The brand ad everyone doubted returned 4.7× its spend.
Make the call at each gate, then see what I actually did and what it produced. In a hurry? Skip to the results.
A brand nobody in India knew.
Uplers was known to NRI founders hiring in India, but barely known to founders building in India. The team's belief was simple: brand doesn't convert.
I was the first brand manager. The marketing director approved the budget; the strategy, audiences, creatives and reporting were mine.
You're the first brand hire. 320K founders have never heard of you, and you have about $20 a day. What runs first?
A brand video: the Hiring Anthem. Nobody fills in a form for a company they've never heard of, so the first job was to be known.
Earn attention before you ask for a form.
Week one. Your ads are running to three hand-picked audiences: IIT founders, founders in stealth (only ~4,200 people) and recently funded founders. What now?
Every creative was running to a different, tiny audience, so I couldn't tell a bad ad from a bad audience. I kept the one that engaged (funded founders), cut the niches, and rebuilt around one pool of ~90K founders and co-founders in Bangalore and Delhi NCR. From then on, every creative was tested against the same people.
Strong engagement, so it stayed.
A niche you can't learn from is just an expensive guess.
Guess the winner. Same ~90K founders, three messages: speed, cost and quality. Which one brought in paying clients?
I ran all three against the same audience and let the results decide.
Don't compete on the claim everyone already makes.
Thousands of founders watched at least half of the anthem. What do you do with them?
I built a list of everyone who watched 50% or more of the anthem and retargeted them on LinkedIn with engagement ads, so the people who already knew us heard from us next. Alongside it, a single-image lead-gen ad sent founders to the website. It ran the longest of anything I tested.
Warm audiences are the cheapest ones you'll ever buy.
Bonus level. Your US test on Reddit costs $1.21 a click in week one. Kill it, keep it, or fix it?
I kept it, refreshed the creatives on 19 June, and tuned it week by week.
Reddit cost per click, by week
A bad first week is data, not a verdict.
The campaign that "wouldn't convert" is the one that did.
- Hiring Anthem (brand video)A recently funded voice AI company: a 20-role deal worth ₹4L
- Cost carousel2 paid clients
- Website lead-gen ad5 sales-qualified leads
- Quality carouselMost website visits in two weeks
- Reddit, US testCost per click down 68%
- IIT, stealth and speed adsCut in week one, which is how the rest got a fair test
Each gate is a real decision from this campaign.
Questions a sharp interviewer would ask.
Is the ₹4L in the bank?
₹1L is collected so far, for the first 5 of 20 roles. The rest is billed as roles are filled. That's why I quote 4.7× on deal value and 1.2× in cash.
How do you know the ad brought the client?
The lead came in tagged as paid LinkedIn, brand campaign, in our lead tracker, during the anthem's first active week on LinkedIn.
What is 4.7× measured against?
The ₹4L deal divided by the ₹84,931 spent on the brand campaign that brought it in. It's the brand campaign's return, not the whole ads program's.
What didn't work?
The niche audiences (IIT founders, founders in stealth) were too small to learn from, and the speed message couldn't stand out because every agency claims it. I cut them in week one.
Who owned it?
The marketing director approved the budget. I owned the strategy, audiences, creatives, testing and weekly reporting, running LinkedIn ads for the first time.