Sign UpLog in
All case studies
CloakedCustomer Story

How Cloaked cut CPA 29% while scaling spend 80x with Hypesonic

#1
lowest cost per acquisition in Cloaked's ad account, as of the most recent 30 days
29%
lower cost per acquisition than in its first month on the account, eleven months earlier
80x
the increase in spend on Hypesonic's ads over the past eleven months
Pulkit Gupta
Hypesonic enabled us to scale by keeping CPA consistent as we increased ad spend. We now have the ability to deploy more capital and grow faster.
Pulkit GuptaHead of Growth at Cloaked

The problem

The expensive part was the ideas that never got briefed

In February 2025 Pulkit Gupta found the best ad idea of his year in a news story, and could do nothing with it. A huge company had just been breached, and a large volume of personal data was suddenly loose, with little accounting of who had it. Pulkit runs growth at Cloaked, a company whose entire product is keeping personal data away from people who should not have it. He needed topical content out fast, an ad live in three or four days.

He briefed his agency. A week later they returned with scripts and concepts, which he reviewed and green-lit. The week after that, a final script with a creator attached. The week after that, a draft video. Somewhere in the fourth week the story stopped being news, and the ad that would have run on it was quietly abandoned before anyone filmed a frame.

Between an idea coming to me, me briefing my agency, and the ad hitting my account, it’s already four to five weeks. By which some of these ideas actually get stale.

One lost story is a mere rounding error, but the structural issue is much more serious. When finding out whether an angle works costs five weeks, only the angles you are extremely confident about clear the bar. Those are, by construction, the angles every competitor in the category is already running. Topical work was simply off the table, and he says so flatly: “Basically, I could not run topical ads. It just wasn’t possible to get the ad out in time.”

The same physics throttled everything downstream. Cloaked buys on four platforms with genuinely different audiences on each. When twenty videos was the ceiling of what one agency cycle could produce, all twenty got made for Meta and uploaded everywhere else unchanged. “Which is not the most optimal way to run performance marketing,” he notes.

The skepticism

The failure modes he was watching for

The burden. Pulkit was skittish about AI marketing, and had every reason to be. Over the previous year he had tested nearly every AI creative product on the market, and most of them failed in the same direction: they moved work onto him. He names one of the largest AI marketing tools at the time. “Their process was a headache, because they put all of the burden on me. I’m supposed to be making it easier for myself. Why would I start working with an agency and then put all of the burden back on myself?”

The compliance risk. Every regulated advertiser carries this objection and Cloaked carries it more heavily than most, because the product it sells is trust. An AI avatar that invents a product claim is the kind of thing that makes a compliance team queasy, and it stops being a creative problem the second it goes live. Pulkit says plainly that he expected this to be the failure mode: “I was afraid you’d get details about my product wrong.” The hallucinated claim that puts compliance on alert, he says, never arrived.

I was surprised to see you were spot on, with fidelity, with getting the product details right, with making the right marketing claims.

The public tell. A legal team can be walked through a script before anything ships. The people scrolling past the ad cannot, and they have become fluent at recognising synthetic media and saying so in public, in the comments, where every other prospect can read it. For a brand whose product is trust, that verdict costs more than a rejected claim. It is the second thing Pulkit was watching for, and it did not happen either: “users spotting it, this is AI, I’m calling you out. That didn’t happen.” In the end, he put the whole test simply.

I have two questions when deciding to run an ad: will it work for my compliance team? Will it work for customers? Both of those stakeholders are happy, because you guys’ content is top notch.

The solution

What actually happens before a batch reaches Pulkit

Getting set up. Turning Hypesonic on did not look like bringing in a new agency. There was no roster to build, no production calendar to plan around, no weeks-long ramp before the first batch. Getting it running on Cloaked’s account meant connecting what already existed, the ad account itself, the brand and compliance rules, and letting the system take it from there.

What it watches. The system runs on three inputs that never sat on the same desk before: what is already working in Cloaked’s own account, what competitors in the privacy category are currently scaling, and what is trending on the platforms Cloaked buys on. Every ad in that pool gets broken into its parts, the hook, the claim, the format, the pacing, the call to action, and indexed against how it actually performed rather than how it looked.

What it finds. It surfaces the angles Cloaked was not yet testing but the category was already scaling, the kind of gap you only notice if you are watching every competitor’s account continuously rather than checking in once a quarter.

What it builds. It takes the parts of Cloaked’s own winning ads, the hook that held attention, the claim that converted, and recombines them into new creative instead of starting from a blank page each time.

What it checks. A compliance and brand filter runs before anything ships, so nothing reaches Pulkit’s account carrying a claim the product cannot back up.

The results

From ten videos to the best-returning creative in the account

The loop. There is no weekly call anymore. Batches land on a regular cadence, Cloaked’s growth lead reviews them inside the Hypesonic platform, leaves comments, and approves. Pulkit is not in it day to day, which is the clearest measure of whether it works.

You guys make ads for us, the ads go live, we see purchases, and both of us are happy.

The volume. Cloaked started at ten videos a week with Hypesonic. It now averages about 150 assets a week, cut separately for every platform rather than made for Meta and reposted, and tagged by theme so Cloaked’s landing pages match whatever a visitor actually clicked. Hypesonic creative has served just under 100 million impressions in the account.

The numbers. The money behind Hypesonic creative grew roughly 80x in eleven months, and the cost side never buckled the way it usually does when volume climbs that fast. Cost per acquisition is 29% below where it started. Reach costs about 21% less than the human-production agencies in the same account, a gap that compounds because Meta prices delivery on how creative performs in its own auction. And the customers arriving through it are worth more: the average first purchase runs about 22% higher in value through Hypesonic than through the largest agency Cloaked works with.

What comes next

Wider, not deeper

In August, Hypesonic became the lowest cost per acquisition of any creative partner in Cloaked’s account.

Pulkit thinks most of the market has not caught up to what good AI marketing can actually look like. “A majority of marketers are still seeing AI in marketing as a new concept,” he says, still visibly surprised by it. He does not think that lasts, and he ties it to Meta’s own ranking system rather than to any argument of his own: “Post-Andromeda, this is fast becoming essential. It’s not ‘do I have AI?’ You have to have AI in your marketing stack.”

Going wider. Asked where the partnership goes next, Pulkit reframes the question. “Instead of going deeper, we can go wider with the future,” he says. The deeper version, replacing the account’s human-made creative altogether, is what he is ruling out. “We don’t want to replace human UGC,” he says. What he wants is more of what already works: content on more platforms, in more segments, and ideas he would never brief a human agency for, because it costs so much less to test them on Hypesonic first.

The breadth itself is the point. Testing broadly and cheaply enough is what makes real customization possible instead of one script for the whole audience, which is why he keeps coming back to something Meta itself has been saying.

The one thing that Meta keeps saying is that creative is the new targeting. That’s the future we get to. Literally a creative for each potential customer.

It’s the same math that took Cloaked from ten videos a week to 150: the cheaper a test gets, the more targeting becomes possible. That’s the bet Hypesonic is built on.

See it on your own account

Find out what Hypesonic would test first.

Sign up now