September 24, 2025
Anshuk Aggarwal skipped every marketing course he could at ISB. A decade later he runs AdYogi, and his argument is that most brands are still feeding their ad platforms the wrong number.
Anshuk went into consulting after ISB, took a year off to build a data business, and found that brands didn’t want data. They wanted someone to use it and bring them customers. That nudge turned into AdYogi, which today has 12 to 13 brands in its portfolio doing over ₹100 crore a year on their D2C websites alone.
He spoke to Ragini about the ₹5,000 campaign that started it, why cash on delivery is a trust problem rather than a payments problem, and the signal he thinks every brand should be sending back to Meta.
At ISB I would purposefully skip the marketing courses. I thought marketing was faff and very generic, and I did finance because it would give me good grades. Ten years later I have invested a decade of my life in digital marketing.
Our first client came through a Google startup boot camp connection: a boutique brand called Fancy Pants. The budget was ₹5,000 for the whole month, and both of us, Rohin Mittal and me, sat on the dashboards optimising it. I still don’t know what we were doing for a month, but we got her 18 sales. If we hadn’t got those, I don’t think we would have continued down the e-commerce path at all.
If you are just starting out, you need someone to help with pricing, cataloguing, logistics, the payment gateway, setting up your Facebook and Google pages. We are not able to do that. That is where Fynd comes into the picture, and Fynd does a fabulous job of it. We come in as the marketing partner.
A good ballpark is that if you are spending ₹5 lakh a month, you are probably already doing more than ₹10 or ₹12 lakh in monthly revenue. That is a good fit. The journey from ₹1 crore to ₹100 crore is what we specialise in.
One more thing, since people hear performance marketing and think it means running ads. The term implies brand marketing is non-performance marketing, and I don’t believe that. Both give you performance. What people mean is direct response: you run a campaign and measure sales over the next seven days.
Meta has 200 to 300 million people who could be in your target group. How is Meta supposed to know which of them is most likely to buy? Only if you pass the signals from your website back, which is what the pixel is for. Getting your pixel and events configured correctly is the 101 of this.
When you boost a post you are passing none of that back. You are saying I have a nice post, please show it to people who might like it. They see it, they like it, they learn about the brand. They don’t shop.
Ten years ago people used likes and shares as their north star. We were among the first companies in India to say we don’t care about likes. But it goes further. We have seen campaigns optimised for add to cart deliver plenty of add to carts, and the conversion after that drops substantially.
Think about the chain. Add to cart, then purchase, then an online payment purchase rather than cash on delivery, then a delivered order, then a delivered order the person doesn’t return. Your P&L is directly linked to that last one, so pass it back and ask Meta to optimise for it. The AI is very intelligent now. The better the signal you give it, the better it gets.
On a marketplace, the marketplace is protecting you. On a brand’s own website there is always a little discomfort. What if I pay and the order never comes, and who do I chase? Even today, for bulk of the D2C brands, around 60 to 70% of purchases come through cash on delivery, and among new customers that share is higher.
There are two levers. Exclude known COD defaulters using data from checkout partners. And reallocate budget between campaigns: if one gives you 100 orders with 60 on COD and another gives you 100 with 40 on COD, put more money behind the second. Every campaign can now show its COD percentage in the dashboard itself.
If there is a COD charge or an online payment discount, put it on the product page or the announcement bar. Tell people ahead of time. Customers anchor on a price, and if they have decided the product costs ₹1,900 and it becomes ₹1,950 at checkout, you will see drop-offs.
After the order, WhatsApp flows work really well, because a lot of these orders are impulsive. Someone is scrolling Instagram at 12:30 at night, sees a nice pair of shoes and buys them, and by morning they have half forgotten. A message saying it ships today and arrives Saturday reminds them, and tells them to keep the cash ready.
For the longest time Facebook was seen as a way to get sales on your website. Now you can send signals from the point of sale machine in your offline store back to Meta through the offline conversion API, and to Google through store sales integration. For the first time you can close the loop from online to offline and run performance campaigns for store footfall.
Same on marketplaces. Five years ago you could send someone from Facebook to Myntra but had no way of knowing if they bought. Facebook is now integrated with Blinkit, Zepto, Instamart and Myntra. One brand we work with sells sanitary pads: you see the reel, you click, you order on Blinkit, it arrives in ten minutes. Awareness to conversion in half an hour.
Listen to the full conversation with Anshuk Aggarwal on FyndOutWithRagini.
Fill out the form
Share your contact information to get started
Speak to an expert
A member of our sales team will get in touch with you