You're pouring money into ads to get new customers. Meanwhile the customers you already have buy once and vanish. At some point that math breaks. Here's how to know when to shift budget toward keeping the people you've already paid for.
The signal is CAC rising against repeat rate falling
When customer acquisition cost climbs while your repeat purchase rate sits flat or drops, you're paying more for customers who are worth less. That's the moment. If it costs you Rs.800 to acquire someone who buys once and never returns, you're renting customers, not building a business.
Put numbers on it so the trigger is visible. Say your first-order contribution is Rs.600 and acquisition sits at Rs.750. You're Rs.150 short on every new customer, which only works if the average customer buys again. But your repeat rate is 12%, and say a repeat order nets another Rs.500 of contribution. Expected contribution per acquired customer is 600 plus 0.12 times 500, Rs.660, against Rs.750 of CAC. You lose Rs.90 per acquisition and the ads look great on first-order dashboards. That gap between what a customer costs and what a customer returns is where the whole decision lives.
Retention deserves more budget when the cost of a new customer starts to outrun what that customer actually pays you over their life. That's the LTGP to CAC ratio. When it slides toward 1 to 1, acquisition is eating the whole business and retention is the lever that fixes it.
When retention wins the budget
Shift weight toward retention when your second-purchase rate is weak. If most customers never buy twice, no amount of ad spend fixes it, you're just refilling a leaking bucket. Fix the leak first, post-purchase flows, a reason to come back, a product experience worth repeating, then pour acquisition back in.
Leaky bucket arithmetic makes the point brutal. Two stores spend Rs.10,00,000 a year on acquisition. Store A retains 35% of customers, so every 1,000 acquired becomes 1,350 orders of volume over time. Store B retains 5%, the same acquisition money returns 1,050 orders. Similar stores, similar product, one earns nearly a third more from the identical spend. That's not an ads difference. That's what the first order is worth after it ends and the growth compounds from there.
Retention also deserves real budget when acquisition gets expensive. In a crowded market where everyone is bidding on the same cold traffic, an email flow that drives a second purchase from people who already trust you is often the cheapest revenue in the business. Nobody else can bid on your customers.
"Shouldn't I just focus on new customers until I'm bigger?"
The most common objection, so here's the straight answer. Early on, yes, acquisition dominates, because you have no base to retain and every rupee buys learning about who buys. Retention spend on fifty customers is mostly perfecting craft nobody's around to receive.
Objection handled, the honest version: even early, don't buy new customers on the assumption they'll repeat later. Cheap flows exist from day one, a post-purchase sequence and a replenishment reminder cost almost nothing and decide a pure repeat rate early. Track second-purchase rate from your first hundred customers, because that number is the evidence that decides everything later.
What surprises founders who finally check: the same acquisition spend works twice as well on a store whose first cohort comes back. The reason won't shock you. A returning customer already trusts the store, skips the consideration that makes cold ads expensive, and those conversions show up in reports as cheap revenue no competitor can bid against.
The balance that actually works
The mistake is treating retention as free. It isn't. It takes real spend on flows, offers, loyalty, support. Underfund it and your "retention strategy" is just an abandoned-cart email and a hope. Budgeted retention gets produced monthly, with an owner, even if the owner is you and the tool stack is three automated flows and a spreadsheet.
As your repeat rate builds, retention should take a bigger share, usually somewhere in the range where a solid brand might run 70/30 acquisition to retention early, moving toward 50/50 as the customer base compounds. Numbers shift with margins and repeat behavior, so treat those as shapes and let your own ratios tension it out. No percentages, no drama; read the two ratios and fund whichever lever returns more of what you need.
The honest test
Ask what a returning customer is worth versus what a new one costs, and that worth is your LTV, their lifetime value. If retaining is dramatically cheaper than acquiring, and your repeat rate is soft, put money there. Acquisition grows the top of the funnel. Retention decides whether any of that growth turns into a business or just churns out the bottom.
The move, in order
Sequence matters here, so take it as a running order rather than a menu. One, count your second-purchase rate by cohort. That number decides everything downstream, and most stores have never computed it. Two, if repeat is soft, fix retention first and hold acquisition flat while the fixes land, because pouring spend through a leaking funnel is how quarters disappear. Three, if repeat is solid and CAC is stable, keep funding acquisition and add retention spend gradually to compound what acquisition already built. Four, recheck the LTGP to CAC ratio every quarter, because the right share moves as the market moves, and a ratio that was healthy at 3.5 can quietly sit at 2 a year later without anyone changing a setting.
And one discipline buys more than any split: track retention spend against the revenue it produces, the same way you track ad spend. It keeps the retention budget honest, stops it becoming a cost bucket nobody defends, and gives you the evidence to argue for more of it when the ratio says the moment has come.
I'm Hirdesh Matai, founder of HDM GRO-ads, the science of e-commerce advertising.