IGP's Priyesh on sending fewer WhatsApp messages and earning more from each one

Half product, half service

IGP started with NRIs sending Rakhis to their families in India, then moved into birthdays and anniversaries through cakes and flowers. The team saw a wide experience gap in these categories and was among the first to promise that what you see is what you get.

Priyesh describes the business as 50% product and 50% service. IGP gets products made to its specification, trains the people who make the bouquets and cakes, and delivers them itself from its Customer Fulfillment Centers, the first of which opened in 2018. Across eight facilities, it knows in real time which florist made which bouquet, how customers rated it and how long it took. The same data tells marketing what each customer is fussy about.

Credit only for what marketing adds

About 35% of IGP's consumers each year are new, and close to two thirds are repeat buyers. Upwards of 75% to 80% of spend goes to Google and Meta. Among returning customers, roughly 70% come through CRM and free channels and about 15% through Google and Meta, which means IGP still pays to bring back its own customers.

Priyesh uses last touch attribution for day to day operations but judges spend on incremental return. Take 100 customers likely to convert today, message 50 and hold back the other 50. If 20 convert in one group and 10 in the other, marketing gets credit for 10.

Fewer WhatsApp messages, better returns

Gifting is a low frequency business, with 6 or 8 occasions per customer a year, so IGP stopped mass blasting early. When WhatsApp raised its prices, that discipline let the brand treat the change as an opportunity.

Last year IGP sent only about 40% of the messages it had sent the year before, a cut of almost 60%, and earned almost 25% to 30% higher revenue from the same customer base. Revenue per message is up 3X to 4X. Assuming WhatsApp's cost rose by around 30% to 40%, he puts the ROI gain at perhaps 30% to 40%.

The method is cohorting. IGP looks at life events, at what people do during their journey such as abandoned carts, and at timing. Combined, these bring the send count down to a couple of percentage points of the base. IGP runs 100 plus cohorts today and expects that to reach 1,000 within about a year.

The mistake he sees D2C brands make is treating WhatsApp as just another retention channel, blasting the same 20% off message and celebrating a 90% to 95% delivery rate. Revenue does not compound with send volume. Unsubscribes and blocks do.

"If you don't have anything meaningful to say, better not knock the door," he says.

Why a little friction helps

Gifting is a high anxiety purchase. Two years ago, every third or fourth IGP customer called the call centre after ordering to check whether it would arrive on time. Priyesh treats WhatsApp like a shop counter, where buyers can ask questions before and after they pay.

A few months back, IGP gave its A+ customers a dedicated number that handles their questions and every step of the post order journey, down to a real image of the bouquet going out. Orders per customer in this cohort are up almost 28% to 30% on the same period last year, with no additional marketing cost.

On other channels, he sees email moving to the top of the funnel, with AMP and interactive formats keeping engagement roughly flat despite falling delivery rates. App push notifications, along with WhatsApp, are among IGP's fastest growing channels.

What AI agents got right and wrong

AI has made the analysis after big occasions like Valentine's Day and Rakhi far more efficient. IGP also uses it to split predefined cohorts into smaller sub segments and suggest what to talk about, whether flowers, a personalised product or a sipper. For the best optimised cohorts, CTR moved from sub 3% to close to 4.5%, an increase of almost 50%, with a 25% to 30% lift at an aggregate level.

Two experiments fell short. Content created fully by AI over optimised for clicks and drifted from what the brand stands for. AI also has not yet surfaced hidden insights beyond the analysis a team already knows to run. For now, Priyesh says, humans remain essential for the brand voice, the emotion and the strategy.

From campaign operators to campaign strategists

Priyesh does not expect agents to replace 40% to 50% of his team in three years. The roles will change, and the same team could provide 10X the service. His job as a leader is to give the team confidence to experiment in grey areas, knowing half the experiments are bound to fail, and to scale the ones that work.

"Execution will go away, but value will be compounded only by the people who are using them rightly or more strategically," he says.