From Guesswork to Data-Driven Creative: How Marpipe Uses Buyer Intelligence to Transform Ad Performance

For years, the creative side of marketing has been the last untouched frontier of automation and analytics. Media buying became algorithmic. Targeting became precise. But deciding what an ad should actually look like? That was still done the old-fashioned way — by gut feel and “this looks cool” opinions.

Dan Pantelo, founder and CEO of Marpipe, set out to change that.

What started as a performance marketing agency for venture-backed startups evolved into a SaaS platform that uses multivariate testing to scientifically determine which ad creative actually drives performance — and why.

Lesson 1: Creative Is the Ultimate Performance Lever

While running his agency, Dan noticed a pattern: Changing small details in an ad — a background color, a product angle, even swapping a cactus for a palm tree — could cause major swings in performance.

“If you only looked at the numbers, you’d think you were looking at totally different ads,” Dan recalls.

That insight reframed creative as the single most influential driver of ad performance, outranking targeting, budget, and even brand recognition. But most creative decisions were still being made without data.

Lesson 2: Buyer Intelligence Should Start at the Design Stage

The lightbulb moment came when Dan realized that the creative process could be driven by audience data from the very start. Instead of designing one or two concepts, Marpipe could generate every possible variation — messaging, backgrounds, product angles, colors — and test them simultaneously in-market.

The result: brands learn not just which ad works best, but why it works, and which elements resonate (or fall flat) with their audience.

Agency parallel: In SaaS web design, we use the same principle — testing layouts, copy, and visuals to see exactly what increases conversion, instead of guessing.

Lesson 3: Spin-Outs Require Focus

Marpipe began as an internal tool for the agency’s own campaigns. Clients loved the results so much they started referring others — many saying, “I don’t need your agency. I just want that tool.”

The demand was clear. Dan raised $1.8M in seed funding (including backing from Adobe and MediaMath), shut down the agency, and went all-in on Marpipe.

“You can’t build two businesses at once. You have to rip the band-aid off.”

Lesson 4: Make High-Level Tools Accessible

Competitors existed, but they were built for enterprise, locked behind “talk to sales” gates and $100K+ ad spend minimums. Marpipe flipped the model: a product-led growth approach, with a freemium tier and self-serve onboarding designed for the actual operators running campaigns day-to-day.

This lowered the barrier for adoption, created viral word-of-mouth, and put the product in the hands of the people who would champion it internally.

Lesson 5: Creative Data = A Predictive Advantage

As more campaigns run through Marpipe, the platform is building a first-of-its-kind dataset: which creative components work for which audiences, products, and verticals.

This opens the door to predictive creative intelligence — telling brands before they launch an ad what their buyers will respond to, much like how your agency can predict which site experience or sales enablement asset will resonate most.

The Buyer-Centric Creative Playbook from Marpipe

1. Treat creative as the biggest growth lever, not an afterthought.

2. Test early and in bulk — don’t rely on single-variation “A/B” testing.

3. Use real buyer data to inform design decisions from the start.

4. Lower adoption friction so more users can experience value quickly.

5. Build data assets that make future performance more predictable.

Why this matters for SaaS founders & tech companies: Marpipe’s journey mirrors what we do for SaaS brands: replace guesswork with measurable buyer intelligence — in web design, marketing strategy, and sales enablement. Whether it’s an ad, a landing page, or a sales deck, the most effective version isn’t found in a brainstorm — it’s revealed through structured testing and audience-driven iteration.