concentration falling as new wins outrun roll-offs, (2) Analytics & Automation now a US$90M annualised book running as a separate platform business inside the group [14], and (3) ROE in the high-20s while equity has been actively returned via buybacks. None of those alone proves a durable moat. Together, they describe a franchise that is at least temporarily out-earning its cost of capital.

Unit economics and the operating cadence

Utilisation, attrition, top-10 concentration and DSO from the Q4 FY2026 call [5]; BPaaS and onshore mix from the Q4 FY2026 deck [18]. The three numbers an investor watches every quarter are:

• Utilisation in a 70–77% band — Q4 FY26 came in at 74%, down from an eight-quarter high of 76.5% in Q3 [5]. Above 76% the business is running hot; below 70% it is investing ahead of demand.
• Attrition running 18–25% — Q4 FY26 at 21% is well below the cycle peak of 28.8% seen in Q2 FY2025. Sustained > 25% is a quality red flag; sustained < 18% suggests demand has cooled.
• BPaaS share stuck around 18% over the last eight quarters — this is the number that has to move for the AI-led pricing thesis to play out at scale. Management is investing accordingly but has not yet shown a step-change.

Margin guidance and the operating-EBITDA shift

A reporting note matters. In Q4 FY2026, management explicitly began separating operating EBITDA (excluding other income — almost entirely treasury/FX) from total EBITDA, signalling that this will be the metric they want analysts to focus on [5]. Operating EBITDA for Q4 FY26 was 25.7%; full-year operating EBITDA was 25.6%; including other income, FY26 EBITDA was 27.3% [19]. Management's forward guidance: stay in a 24%–28% operating EBITDA band while growing revenue in the "top quartile" of the BPM peer set [20]. That is a tight band. If a margin print drops materially below 24%, that is the structural concern (likely a pricing or mix shift); above 28% would suggest the AI productivity benefits are landing.

Geography, currency, and hedging

92% of revenue originates in the US and Western Europe, with 86% billed in USD, 9% in EUR, and 5% in GBP and other [21]. The company hedges ~80% of forward USD receivables; FY2027 hedges sit at an average of ₹89.89/USD [22]. The hedge book matters for two reasons: it caps both upside and downside from INR/USD moves, and it dampens the reported revenue and margin volatility that would otherwise wash through quarterly results.

North America dominates at ~76% of revenue with the UK and Europe contributing another ~16% [26]; together the US and Western Europe make up 92% of consolidated revenue, with 86% billed in USD [16]. What this geography means for the underwriter: eClerx's business cycle is the US corporate-tech-budget cycle, not the Indian GDP cycle. Read the JPMorgan, Bank of America, Verizon, Comcast and Adobe budgets, not the Reserve Bank of India outlook.

Capital allocation: a textbook compounder's playbook

The company has functionally no debt, runs net cash, and converts ~85% of net profit to free cash flow [5]. The disclosed policy: return ~50% of cash generated to shareholders within 12–18 months if it is not required by the business [3]. In practice that means periodic tender-route buybacks as the primary mechanism, plus a modest declared dividend (₹1/share proposed for FY2026 [5]).

The FY2025 buyback of 1.375 million shares at ₹2,800 per share, settled on July 22, 2024 [9], retired ~2.9% of equity at a 30%+ premium to the closing price at announcement — a meaningful contribution to the FY2025 ROE step-up. The FY2026 follow-up announced in Q2 FY26 — a ₹300 crore tender at a floor of ₹4,500/share, with the promoter group not participating [24] — repeats the pattern but at a lower payout ratio. Management's framing of why-buyback-vs-dividend is unusually clear: "from a stock perspective, it improves the EPS… buyback is somewhat more advantageous than dividend, although it takes longer to execute" [24].

Financing cash outflow from the consolidated cash flow statement, dominated by buyback + buyback tax + dividend across the cycle. The pattern is exactly what you want to see from a cash-compounder: cash generated, cash returned, capital base shrinking, EPS growing structurally faster than net income. EPS compounded 18% over five years even as net income compounded 20% — the buyback math accounts for most of the gap, and the FY26 bonus issue confuses the picture but doesn't change it.

