Revenue · last 30 daysMeta + Google
₹92.3L▲ 212%
Performance
marketing
Think, plan, test, learn, think. Budget follows profit, not gut feel.
Read this
part first/
part first/
Performance marketer who also builds the store. Rare combination. Mildly annoying to agencies.
Ads that sell.
Stores that
convert.
Your ads agency blames the website. Your developer blames the ads. Meanwhile, CAC went up again. I do both jobs, so there's nobody left to blame. Including me.
Careatic, my own
D2C brand
D2C brand
100
Orders in the first 15 days. My ads, my store, my money.
Scaled a brand seen on Shark Tank India
₹729 → ₹362
Smotect's cost per order, April to September 2025
Fix the
process leak
Scaling a leaky process doesn't grow the brand. It just makes a bigger puddle.
Open to new brands
Meta Ads
Google Ads
Shopify
AI content
Ads & build
Meta AdsGoogle AdsAmazon AdsShopifyWordPressFigmaCanvaShiprocketPayUDelhiveryEngage 360
AI stack
ClaudeChatGPTHiggsfieldMidjourneyHeyGenHeyclickyn8nSupermetricsWindsor.ai
Your ads aren't
the problem. 
Neither is
your store. It's the
gap between them
Most D2C brands hire an agency for Meta, a freelancer for Shopify and someone's cousin for creatives. That's three people, three reports, and nobody accountable for the only number that matters: the money left after the order ships, doesn't come back as RTO, and the customer actually reorders. I work in that gap. That's the whole job.
For founders who
think big
think big
9 yrs
In the media industry, including performance marketing since my first Meta campaign in 2020
9+
D2C brands grown from scratch as Head of Performance at Weird
100
Orders in 15 days for Careatic, my own D2C brand
−50%
Cost per order at Smotect, ₹729 to ₹362 (April to September 2025)
Three jobs. One
person. Zero excuses.
Normally this takes three vendors and a WhatsApp group where everyone agrees in theory. Take all three, or just the one that's bleeding right now.
Performance
★★★★★ Careatic
Laal Bindi dog kurta
₹2,249 · Sized by breed
Find my dog's fit
AI D2C
stores
Loads in under 2s on a ₹10k Android on Jio 4G. That's who's buying.
AI content
creation
Fresh AI video ads to test, without booking a shoot.
LAMA INJA Wellness Smotect Byogi Fearless Tea Careatic Requil Taggie Clay.co Chalk & Chuckles Rungajaun Tea
LAMA INJA Wellness Smotect Byogi Fearless Tea Careatic Requil Taggie Clay.co Chalk & Chuckles Rungajaun Tea
Case studies
Receipts, not
adjectives
My own brand, a freelance client and a Shark Tank brand. Real numbers, and I'm happy to walk you through how each one happened.
FounderPet apparelAds + store + visuals
I paid ₹50,000 to learn the website was the problem.
Careatic makes dog clothing sized by breed, because a Pomeranian isn't a small Labrador. I hired developers, and the site they delivered looked premium but loaded whenever it felt like it, with tracking that only worked sometimes. I was paying for ads to send people to a loading spinner. So I rebuilt the store myself in three days, made the product visuals with AI instead of booking a shoot, and sold the problem ("will this fit my dog?") instead of a cute raincoat.
- Store rebuilt from scratch
- Breed-first positioning
- AI product visuals
100Orders in the first 15 days
₹2KTotal spent on product visuals
3daysTo rebuild the website
K9 KicksSneakers · D2CStore + Meta ads
Zero to ₹1.2 Cr in three quarters. Starting at ₹1,500 a day.
K9 Kicks had no store in January. I built it first: product costs, shipping, pricing, payment gateway, checkout and CAPI tracking. Then the loop. Month one tested every format and angle, month two carried the learning forward, and by month three it was ₹8L a month at 6× ROAS. Then Instagram video ads took over: five months later, its best month did ₹40L in gross sales, on ₹18,000 a day of ad spend that's still climbing.
- Store built from scratch
- Checkout + CAPI
- Format and angle testing
- Instagram video ads
- ₹1,500 → ₹18,000 a day
₹1.2CrTotal sales, Jan to Oct 2026 (Shopify)
6,843Orders in three quarters
6.5×Blended ROAS, Jan to Oct 2026
SmotectAs seen onShark Tank India
SmotectHead of GrowthMeta · Google · Amazon · Flipkart
₹729 to ₹362 per order. On less budget, not more.
