Business
Figures converted from INR at historical FX rates — see data/company.json.fx_rates. Ratios, margins, and multiples are unitless and unchanged.
Bottom Line
eClerx is a margin-premium specialist KPO, not an IT services firm — it sells domain-trained Indian and offshore analysts on a US-dollar T&M rate card, embedded inside the operating teams of Fortune 2000 banks, telcos and retailers. The right way to think about it: a 26% operating-EBITDA, 29% ROE workflow specialist with capital-markets compliance and customer operations as its two profit anchors, trading at roughly 18× earnings — a multiple that sits below faster-growing peers (Coforge, LatentView) and below larger but lower-margin peers (FSL), in part because the market is uncertain whether agentic AI is a tailwind or a deflator. The single most underestimated point is earnings quality: FY26 generated $81m of free cash flow with FCF/PAT above 100% and OCF/EBITDA at a five-year high, on essentially zero net debt.
Revenue FY26 ($M)
Op EBITDA Margin
ROCE
Free Cash Flow ($M)
1. How This Business Actually Works
eClerx is, at its core, a labor-arbitrage business with a domain-expertise premium and a small but rising IP layer on top. A US or UK bank pays a USD billing rate for an Indian or Filipino analyst who maintains its KYC files, reconciles its trades, or runs its digital-shelf merchandising. eClerx's job is to deliver that hour for less than the billing rate — and to make sure the work is good enough that the bank's operations head will not switch vendors for a 5% discount.
The economics line up as four levers, in this order:
Sitting on top of the labor model are three productized IP assets that the market often overlooks: Compliance Manager (KYC / financial-crime workflow), Market360 (digital-shelf analytics for retail and CPG), and Roboworx / Cogniflows (agentic-AI orchestrator now used by 15,000+ internal staff and being client-deployed). These products do not yet generate disclosed revenue lines, but they do two important things: they let eClerx win competitive RFPs against larger integrators and they enable outcome-based pricing in pockets — for example, "we cut your KYC refresh cost 50%" — which decouples revenue from headcount.
The bargaining-power structure is asymmetric. Clients are concentrated (top-10 = 59% of revenue), procurement-led, and have outside options. eClerx's defense is embedded tenure — once an analyst pod has been running KYC for a global bank for three years, ripping it out involves regulatory risk the client will not casually take.
The simplest mental model: for every $100 of revenue, $74 leaves as operating cost, $26 stays as operating EBITDA, $17 falls to PAT, and $18 lands as free cash flow. That FCF/sales conversion of ~18% — almost the same as PAT — is what separates this from a generic offshore BPO.
2. The Playing Field
eClerx does not have a single same-shape competitor. It sits in a margin sweet-spot above generalist BPM (FSL, Datamatics, Genpact) and below high-growth specialist analytics (LatentView, Coforge), with revenue scale roughly one-tenth of Genpact and one-fourth of Coforge.
Three things the peer set reveals once you stare at it:
The first is that eClerx and LatentView are the only two with PAT margins above 15% — both make their living on specialization rather than scale. LatentView is paid 30× for that quality, eClerx 18×. The gap reflects growth (LatentView 25% vs eClerx 22%) and perceived AI-disruption risk, not earnings power.
The second is that Genpact, despite being 10× eClerx by revenue, earns roughly half the net margin (10.9% vs 17.2%). Scale in BPM is not the same as scale in software — past a certain point, the additional revenue comes from lower-tier accounts that pull blended margin down. eClerx's "stay small, stay specialist" strategy is economically defensible.
The third is that Coforge looks structurally different — it trades at 36× because of digital IT-services growth that has little to do with eClerx's KPO economics. It is in the peer table by convention, not by economic substitution.
"Good" in this industry means three things together: domain-specialist EBITDA margin (24%+), credible top-line growth (15%+), and OCF/EBITDA conversion above 70%. By that test, eClerx is the only stock in the table that passes all three. FSL fails on margin; Datamatics on growth; LatentView on conversion; Coforge on conversion-to-FCF intensity.
3. Is This Business Cyclical?
eClerx is mildly cyclical with two distinct cycle exposures: enterprise discretionary BFSI spending (capital-markets ops) and US telco/cable capex (customer operations). The cycle hits the income statement in a predictable sequence — ACV bookings → utilization → realized pricing → headcount — over two to three quarters. Margins move in a narrower band than revenue because management protects the bench and clients value continuity in regulated work.
The chart shows the textbook KPO cycle. Growth swings 0%–38% across a decade; operating margin moves within a tighter 22%–31% band. FY18–19 was the structural low (margin from 35% → 22%) — that was an idiosyncratic event tied to a large telco client and CMT margin compression, not a macro recession. The COVID year (FY20) showed almost no revenue impact but a meaningful margin dent; the post-2022 BFSI slowdown caused a one-year growth pause in FY24, which the FY26 result has now lapped.
Three cycle channels are worth distinguishing because they behave differently:
Cycle exposure to watch right now: the proposed US FCC NPRM on offshore call-center restrictions. If finalized, it could force CMT scope back onshore or to nearshore Manila / Mexico / Cairo. eClerx has been pre-positioning (Egypt MoU Nov-25, Philippines expansion) but CMT is the single largest vertical — a regulatory shock here is the most non-cyclical, hard-to-hedge risk in the model.
