Run paid ads early to learn faster, not to grow faster

Matt Swulinski, the growth leader who has run growth at email startup Superhuman and voice dictation startup Wispr Flow, was asked when a young startup should start buying ads. His answer, he argues, is right away, but only once the basics are in place: a product people genuinely love, working conversion tracking, customers already coming in the door, and about 50 conversions so the ad system has enough data to learn who to find.

Most founders, he said, are told paid ads are a dangerous drug that hooks you too early. He disagrees.

He argues that paid ads let a team check whether the product can scale in any real way. PLG is short for product-led growth, software that sells itself because people can try it, like it, and share it.

Swulinski argues that founders misunderstand what the ads are for. They are not for growth at first, they are for learning. In a single week of paid testing, he said, a team can refine its messaging, try different creative, meaning the ad images and videos, test its signup funnel, meaning the steps from ad click to signup, and test its positioning.

That is much faster than the usual slow engine, he argues. Writing posts, talking on podcasts, waiting to see what sticks, and doing user interviews all take time. Do that work too, he said, but run paid alongside it as a faster engine for checking what is true.

He says the answer comes fast. He suggests two or three cycles of launching ads, checking what worked, and trying new versions. A founder will know within two to three weeks whether the campaign can make money or needs a complete rethink. He recommends giving it a full three months, while fixing the small levers that add up: the images and words in the ads, the landing page and the words on it, how fast the page loads, and whether the screenshots in the app store make sense.

Keep refining all of it as you spend, he argues, or results will stop improving.