Smotect sells a nicotine-free formula to help people quit tobacco, and it's been on Shark Tank India. When I took over in May 2025, ₹27L a month was going out with nobody acting on it: 40 to 50 campaigns on ₹1,000 budgets, and winners that were never scaled. I cut the dead spend, let the account rest, rebuilt it into 20 to 25 campaigns with new ads every week, and checked the numbers four times a day. By June, ROAS with Amazon hit 3.2×. By September, the brand made April's revenue on ₹7.8L less spend. I also planned and rebuilt the website along the way.
- Cut dead spend first
- 20–25 campaigns, weekly ads
- 4 checks a day
- Website rebuild
₹362Cost per order, down from ₹729 (Apr to Sep)
3.2×ROAS with Amazon by June, from 2.18×
2.75×Website ROAS by September, from 1.89×
What's leaking
in your brand?
Start free: I'll make one AI ad for your hero product and audit your store or ad account. Both are yours to keep.
Current clients
Live, not
lorem ipsum
Brands I'm working with right now. Click through, because every one of these is real. Most were built with Claude, because AI lets one person ship like a team.
K9 KicksSneakers · D2C
Built their Shopify store with Claude and run their Meta ads. Zero to ₹1.2 Cr in three quarters, and the store and the ads were built for each other.
Website · built with ClaudeMeta AdsRead the case study
Local SparrowFood · D2C
Redesigned the Moringa Super Podi product page with Claude, then ran Meta ads to it. A better page meant the same ad spend bought customers for less.
Product page · built with ClaudeMeta AdsLower CAC
Sankalp Research WorkPhD assistance · Lead gen
Designed a new website with Claude and run Meta lead-gen campaigns that send scholars straight to WhatsApp and calls, not a form nobody fills in.
Website · built with ClaudeMeta lead-gen ads
Zantarra ImmigrationsOverseas manpower · Services
Built the website by hand, the old-fashioned way. Useful to know when your brand needs something a template or an AI can't give you.
Website · hand-built
CareaticPet apparel · My own brand
My own brand, so I'm the client too. I built the website with Claude and run the Meta ads and the brand strategy.
Website · built with ClaudeMeta AdsBrand strategy
This websitePortfolio
You're on it. Designed and built with Claude, from the brand identity to this sentence. If you like how it feels, that's the pitch.
Website · built with Claude
Founder, Careatic · ex-Weird
Hi, I'm Arun
One person. Both
sides of the funnel
I started in 2020 with a digital marketing programme from MICA, my first Meta campaigns and a lot of money spent learning which ads don't work. Then I spent two and a half years as Head of Performance Marketing at Weird in Mumbai, running paid media for D2C brands, and went on to lead growth at Smotect and media at Digiweave.
In 2026 I started my own brand, Careatic, and learned what founders actually go through. I paid ₹50,000 for a website I had to rebuild myself, and shot product photos with AI because studio quotes were out of reach. We hit 100 orders in 15 days. So when I tell you your product page is losing buyers, I've lost them on mine first.
The ideology is simple: think big, spend big, get big. Not reckless big. Tested big. Think, plan, test, learn, think, and put more money behind whatever wins.
WeirdHead of Performance Marketing, 2022–2025
CareaticFounder, breed-specific dog apparel
MICAPG, Digital Marketing · BMM, Advertising
Education & certifications
The paperwork behind
the proof
Degrees don't run ads. But this is where the fundamentals came from, and I keep going back to learn more.
Education
- 2020–21PG in Digital Marketing & Mass CommunicationMICA, The School of Ideas · Grade A
- 2020–21Facebook Marketing CertificationupGrad · Meta Ads, campaign setup & optimisation · Grade A++
- 2014–17Bachelor of Mass Media, AdvertisingKirti M. Doongursee College, Mumbai · Grade A
Certifications
- Apr 2026Brand Creative DirectorRajeev Mehta Academy
- Feb 2026Certified Product MarketerGrowthSchool · Product marketing mentorship
- Oct 2025Google Ads Search CertificationGoogle Skillshop
- Nov 2021PG in Digital Marketing CertificationMICA · Meta Ads, Google Ads
- Aug 2021Growing Business with FacebookFacebook × upGrad
How we work
No magic. Just
the loop
Think, plan, test, learn, think. No secret sauce and no 40-slide deck you'll never open again. Four steps around the loop, and you see every one.
01
Audit
I go through your ad account, Shopify data and unit economics, then tell you what's broken. That includes the parts nobody has mentioned, like the 30% RTO.
Week 102
Build
Fix the leak before pouring in more water: the store, landing pages, tracking, and the briefs your team shoots for the first round of tests.
Weeks 2–403
Launch
Angles, formats and variations, each one a named test. Your team ships new content every week; I read the account every day.
Week 504
Scale
Whatever wins gets more money. Whatever doesn't, doesn't. Weekly reports in plain English, not a screenshot of CTR.
