Merchant intelligence for the AI commerce era

Acquisition costs keep rising. The brands that win won't spend more — they'll know more.

Acquiring a shopper has never been more expensive — and most stores still treat the one they paid for like a stranger. We make your store know each customer from the first second and sharper every visit, so the spend you've already made actually pays back: faster conversion, more repeat, higher lifetime value.

A few honest questions
01

Every shopper you pay to acquire sees the same generic store. Is that really the best first impression you've got?

02

Your best customer has shopped with you a dozen times. Do you know their taste — or are they still just shopper4471@gmail.com?

03

What if your best customers came back because the store knows them — not because you paid to chase them down again?

The math everyone's living with

Acquisition keeps getting more expensive. The only lever left is what happens after the click.

+40%
Ecommerce customer acquisition cost since 2023. It's structural — privacy changes, ad inflation — and it isn't coming back down.
−$29
The average brand now loses ~$29 on each new customer's first order — up from a $9 loss a decade ago.
3:1
The minimum LTV-to-CAC ratio for sustainable growth. Hit it through repeat purchase, or you're buying customers a competitor keeps.

You can't fix CAC. You can only earn it back — and you earn it back by knowing the customer well enough to bring them back without a discount.

Sources: Shopify / First Page Sage / industry CAC benchmarks, 2025–26.
The shift

Every transaction is a customer telling you their taste. Most stores throw that away.

A purchase isn't just revenue — it's a signal. What they chose, what they skipped, what they came back for. Stitch those signals together and a shopper stops being an order number and becomes someone you actually know. That's what turns a one-time buyer into lifetime value: a store that remembers, anticipates, and earns the next purchase instead of paying for it. We turn the signal your customers are already giving you into a store that knows them.

What "knowing" looks like

Knowing a customer isn't a category. It's a person.

Anyone can tag a shopper by what they buy. Real taste is the specifics — and the things they'd never touch. That's the difference between a guess and a fit.

Coffee
"drinks coffee" light roast, single-origin, fruity not nutty, never decaf
Skincare
"buys skincare" fragrance-free, sensitive skin, clean ingredients, won't touch a retinol
Apparel
"likes apparel" organic cotton, relaxed fit, under $80, made in USA
Offers & cashback
"uses offers" chases grocery and gas, ignores fashion, only acts above 5% back, redeems monthly
The biggest platforms on earth already decided this. Discovery isn't search anymore — it's recommendation.
40%
of Google Play installs come from recommendations
60%
of YouTube watch time comes from recommendations
Source: Google — developers.google.com, ML recommendation systems documentation
What we do

Relevant on arrival. Sharper every session. Eventually, a store that just knows.

Not another popup or discount engine. An intelligence layer that reads who the shopper is and gets more precise the longer they engage.

FROM THE FIRST SECOND

Relevant on arrival

A brand-new visitor with zero history still carries signal — where they came from, device, time, intent. We read it and make the first visit feel like the fifth. No generic feed.

WITH EVERY INTERACTION

Sharper every session

Clicks, skips, dwell, purchases, returns — each one teaches the store more. Relevance compounds. The experience adapts in real time instead of staying frozen.

OVER TIME

A store that knows their taste

Not "likes apparel" — the real, attribute-level understanding your best salesperson would have. Deep enough to power browse, search, and discovery alike.

And when they talk to an agent

More shopping is happening through AI. When your customer asks an agent, they shouldn't be met as a stranger.

Here's the quiet advantage: knowing your customer is the hard part. Solve it for humans and you've solved it for agents too — the same understanding that personalizes your store is exactly what an agent needs to represent your customer well. Get the human right, and the machine comes for free.

Why us

ForYouLabs is built by product and data veterans — pairing two decades of consumer-commerce leadership with deep personalization and ML expertise.

Founder Aviral Gupta has spent 20+ years on both sides of this problem. At Amazon, he ran real retail businesses — a $100M apparel P&L, fashion across Japan, the #1 national baby brand and Amazon's first private label, and retail operations spanning the US, Australia, and India. He knows firsthand what it takes to run a store and win a customer.

Then he built the answer. At Fetch, he ran the commerce and discovery surfaces serving 18M monthly users — roughly 60% of company revenue — and is lead inventor on the patented personalization system behind ~360M monthly interactions. At Trashie, he led taste-driven recommendation, turning customer signal into personalized discovery. Earlier, product strategy and operations leadership at LinkedIn.

On the data and ML side, ForYouLabs is advised by Raj Prazad, former SVP of Data at Fetch — bringing the modeling and infrastructure depth that turns shopper signal into a store that knows them.

This isn't a bet on a new idea. It's the next version of work that's already shipped at scale.

$100M
Apparel P&L run at Amazon
4
Retail leadership across 4 countries
18M
MAU on personalization surfaces at Fetch
~360M
Monthly interactions, patent lead inventor

Ready to actually know your customer?

We're working with a small number of pilot partners to prove out measurable lift. A short call to see if it's a fit — no deck, no pressure.

15 minutes · we'll talk through your funnel and where the opportunity is