He decided the retailer's algorithm was a customer too - and nobody was selling to it.
Jason Kowalski. The look of a man who has read the Best Seller Rank so you don't have to.
The Dispatch
Most people who spend on influencers are counting the wrong thing. They tally likes, reach, impressions - the confetti of the internet - and hope some of it lands as a sale. Jason Kowalski runs ProductWind on the opposite instinct. His company coordinates thousands of creators not to be admired, but to be counted: units moved, reviews earned, ranks climbed. The audience he cares about most doesn't have a face. It has a ranking function.
ProductWind is an influencer technology platform built for one specific arena - brands that live and die on online retailers like Amazon and Walmart. Kowalski's argument is that those marketplaces are run by algorithms, and algorithms are a kind of buyer. They decide which products get seen, which get buried on page five, which earn the badge that turns a browser into a buyer. Market only to the shopper and you're speaking to the room while ignoring the bouncer at the door.
That framing - the retailer as a second customer - is the whole company in a sentence. Everything else is machinery built to serve it.
The Big Idea
Retailers are getting smarter, and Kowalski's bet is that as they do, the old marketing playbook quietly stops working. The classic move - flood consumers with a message and wait - assumes a human is doing the choosing. On a modern marketplace, the first gatekeeper is code. It reads sales velocity, review flow, and search behavior, then decides what deserves the front page.
So ProductWind categorizes creators by what they're actually good at, then orchestrates them like an ensemble. A launch isn't a post. It's a sequence - coordinated bursts of activity engineered to look, to the algorithm, exactly like a product catching fire. Traffic rises. Conversions follow. The system notices. The badge appears.
Where It Came From
Before ProductWind, Kowalski spent his Amazon years as Head of Product Management for Promotions & Deals, building out the retailer's self-service marketing platform. That is a rare vantage point. He wasn't guessing at how the machine rewarded products - he was inside the room where the machine was assembled.
He was also, on his own time, an Amazon seller. So he felt the gap from both sides: the platform builder who knew what moved the needle, and the merchant staring at a marketing bill with no clear line to a single extra sale. Brands were pouring money into influencer campaigns and had almost no idea whether any of it translated to retail. That disconnect became the founding premise.
After Amazon he served as Vice President of Product Management at Incode Technologies, before starting ProductWind in 2020. He studied at Cornell University, and earlier in his career worked at Revolution Prep - the kind of teaching-and-selling apprenticeship that tends to leave a person comfortable pitching a stranger cold.
Which came in handy. In the earliest days, before the product was finished, Kowalski walked trade show floors and pitched potential customers to their faces. No demo, no polish - just the idea and the nerve to sell it.
The Unfashionable Part
The tidy startup story raises money, then chases revenue. ProductWind ran it backward. For its first 18 months the company was bootstrapped and profitable - not profitable eventually, profitable from the start - while revenue grew tenfold. Only then did the venture money arrive: $1.67M in 2021 led by Early Light Ventures, then $3.5M in 2022 led by Tenacity Venture Capital and Broom Ventures.
To do it, Kowalski recruited a co-founder who could build the machine underneath the pitch: Tom Hirschfeld, a former Uber engineer. One founder sold the vision on trade show floors; the other turned influencer chaos into automated, coordinated signal. The proof was blunt and repeatable - client products dragged from the graveyard of Amazon page five up to page one.
Treat the retail algorithm as a customer with its own strategy, separate from the shopper.
Sort creators by strength and fire them in coordinated sequences that read as momentum.
Move the metrics algorithms reward - velocity, reviews, rank - not the ones that only flatter.
The Decoder
If You Only Remember Six Things
Follow The Trail
Pass It On
A good idea deserves 200 influencers. This one will settle for you.
Compiled from public sources • Facts verified where possible, and left out where not.