The Anatomy of a High-Converting Ad (What Actually Makes People Buy)
Insights from Harry Delmege · Foxwell Meetup, Lisbon – March, 2026
Why This Matters
In the world of advertising, many agencies are all about the latest trends and flashy new ideas. But Harry Delmege takes a different approach. He looks for what’s already resonating in a brand’s account, borrows that effective messaging, and then presents it in ways others often shy away from.
For example, he crafted long-form, story-driven ads inspired by real Reddit posts, and those ended up outperforming everything else, even when the client was initially uncertain about them. The key insight here isn’t about adopting a new strategy; it’s a powerful reminder that the most engaging ads connect with us on a personal level, almost like a friend sharing a great secret.
Harry Delmege built a 12-person agency with 17 clients in under a year by doing something most agencies are too afraid to do: writing longform, story-driven ads that marketing managers many times shy away from. Spoiler alert: they crush.
At the Foxwell Meetup in Lisbon in late March of 2026, Harry shared the steps he uses to find winning ad ideas and creatives. Here’s his process:
Start where the money already is
When Harry gets access to a new ad account, he does one thing first: sorts by spend. Last 7 days, last 14, and so on. He's looking at the top-spending ads and asking a simple question: what desire is this already speaking to?
For a toothbrush brand, the answer was obvious. Women were self-conscious about their smiles. The messaging was working. The format, though, was a generic mashup of fast cuts, B-roll, and a few creators. Functional, but nothing special.
His move: Take the messaging that's already proven and repackage it in a completely different format. In this case: long-form, first-person, narrative-driven UGC.
"The messaging is already there. If they've scaled up to a certain number, the messaging is working. You can take that and put it in a new style of content."
— Harry Delmege
Most agencies overthink it and try to reinvent the wheel. Harry starts simple with the account's own proven winning tactics.
The story structure that's quietly running everything
Both ads Harry showed followed the same skeleton.
How he structures his ads
Start with a relatable, small problem. "My hygienist said my teeth were yellowing."
Bring in a trusted expert, like a dentist or hygienist.
Introduce a friend or 'neighbor' who’s already solved the problem and is happy to share their secret.
Share the 'aha' moment: reveal what actually works in simple, insider terms.
Explain why past attempts failed, without faulting the viewer.
Show real results and reactions from others as proof.
Harry’s twist: add a second character, the sympathetic peer who’s found the answer first. Every niche has one. People trust friends more than experts.
The strategies you’re missing are probably being discussed here.
Join the conversations you wish you were already in.
The research process: Reddit first, AI second
Real stories come from real people. Online tools can help you get up to speed quickly, but to find the stories that stick, you have to go to real communities like Reddit.
Both toothbrush ads were built from real Reddit posts. Someone genuinely wrote about their hygienist shaming their teeth. Someone else had the dentist conversation. Harry's team found those stories, verified they were emotionally resonant and believable, and then scripted around them.
The rule: If a strategist brings Harry a creative angle without proof from manual research, he won't approve the ad. One compelling story is enough (it doesn't need to be ten),, but it has to be real and emotive.
The bar for proof isn't volume. It's emotional weight. One post that clearly captures something people feel can carry an entire campaign.
Demographics shape the story
The ads ran long. Four and a half minutes for the original cut. A 50+ woman on Facebook will watch that. A 22-year-old on Instagram Reels won't, and that's fine (and by design). The opening line of the ad does the filtering for you: "If you're a woman over 50 and you've tried everything for whiter teeth..."
Understanding who's actually buying (and where they spend their time) shapes every creative decision. The toothbrush ads weren't on Reels being watched at double speed. They were built for an audience that reads comments, watches videos through, and makes considered purchases. Length was a feature, not a bug.
Harry also flagged something he sees constantly in ad accounts: men writing ads for women's markets with a problem-agitation-solution hammer. It doesn't land. Women's purchase decisions are often driven by justification and validation, not by pressure. The ads that work acknowledge past failures, explain why they weren't the viewer's fault, and make the solution feel like an obvious next step and not a sales pitch.
The thing the marketing manager hated
When the brand's marketing manager first saw the scripts, she said they were too boring. No one would watch. No one would buy, but Harry ran them anyway and they became the top-spending ads in the account.
This isn’t about tuning out feedback or not listening to members of your team. It’s just that story-driven, longer ads can seem totally off to folks used to punchy, six-second spots. The ads that work often feel the least like ads when you’re close to them.
"A lot of people are really scared to write scripts like these. Just don't be scared of this type of content and go in."
— Harry Delmege
Three things worth replicating from this
Audit by spend, not by metrics you like. Sort by amount spent. Find what's working. Understand the desire behind it before you build anything new.
Find the real story before you write the fake one. Reddit, TikTok comments, reviews. The best scripts are built on things that actually happened to real people — then scripted to land harder.
Justify the failure. Don't punish it. Especially in women's markets: your job is to make the viewer feel validated, not stupid. The mechanism (40,000 vs. 7,000 brushstrokes) does the explaining so you don't have to shame anyone into buying.
Most newsletters aren’t worth it. This one is.
Get the insights. Skip the fluff.