OngoingPricing
Small fee. No
small budgets.
You scrolled straight to the prices. Fair, so would I. One warning before you do the maths: if the plan is ₹500 a day and a hundred orders, close this tab. ₹500 a day is ₹15,000 a month. A hundred orders on that is a customer for ₹150.
Start here
Start free
Free
Before you pay me anything: one AI ad for your hero product, and a written audit of where your current money is leaking. Both yours to keep, even if you hire someone else.
- One AI video ad for your hero product
- Structure, angles, formats: what's really being tested
- Pixel and CAPI: what's really being counted
- The leak list, in the order I'd fix it
One-time
AI D2C store
₹30,000complete store + GST
I once lost ₹50,000 on developers. Then I rebuilt my store myself with AI in about three days. You're paying for the second part.
- Product pages built to the same standard
- Pixel, CAPI and GA4 that count what actually sold
- Yours to edit, no developer on retainer
Per content
AI content
₹1,000onwards
AI video ads for your products, made in your brand's look. Fresh creatives to test, without booking a shoot.
- 12 or 20-second videos
- In 720p or 1080p
- Your first one is free (see Start free)
The loop
Performance marketing
₹20,000/ month + GST
Think. Plan. Test. Learn. Think. Everything lives inside this loop. I run it on your ad account. Your team makes the content I brief, or we make it with AI from ₹1,000 a video.
- Daily account reading, Meta and Google
- Briefs: the angle, format and variation to test next
- Budget follows profit, not gut feel
- A plain-English report: what won, what died, what's next
Coming soon
Arun's AI course
₹499to pre-book your seat
Course fee at launch: ₹20,000 + GST. Your ₹499 is adjusted against it and is fully refundable.
Rather do it all yourself? Learn everything I do for clients, with Claude and the AI tools I use every day. Pre-book your seat for ₹499 and you'll hear first when it opens.
- Build your D2C store with Claude
- AI product visuals
- Meta and Google ads, step by step
- Tracking and reports you can trust
A freelancer can work for less. Your brand and Meta can't. Sapna dekha hai toh build it strongly, and if the budget isn't there yet, wait for the right time. We work on a pay-in-advance model, and these prices are for my first 10 clients. All prices exclude GST. Ad spend goes straight to Meta or Google, never through me.
Clicks are cheap.
Customers aren't
anyone can buy traffic —
I build what keeps it
Let's find
where your
money leaks
If you've read this far, you've read more of this site than most founders read of their own ad account. Might as well talk.
FAQ
The awkward questions,
first
Most agencies save these for the third call. Here they are before the first.
Can I start with ₹500 a day?
Buddy, not if you want hundreds of orders. The maths is in the pricing section, and it isn't kind. Money generates money: put none in, see none come out.
So what's a real budget?
Around ₹1.5L a month on Meta to test properly. Under that we're guessing, and you'd be paying me to guess.
Then why is your fee so low?
Because I'd rather your money go to Meta than to me. A heavy agency retainer for performance alone eats the budget the loop needs to learn. My fee stays small so your ad account doesn't have to.
Do you make the content?
Yes, two ways. AI video ads start at ₹1,000 each, and your first one is free, so you can test it against what you run now. For real shoots there's a studio and a small team, and if the budget matches, we'll shoot it. But I'll try to talk you out of that first. Make it yourself for one quarter, with my brief on every piece: the angle, the format, what it has to prove. You'll learn what one good ad really costs: the idea, the retakes, the edit, the version that flopped. After that, nobody in your office says "sirf camera leke shoot hi toh karna hai" ever again.
What happens in month one?
The burning part. Testing, burning, testing, burning, testing, earning. INJA Wellness started at 0.7× in month one and got to 3.5×, which for most D2C brands is breakeven. Anyone promising 5× in week one is counting something you can't bank.
Why one person?
Because the person you talk to is the person reading your account every day, and the person who built the page your ads land on. If ROAS drops, there's exactly one person to blame. When I'm full, I'll say so. No intern handover.
My website is fine. Just run the ads?
Sure. I'll still look. Most "fine" websites fail the basics: shipping, returns, COD and "will this fit me" answered before the scroll. If yours passes, we skip it and save you the money.
Meta or Google?
Meta makes people want it. Google catches the ones already searching. Anything else needs a reason, not a trend.
Revenue share?
Not on day one. On day one the tracking is usually broken, and you can't share revenue nobody can count. After three months of clean data, ask again.
How fast can you start?
Audit in about a week. Campaigns live in about three, depending on how broken the tracking is. It's always more broken than you think.
I'm not ready yet. Now what?
Then wait. Seriously. Save the budget, fix the store, and come back when you can run the loop properly. Starting late is cheaper than starting broke.