How Purposeful Friction Halved Jenfi's Cost per Lead After iOS 14.5
By Kevin Veitia
- Client
- Jenfi, a Singapore fintech offering revenue-based financing
- Markets
- Singapore, Malaysia, Indonesia, Vietnam
- Channels
- Meta Ads (primary), Google Ads, LinkedIn Ads
- Timeframe
- 2020 to 2024
- Engagement
- Kevin led global growth: paid acquisition, regional strategy and the marketing team
The challenge
Jenfi gives revenue-based financing to growing e-commerce and SaaS companies across Southeast Asia. It's B2B fintech, so the job was leads: business owners who wanted growth capital, at a cost sales could actually close.
Kevin led Jenfi's global growth from 2020 to 2024, from his desk in France, while the company went from a handful of people to hundreds. Campaigns rolled out across Singapore, Malaysia, Indonesia and Vietnam, and spend grew to six figures a month.
Every market was its own animal. Singapore was the most expensive and delivered the best leads, which makes sense: it's the region's financial hub, and a lot of companies keep their headquarters there. Vietnam grew fastest and became the heart of the program.
Then Apple shipped iOS 14.5 in 2021. Once iPhone users could opt out of tracking, Meta lost most of its view of what happened after someone clicked through to a landing page. That's exactly how Jenfi's funnel worked, and performance took the hit. Before the fix, Vietnam was running at about $4 per lead, with about 30% of leads marketing-qualified.
The call that changed everything
Jenfi's CEO set us up on a call with one of the company's investors, a product leader with deep experience at Meta. Her diagnosis was blunt: after iOS 14.5, Meta simply couldn't see events that happened off its platform anymore.
Her fix was Meta's instant forms. Kevin scoffed, and she heard it. Every performance marketer has run instant forms and gotten buried in junk leads.
Then she flipped it. Instant forms are junk because marketers build them for speed: pre-filled name and email, one tap, done. The fix is to add friction. On purpose.
What we did
Added purposeful friction
Purposeful friction means designing the form so people have to stop and actually engage with it. Instead of leaning on pre-filled fields, we added five to seven custom questions that made a business owner think about their business, and what they needed, before hitting submit.
Here's the part that matters. An instant form is first-party. Meta sees every step: form opened, question one started, question one dropped, question two started, all the way to completion. That's a rich, reliable signal to optimize on, exactly what it couldn't get from a third-party landing page anymore. Fewer people finish a longer form. But the people Meta learns to find are the ones willing to think about their answers.
Meta's Conversions API, which sends conversion data straight from your server to Meta, was built to patch the same hole for landing pages. It helps. It's still a partial picture. An instant form keeps the whole conversion inside the platform, where the data is complete.
Turned the form into market research
Her second piece of advice: write the questions like a marketing survey. Every completed form tells you who's applying and why. Add them up and you can see which kinds of businesses are responding and what they're after. We fed those patterns straight back into targeting and into the next round of creative.
Optimized to the signal we could trust
For a lending product, the dream optimization event is a sales-qualified lead, and we were sending offline conversions from HubSpot back to Meta. But sales-qualified data is only as good as the CRM stages behind it. In a company growing as fast as Jenfi, with local sales teams changing all the time, those stages weren't updated consistently enough to train an algorithm on.
So we optimized to the marketing-qualified lead, the cleanest signal marketing controlled, and kept working with sales on stage definitions. Almost every B2B company has this fight. The algorithm learns from whatever you feed it, mistakes included, so marketing can only optimize as well as the CRM is kept.
Tested local assumptions instead of obeying them
Localizing the language and each country's financial advertising rules? Non-negotiable. Localizing the look? Different question. When we expanded in Vietnam, the local team wanted creative built on cultural symbols, like the cicada as a sign of renewal, and on the color red, even though Jenfi's brand was blue and teal.
Reasonable hypothesis. So we tested it. The data didn't back it up, and the on-brand creative won. We've seen the same thing in multinational campaigns at Canva, Ground News and Jenfi: good creative travels much further than most teams expect. Local intuition is a hypothesis to test, not a conclusion.
Same goes for the people in the ads. Across thousands of variants, for Jenfi and other clients, ads with a model looking straight at the camera consistently won, and a female model tended to pull a stronger response from men and women alike. We don't treat that as gospel. We treat it as the starting point for the next test.
Results
Switching to instant forms with purposeful friction changed the economics in Vietnam, the program's main market:
- Cost per lead fell from about $4 to about $2
- The marketing-qualified lead rate rose from about 30% to about 40%
- Cost went down and quality went up at the same time, after iOS 14.5 had pushed both the wrong way
- The playbook was repeated across Singapore, Malaysia, Indonesia and Vietnam
Cheaper leads that were also better leads means the cost of each marketing-qualified lead fell by more than 60%, from roughly $13 to roughly $5.
"Kevin was instrumental in our fintech's early marketing success, building impactful digital campaigns and expertly guiding our regional marketing strategy. As we grew, his ability to grasp cultural nuances and navigate country-specific challenges was essential as we expanded into Indonesia and Vietnam. Organized, resourceful, and consistently positive, Kevin provided valuable assessments. I highly recommend him as a skilled and adaptable marketing professional."
Jeffrey Liu, Jenfi
What you can take from this
- If the platform can't see your conversion, move the conversion to where it can.
- Friction can be a feature. Five to seven real questions filter for intent and feed the algorithm a better signal.
- Write lead form questions like a survey. The answers are free market research.
- Optimize to the cleanest signal you control. Fix your CRM stages before you ask an algorithm to learn from them.
- Localize the language and the compliance. Test before you localize the look.
Want a second set of eyes on your account? Book a free 30-minute consultation and we'll walk through what's working, what isn't, and where the next win is. Curious what it costs first? Our pricing is public.
Book a Free Consultation