M&A — the small, capability-led playbook

The disclosed M&A philosophy is very clear: "M&A would either be capability focused, which is on the horizontal axis, or in a vertical focus, where we have a white space of a capability that we are strong in" [25]. Translation: small bolt-ons of platform IP or geography, not transformational scale acquisitions. The track record matches — Personiv (the F&A book for US SMBs) and CLX Europe (luxury content/digital) are the two main acquisitions of the last decade, and both are now visible segments inside the consolidated numbers without dragging down group economics.

How eClerx stacks against the BPM peer set

The peer set the indexed corpus auto-selected is not uniformly comparable. Sagility is essentially a US healthcare BPM pure-play; Hinduja Global is voice-heavy contact-centre work; Allsec/Alldigi is HR/payroll-skewed and small-cap. The peers that genuinely share eClerx's business model are Genpact, EXLService, and Firstsource — multi-vertical, knowledge-heavy BPM with significant BFSI exposure and a productised platform layer. Read the table with that hierarchy in mind.

Peer financials translated to USD at FY-end rates; figures from each peer's own latest disclosed financials (see data/competitors/<ticker>/financials/). Three takeaways:

1. Among the true peer set, eClerx is the highest-quality franchise on net margin (17.2%) and matches EXLService on ROE (27.6% vs 27.5%) — at a fraction of their scale. Genpact is the volume leader by an order of magnitude but trades premium-segment economics for scale.
2. Sagility and Hinduja Global expose the reality that "BPM" includes both well-run knowledge franchises and structurally challenged voice/contact-centre operations. Hinduja's 0.1% net margin in FY2026 is the cautionary tale; Sagility's healthcare-only vertical is a different business entirely.
3. eClerx's net margin / ROE quadrant most closely resembles Alldigi (Allsec) and EXLService, not the broader Indian BPM peer mean. The bubble chart below shows the dispersion:

eClerx sits in the high-margin, high-return upper-right quadrant alongside Alldigi and EXLService — and well clear of the voice/contact-centre cluster. That is the single best one-chart proof that this is a knowledge-heavy KPO, not a labour-arbitrage BPO.

The AI inflection: where the equity story actually lives

The next five years for eClerx are about whether the productised, agentic-AI-led layer eClerx is building can replace pure-FTE pricing fast enough to offset (or capture) the AI deflation in commodity BPM. Three specific data points frame the question:

1. Compliance Manager (KYC) is already outcome-priced, with cost-of-refresh down 50% via agentic AI [17]. This is the cleanest existing proof that eClerx can keep enough of the AI productivity surplus to defend margins. Management's framing: clients "who are confident in achieving the outcome are not keen in sharing the outcome" — i.e. some are willing to take outcome-priced contracts because the savings vs status quo are large enough that both sides win [17].

2. First large-scale agentic AI win in Q4 FY2026, with deployments planned for Q1 FY2027. Won in HiTech/M&D/Retail; management has cautioned that meaningful revenue ramp is more likely Q4 FY2027 onward [14]. That timing matters: it caps near-term Y1 revenue contribution but suggests a more durable wedge if delivery goes well.

3. The 15–20% natural project roll-off rate has not accelerated due to AI, per the CFO: "AI is another productivity tool, and it takes time to implement… we don't see that number changing much" [23]. That's the bull case to hold against the bear claim that agentic AI is about to vaporise BPM revenue.

Risks an investor must price

The single highest-priority risk is customer concentration. Even though top-10 has been moving in the right direction, two customers individually contributed >10% of group revenue in FY2025 — together ₹9,263 million of the ₹33,659 million top line [26]. That is a real tail risk that no other moat element fully neutralises.

How an intelligent investor should value this business

The right way to value this business is to treat it as a Fortune-2000-services cash compounder with a documented 50% payout policy, then cross-check against an EBITDA multiple of a true peer (EXLService, Genpact). The two anchor questions to answer first:

1. What FCF margin do you believe is sustainable across an AI transition? Bull case: ~18% holds because productised IP captures the productivity surplus. Bear case: drifts to ~13% (closer to Firstsource) as commodity BPM gets re-priced and Personiv F&A dilutes mix.
2. What sustainable ROE do you believe? Bull case: 25%+ (today's level), implying re-rate optionality. Bear case: drifts to 18% as AI deflation eats the margin and concentration eventually forces a defensive pricing posture.

