Why retention is the metric founders should pitch — and almost no one does.

Anand's most direct line in the conversation: "retention is probably ninety-nine percent of business value, customer acquisition is something that of course you do — but without retention it's just zero." The framing is unsentimental and correct. A startup that has to double revenue every year, but whose current customer base contributes nothing to next year, is riding what he calls "a tiger" — every quarter, the entire revenue base has to be re-acquired before any growth can be claimed. That is not a business; it is a CAC treadmill priced at venture multiples.

The Gillette example does the work. A customer acquired at twenty pays the company across roughly fifty years, migrates from two blades to three to five to six with vibration, and brings a younger brother and an eventual son into the same brand. The revenue grows without acquisition. That kind of compounding is not a marketing outcome; it is the entire reason large consumer companies are worth what they are worth.

The phrase belongs to Anand and travels well beyond consumer brands. A startup that promises three-times revenue growth every year, and whose current customers churn before the year ends, has implicitly committed to acquiring not just the growth — but the entire base, again, at next year's CPM. The 2018-2021 cohort of Indian D2C brands ran this trade openly. The 2022 funding winter wrote the receipts. A long list of brands that looked like category creators when CAC was subsidised turned out to be re-acquisition machines when it was not.

The diagnostic is simple. Project current customer revenue forward, holding new acquisition at zero. If the line collapses inside two quarters, the business is acquisition-led, not retention-led. The valuation multiple should reflect that — and in 2024 onward, it finally does.

Anand returns repeatedly to the razor-blade business as the canonical retention model. A first purchase at twenty, two blades. By thirty, three blades. By forty, five blades. By the late fifties, a vibration-mode six-blade. Revenue per customer climbs through the product ladder; the company spends almost nothing on re-acquiring them. The compounding is not in marketing — it is in product upgrade paths designed to fit the same customer at every life stage.

The implication for an Indian D2C brand is sharp. The hero SKU is the entry-point, not the entire catalogue. The roadmap has to include the cohort's next purchase, and the one after that. Perfora — the oral-care brand from Shark Tank Season 2 — went from five SKUs at month six to 100-plus SKUs today, exactly because the founders designed the brand as a household over time, not a toothbrush at a time.

Anand extends the Gillette frame one step further: a customer is also a household. "My son when he grows up will also probably use Gillette," he says. "My younger brother started using it because I was." Tanishq jewellery survives on the same loop — a mother teaches a daughter-in-law, a daughter learns from a sister. The lifetime value isn't fifty years; it is generational.

Almost no early-stage deck prices this in. LTV models stop at year three or five because that is what the data supports. The point is not to fake longer numbers — it is to design the brand experience so the next-generation handoff is built in. Packaging that survives a kitchen-shelf decade, brand voice that crosses age cohorts, a refer-a-relative motion that isn't a Q4 campaign but a year-round system. The brands that get this right end up with multi-decade share that no challenger can buy back.

Before any cohort curve is drawn, Anand wants to know the replenishment clock of the category. A razor blade replenishes in eight to twelve weeks; a toothpaste in eight to ten; a mattress in five years; a piece of luggage in seven. A demo-day startup with six to twelve months of operating history can demonstrate retention in the blade category — and cannot, structurally, in the mattress category. Pretending otherwise wastes both sides' time.

The harder discipline is for mattress-clock founders to design retention proxies: NPS, usage frequency, comfort-after-purchase, regret signals, did-someone-in-your-family-buy-one. These proxies are not the same as a repeat number, but they are honest leading indicators for one — and they are the only fair audit at seed for a long-replenishment business.

Anand all but writes the demo-day exam paper for the audience. On a base of 500 to 1,000 customers — past that, more customers do not improve the diagnostic until product-market fit is in print — he wants four readings. Net Promoter Score after purchase. Usage after purchase. Comfort after purchase. Repeat behaviour in buying. He says it almost as a single phrase, and the phrase is the cheat-sheet: "that's the cheat sheet guys, I hope you guys prepare for it and wow on that day."

The order matters. NPS is the cheapest signal and reads in a week. Usage is what produces the second purchase. Comfort is the absence of regret. Repeat is the only one that is real revenue. Most founder decks lead on the last and skip the first three; the diligence pattern Anand describes inverts that — and is the right way to read product-market fit at seed.

