ROAS Reality Check
Ads Manager reports revenue, not profit. This works out the ROAS a store actually needs once product cost, shipping, payment fees and returns are counted, and the most it can pay for an order.
I find where money leaks between an ad and a sale.Then I build the fix.
Computer science engineer with 5+ years in paid performance marketing. At PHD Media I run paid social for brands like Red Bull, Lipton and GO Transit. I also build tools that catch the money leaks ad dashboards miss, for teams of any size.
- Millionsin ad spend managed for global brands
- Under 30¢per purchase, by running conversion campaigns on platforms usually bought only for reach
- 7ad platforms run, from Meta to Reddit
Day jobSocial Manager, Paid PerformancePHD Media (Omnicom), Toronto
Utkarsh Jukaria
Ads Manager says your ROAS is 3x. Your bank account disagrees.
$100in sales at a 3x ROAS
$100left as profit
Keep scrolling to watch where it goes.
Big brands pay a team to catch this. I'm on that team, and I build the same checks for shops that can't afford one.
Selected work
Things I've built
Each one started with a problem I kept running into. AI writes the first draft, then I test it until it holds up.
Discount Stack Check
One extra tick box lets a sale, two codes and free shipping stack below cost. This rings up the worst cart your Shopify discounts allow, before Black Friday shoppers find it.
Register to Store
Small makers count stock on paper. Snap a photo of the page and Gemini reads it, in English, French or Hindi, into a Shopify product file. Anything it isn't sure about gets flagged for a person to check.
Campaign naming tool
An AI-assisted tool I built at PHD Media that names campaigns and audiences the same way across five ad platforms. My team now uses it, and reporting rolls up without anyone fixing names by hand.
Internal tool at PHD MediaThe Attribution Argument
One $120 sale, four channels, five attribution models. A scroll-through explainer of why every platform claims the same order, with a builder to test your own customer journeys.
Career
Laid out like a media flowchart
Agencies plan campaigns on a flowchart: one row per line item, one bar for when it runs. Here's my career on one. Select a row to see what I did there.
Swipe the chart sideways to see every year.
Education
- B.Tech, Computer Science and EngineeringGraphic Era University, India, 2017 to 2021
- Post Graduate Diploma, Business AdministrationYork University, Toronto, 2023
- Post Graduate Certificate, Social Media MarketingSeneca College, Toronto, 2022
Certifications
- Meta Certified Media Buying ProfessionalMeta
- Google Ads Search and Google Analytics 4Google
- TikTok and Pinterest Media BuyerTikTok, Pinterest
Skills and platforms
Both sides of the funnel
An engineering degree for building, and five years of spending real money on ads. Seven platforms, two kinds of client: global brands on multi-million-dollar budgets, and lean accounts where every dollar has to earn its keep. Select a side to see where it spends.
- MetaBig brandsSmall business
- Google AdsSmall business
- TikTokBig brandsSmall business
- SnapchatBig brands
- PinterestBig brandsSmall business
- RedditBig brands
- LinkedInSmall business
Meta, TikTok and Pinterest pull their weight on both sides.
Paid media
Full-funnel planning, bidding, audiences, creative testing, ROAS and CPA across the seven platforms above.
Measurement
GA4, Google Tag Manager, Meta Pixel and Conversions API, Looker Studio, attribution, A/B testing and Excel.
Building
Claude, the Gemini API, HTML and CSS, working knowledge of JavaScript and Node, and WordPress.
Tools and languages
HubSpot, GoHighLevel, Zapier and Canva. I speak English and Hindi.
FAQ
Questions recruiters ask first
Are you authorised to work in Canada?
Yes. I'm a Canadian Permanent Resident, based in Toronto.
How technical are you, really?
I have a B.Tech in Computer Science and a working knowledge of JavaScript and Node. The tools on this page were built with AI assistance, then tested by me until they held up. I'm comfortable with tracking setups, APIs and data.
Do you use AI in your work?
Yes, to move faster on drafts, code and analysis. Every number and every line of code still gets checked by me before it goes anywhere.
Blog
Notes from inside the ad account
Plain-language guides to the numbers that decide whether ads make money: break-even ROAS, attribution, discounts and where AI actually helps.
AINov 11, 20269 min read
AI in performance marketing: 9 use cases that earn their keep, and 4 that don't yet
A working list from inside the ad account: where AI saves real hours in paid media, where it quietly makes things worse, and how to start safely.
MeasurementNov 9, 20268 min read
Why Meta, Google and TikTok all claim the same sale
Add up your ad dashboards and you get more revenue than your store took. Here is how attribution models and windows work, and what to trust instead.
ShopifyNov 6, 20267 min read
Shopify discount stacking: how three good offers add up to a loss
A sale, a welcome code and a loyalty code are each sensible alone. Combined, they can sell below cost. Here is how to find your worst cart before BFCM.
Freelance
A few freelance projects a year
Alongside my full-time role, I take on a small number of projects each year, mostly for small businesses and non-profits that want their ad spend to work harder.
- 01Paid ads setup and audits
Meta, Google, TikTok and Pinterest, with a clear read on what's making money and what isn't.
- 02Landing pages that convert
WordPress sites and landing pages built for paid traffic, fast on phones.
- 03Tracking you can trust
GA4, Google Tag Manager, Meta Pixel and Conversions API, set up so the numbers match reality.
- 04Small AI tools
Custom tools that take a repetitive task off your plate, like the ones above.
Hiring? Let's talk.
I'm a paid social manager who also builds the tools. My résumé and LinkedIn are below, and email is the fastest way to reach me.
hello@utkarshjukaria.comHave a freelance project instead? See what I take on.