The honest answer: eClerx is high-quality enough that it deserves a premium to the Indian BPM mean, but not high-quality enough that you should pay an unconstrained software-services multiple. The right framing is: a P/FCF in the high-teens to low-20s is justifiable today given the visible cash-return discipline, low capital intensity, and the early proof of outcome-based pricing in Compliance Manager. Anything materially above that is paying for the agentic-AI story that has not yet shown up in the numbers — a story that may yet play out, but for which the current pipeline (one large win in Q4 FY26, modest BPaaS share) is not sufficient on its own.

The five things to remember

References

1. eClerx Services Limited — FY2025 Annual Report, MD&A Infrastructure section (seat capacity, geographies) — p.92
2. eClerx Services Limited — FY2025 Annual Report, Chairman's Message (two-axis model, capital return) — p.13
3. eClerx Services Limited — Q2 FY2026 Earnings Conference Call Transcript, capital allocation policy — p.14
4. eClerx Services Limited — Q4 FY2026 Earnings Conference Call Transcript, CEO opening remarks (FY26 financials, Q4 EBITDA, ACV) — p.2
5. eClerx Services Limited — Q4 FY2026 Earnings Conference Call Transcript, CFO commentary (utilisation, attrition, DSO, FCF, dividend, operating-EBITDA shift) — p.4
6. eClerx Services Limited — Q4 FY2026 Earnings Conference Call Transcript, CEO on top-10 client concentration — p.3
7. eClerx Services Limited — FY2025 Annual Report, Who We Are (productized services, 19,000 employees, Fortune 2000) — p.5
8. eClerx Services Limited — FY2025 Annual Report, MD&A Results of Operations (cost stack, EBITDA margin) — p.96
9. eClerx Services Limited — FY2025 Annual Report, Directors' Report (client-base segmentation, FY25 buyback at INR 2,800) — p.35
10. eClerx Services Limited — Q4 FY2026 Investor Presentation, vertical revenue mix — p.6
11. eClerx Services Limited — Q4 FY2026 Earnings Conference Call Transcript, CEO on vertical readout (Personiv F&A, Fashion & Luxury, CLX) — p.3
12. eClerx Services Limited — FY2025 Annual Report, MD&A Business Performance (BFSI, Fayetteville FCC CoE) — p.91
13. eClerx Services Limited — FY2025 Annual Report, MD&A Customer Operations (QA360 launch, CX) — p.93
14. eClerx Services Limited — Q4 FY2026 Earnings Conference Call Transcript, Analytics & Automation $90M book, first large-scale agentic AI win — p.2
15. eClerx Services Limited — FY2025 Annual Report, Awards and Recognition (Everest PEAK Matrix, Globee, BIG Innovation) — p.8
16. eClerx Services Limited — FY2025 Annual Report, MD&A Opportunities, Threats, Risks (client concentration, geography, currency) — p.94
17. eClerx Services Limited — Q4 FY2026 Earnings Conference Call Transcript, CEO on Compliance Manager / outcome-based pricing — p.6
18. eClerx Services Limited — Q4 FY2026 Investor Presentation, Key Revenue Metrics (BPaaS, onshore mix) — p.9
19. eClerx Services Limited — Q4 FY2026 Investor Presentation, KPIs (operating EBITDA, EBITDA, EBIT, PAT margins) — p.2
20. eClerx Services Limited — Q4 FY2026 Earnings Conference Call Transcript, FY27 outlook (top-quartile growth, 24-28% EBITDA) — p.5
21. eClerx Services Limited — FY2025 Annual Report, MD&A Opportunities, Threats, Risks (currency concentration) — p.94
22. eClerx Services Limited — Q4 FY2026 Investor Presentation, Hedge Updates — p.10
23. eClerx Services Limited — Q4 FY2026 Earnings Conference Call Transcript, CFO on roll-offs and AI productivity — p.7
24. eClerx Services Limited — Q2 FY2026 Earnings Conference Call Transcript, INR 300 crore buyback announcement — p.4
25. eClerx Services Limited — Q2 FY2026 Earnings Conference Call Transcript, M&A philosophy (capability vs vertical) — p.11
26. eClerx Services Limited — FY2025 Annual Report, Consolidated Notes, Segment Information (geographic revenue split, two customers >10%) — p.244