A subtle line worth catching. Anand says, almost in passing, that the seed-stage retention read does not get better as the sample crosses a thousand. "It's irrelevant whether you have a thousand customers or ten thousand." A brand isn't going to IPO with ten thousand customers; the value of the cohort read at seed is in the shape of the curve, not the absolute size. Beyond a thousand the data smooths, but it doesn't reveal anything new.

The implication for founders is to stop using "we'll show you retention once we hit X-thousand users" as the answer when X is over a thousand. The data is already there. The honest thing is to publish what the first thousand cohort did, accept the noise, and let the curve speak.

Nyasha's example does a lot of work. Perfora, the Shark Tank Season 2 oral-care brand, started measuring retention at five to six months in, when it had only five or six SKUs. Today it is at 100-plus SKUs across the oral-care category — and the retention discipline started at the front of the journey, not bolted on after a Series A. The reason oral care reads cleanly: toothpaste is replenished every six to ten weeks, toothbrushes every two to three months, and a successful brand graduates the customer into mouthwash, floss, and adjacent SKUs without ever re-acquiring them.

The category sets the curriculum. A toothpaste brand that builds retention discipline early can scale a hundred SKUs against the same household. A toothpaste brand that builds growth discipline first and retention later usually finds the retention reading too late — when the second-purchase rate is already structural, not coachable.

Nyasha's clarification on the 300-startup cohort is structural. WebEngage's program has worked with D2C, fintech, social, dating, content — and in each category, retention is the load-bearing metric, just defined differently. A fintech app measures session frequency on the right cadence: daily for payments, weekly for trading, monthly for lending. A social app measures viral loop — without organic re-entry, the platform is in a paid-acquisition death spiral because new social users are the most expensive in Indian digital advertising. A content app measures session depth and minutes per user.

The mistake is to think retention discipline is a consumer-brand idea. The opposite is true: in fintech the cost of acquiring a fully-KYC-verified user is two-to-three thousand rupees on average; in dating the cost of a verified active user runs higher still. The arithmetic only closes if those users come back. Channel infrastructure — email, SMS via DLT, push (post-iOS-ATT damaged but still useful for Android), WhatsApp Business API — is what produces the come-back. Retention is the metric; channels are the levers.

The single most consequential platform change for mobile retention in the last five years was Apple's App Tracking Transparency framework, shipped with iOS 14.5 in April 2021. The ATT prompt — "Ask App Not to Track" — collapsed deterministic attribution for Meta and Google ads, killed the IDFA-based retargeting model that propped up CPI economics, and structurally damaged push-notification opt-in rates on new installs. Adjust and AppsFlyer benchmarks since then put D30 mobile retention in the 6-10% range industry-wide; D90 falls to single digits for most categories outside fintech and OTT.

The retention practitioner's response — visible in WebEngage's own product evolution and at MoEngage, Braze, CleverTap — has been to push the work earlier in the lifecycle: in-app onboarding, day-one habit triggers, WhatsApp Business API as the de-facto re-engagement channel for India because it does not depend on push opt-ins, and email lifecycle that assumes Android-first delivery. The brands still running their 2020 push playbook on 2026 cohorts are quietly losing the second-purchase economics.

An unspoken constant across the retention conversation: the single channel that has made the largest difference to Indian D2C and fintech re-engagement since 2022 is WhatsApp Business API. Open rates run in the 80-90% range against email's 15-20%, click-through is roughly ten-times email's, and the channel is messaging-app native — which means it lives where the Indian customer already is. The structural advantage compounds in tier-2 and tier-3, where vernacular WhatsApp messages out-perform English email almost universally.

The trade-off is friction. WhatsApp Business API templates require Meta approval, the per-message cost is non-zero, and over-sending burns the channel faster than email — once a number is muted or reported, it is structurally harder to recover than an email address. The brands that have built durable WhatsApp lifecycle automation have learned to use the channel for moments of real customer value — order updates, refill nudges, abandoned-cart at the right minute — and to let email and SMS carry the longer-tail communication. The discipline isn't volume; it is signal-to-noise.