4. The Metrics That Actually Matter
Five metrics carry almost all the signal for this business. The income statement is a lagging indicator; the operating disclosures are where the next four quarters get decided. Standard ratios (P/E, ROE) tell you what the stock is — not what is happening to the business.
The heatmap shows the FY24 trough — ACV slowed, margin held but did not expand, concentration stayed elevated — and the FY26 broad recovery across every operational axis simultaneously. The recovery is what makes FY26 a credible new baseline rather than a one-off. The single weakest cell is FY25 operating margin (came in at the bottom of the guidance band) and the single biggest improvement is FY26 OCF/EBITDA at 75%, the highest in five years.
The hierarchy that matters: New ACV growth → utilization → margin → PAT → cash. If a young analyst checked just two numbers each quarter, they should be the ACV figure (USD millions) and the OCF/EBITDA ratio. P/E is what the stock costs; ACV is what the business is doing.
5. What Is This Business Worth?
The right valuation lens for eClerx is earnings power × cash conversion through the cycle, capitalized at a multiple that reflects specialist-niche moat plus mid-cycle growth. This is not a sum-of-the-parts story — there are no listed subsidiaries, no investment book of strategic stakes, no distinct businesses with materially different economics. Customer Operations, Financial Markets, and Digital share the same revenue mechanic, the same cost structure, and the same labor pool. Trying to value them separately would invent precision the company itself does not disclose.
Value, therefore, comes down to the durability of three numbers: top-line growth, the 24–28% operating-EBITDA band, and free-cash-flow conversion.
The peer table in Section 2 is also the valuation matrix. At 18× earnings, eClerx trades at a discount to LatentView (30×), Coforge (36×) and FSL (23×), despite higher ROE than all three and a more cash-generative model than two of them. The discount is real and has two probable explanations: lower top-line growth than LatentView and lingering investor uncertainty around AI deflation. A re-rating to 22–25× is justified if FY27 delivers (i) operating margin held above 25%, (ii) ACV growth maintained at 20%+, and (iii) the first quantified disclosure of agentic-AI revenue.
FCF/PAT above 100% in two consecutive years is the single best evidence that the earnings are real and that the buyback-led capital-allocation policy is sustainable. The bear case requires explaining why this conversion rate would mean-revert downward — receivables / DSO is the only credible reason, and Q4 FY26 DSO of 81 days, while up from the ~60-day historical norm, has stabilized.
The valuation framing in one sentence: eClerx is priced as if 22% revenue growth and 26% operating margin are not durable; the buy case is the bet that they are, helped rather than hurt by AI productization. The lens to underwrite is P/FCF and earnings yield through cycle — not SOTP, not multiples.
6. What I'd Tell a Young Analyst
Five things to internalize, in order of how often they matter:
Watch ACV bookings before revenue. New ACV ($46m in Q4 FY26, ~$170m for the year) leads revenue by two to three quarters. If you spot the ACV figure deteriorate while the press release celebrates the latest revenue print, you are looking at a vendor that will disappoint next quarter. This is the one number management has trained the market to focus on, and it is right to do so.
Treat top-10 client concentration as a quality signal in both directions. Falling concentration during growing revenue (the FY26 pattern — top-10 from 63–64% to 59%) is healthy diversification. Falling concentration during flat revenue is a warning — it means clients at the bottom of the top-10 are leaving. Always read the trajectory in the context of growth.
The AI question is the most important single thesis variable. Two opposite scenarios are live: (1) agentic AI deflates BPM unit economics, eClerx becomes a 15% margin business in 2028; (2) eClerx successfully sells outcome-priced AI-enabled workflows (the first large win was in Q4 FY26), and margin holds at 26%+ with revenue growth helped by capturing the savings as price. Mid-2027 will be the inflection — when AI revenue first gets quantified, or when it does not.
Distrust the 91% T&M / 9% fixed-price split as the future state. Outcome-based pricing on AI-enabled contracts will scramble that mix. The KPI to watch is whether average revenue per employee (revenue ÷ headcount) starts rising materially. If it does, the productized model is working. If headcount keeps growing in line with revenue, eClerx is still selling FTEs, just under new packaging.
The capital-allocation question is where you might be surprised. Management has stated buybacks are the preferred channel and the ESOP trust has been a quiet but significant buyer ($27m in FY26 alone). This benefits per-share metrics without showing up as a headline buyback. Combined with bonus issue, low dividend ($0.01/share), and zero net debt, the math says shareholders are being returned to without it being legible from headline disclosure. Track diluted share count carefully — that is where the truth lives.
What would change the thesis: a final-rule US FCC NPRM that forces CMT scope onshore, a Capgemini-WNS aggressive BFSI push that wins eClerx's largest client, or a two-quarter sequence of falling ACV alongside rising DSO and top-10 concentration. None of those is present today; all of them are observable in the quarterly investor pack before they hit the income statement.
The market may be underestimating the durability of eClerx's earnings quality — FCF/PAT above 100% for two straight years, 75% OCF/EBITDA conversion, ROCE rebounding to 35%, and a buyback-friendly capital structure. The market may be overestimating how quickly AI will compress mid-tier BPM margins for embedded specialists. Both calls turn on the same observable: whether agentic-AI deployments show up as new revenue or as price concessions over the next 18 months.