Since the Telecom Regulatory Authority of India's 2019 commercial-communications regulations took effect through 2020-2021, every transactional and promotional SMS in India runs on the Distributed Ledger Technology (DLT) registration framework. Senders register IDs, templates, and consent on operator blockchains; unregistered traffic is filtered out. The friction killed grey-market spam-SMS overnight and forced legitimate brands into a slower, more deliberate cadence. The unintended consequence is that SMS has stayed valuable in India where it has commoditised elsewhere — delivery rates remain in the 95%-plus band, the channel is regulated into a permission discipline.

For early-stage founders, DLT compliance is a one-time set-up cost that almost no one wants to do and everyone has to. The compliance work — sender-ID registration, template approval, consent capture — is the new minimum viable product on the channel layer, and the retention software stack that does it cleanly out of the box is the one founders end up paying for.

The retention stack underneath WebEngage and its peers — MoEngage, Braze, CleverTap, Iterable, Customer.io — is a Customer Data Platform (CDP) layer that unifies events from web, app, payments, support, and offline. The segmentation engine runs queries against the CDP: "customers who bought once, opened a WhatsApp message in the last 14 days, and viewed product page X but didn't repeat" becomes a journey trigger. The journey builder fires a sequence: a WhatsApp nudge at hour two, an email at day three, an SMS at day seven if still dark.

The maturity ladder for an early-stage brand is honest. Year one: events into a CDP, basic segmentation, lifecycle email. Year two: WhatsApp Business API, multi-channel orchestration, win-back automations. Year three: predictive models, ML-driven send-time, segment-level pricing. The brands that skip the foundation and start at year three usually find that their cohorts are too small for the models to be honest — and the models become a way of optimising noise.

One of the under-discussed audits. By Series A, a serious consumer or fintech company has a named retention or lifecycle marketing function — usually one to three people reporting to the head of growth, with their own quarterly cohort goals. Companies without that function are running retention as a side-task of performance marketing, which means it gets the leftover budget and the leftover attention. The cohort curves usually show it.

The diligence question for an investor: "show me your org chart, who owns the D30 number?" If the answer is "the growth team also does it," the retention motion is not yet a function; it is an aspiration. The brands that grew past 100 crore in revenue almost all installed the function at 10 crore — early enough that the cohort discipline coloured the product roadmap, not just the marketing roadmap.

Anand's quietest argument is the most important. The pre-2022 cohort of Indian founders, he says with some bluntness, included "people who wanted to escape their jobs, people who wanted to play startup-startup, and a few gold-diggers who wanted to raise money and get rich quick." The post-2022 cohort, he insists, is structurally different — founders who run twenty customer calls a day, share their personal phone number in welcome messages, lead with "did you try my product" before "do you want to fund me."

The retention metric, in this frame, is the diagnostic for missionary intent. A mercenary founder optimises the funnel because the funnel is what gets headline-printed. A missionary founder optimises the cohort because the cohort is what gets the customer back. Anand thinks the pendulum has swung; the role models have shifted; the bar has risen. The conversation reads, in places, like a quiet manifesto for the post-winter cohort.

Anand's anecdote about Carwale — the auto-marketplace where he served on the board — is the conversation's clearest culture story. He went in once to buy a car himself, asked the CEO who to talk to, and was told he could ask anyone. "Everybody has test-driven every car that comes to the market. Being a car lover is the first quality everybody has to pass." Hiring filtered for passion; passion produced metrics literacy across the org; metrics literacy meant retention conversations weren't owned by one person.

The transferable point is operational. Retention is not a metric a single department owns; it is a culture installed at hiring. Anand's broader recommendation — share every metric with every employee, even those not responsible for it, because context produces suggestions that hierarchy doesn't — extends the same principle. A company in which only the founder sees the cohort curve is a company in which only the founder fixes the cohort curve.

Anand drops the line almost in passing — "whoever funds you is your son of a... it doesn't matter. Beauty lies in the eyes of the beholder." — and the line carries the season's investor mood. Founders should stop ranking funds; should stop celebrating fundraises as the win; should stop measuring themselves against rounds. A fund that nobody else wanted to back the category often turns into a brand without competition, because the rest of the market said no. The fund that everyone wants to back is the one that produces ten competitors inside eighteen months.

The structural extension is the conversation's most quoted line elsewhere: "funding follows success; success does not follow funding." Retention is the cleanest measure of that success because it is the cheapest signal that customers chose, paid, came back, and brought one more household member. A founder who can show that — to a thousand customers — has more standing in the room than a founder pitching the next round of a company that hasn't yet proved second purchase.