Teikametrics

It is two in the morning and somewhere a brand manager is staring at an Amazon dashboard that will not tell them why their best product stopped selling. The bid was right. The listing looked fine. Then a competitor changed one keyword and the whole thing quietly collapsed. Teikametrics built a company around that exact 2 a.m. feeling.

The marketplace is the largest store ever built, and also the most indifferent. Amazon, Walmart, and TikTok Shop do not care whether your business survives the next algorithm update. They publish rules in a language nobody fully speaks, change them without notice, and reward the sellers who can do the math fastest. Teikametrics, a roughly 190-person company headquartered at 280 Summer Street in Boston, exists to do that math - continuously, automatically, and on behalf of brands that would rather make products than reverse-engineer an ad auction.

A casino where the house writes the odds

Here is the uncomfortable truth about selling online: the marketplace knows everything about you, and you know almost nothing about it. Every search, every conversion, every abandoned cart feeds a ranking system you cannot see. Sellers respond the way humans always do when faced with a black box - they guess, they overspend on ads, and they call it strategy.

That was the gap. Most tools treated advertising, organic search rank, and inventory as separate spreadsheets owned by separate people. Teikametrics' argument is that they are one system: ads drive sales, sales drive rank, rank drives more sales, and inventory either feeds the loop or starves it. Optimize one in isolation and you can win the battle while quietly losing the war. The keywords in the company name come from the Japanese word teika - market price - which is, conveniently, the one number every seller is trying and failing to get right.

$900, two suitcases, and a dorm room

Alasdair McLean-Foreman did not arrive at this problem from a McKinsey deck. He arrived in America with $900 and two suitcases, captained Harvard's track team, and - in the very best entrepreneurial tradition - started a sporting-goods business out of his dorm room in 2001. By 2003 it was one of Amazon's first third-party sellers in its category. He was, in other words, the customer long before he built the product.

Selling on Amazon in those years meant living and dying by price, so he wrote software to reprice his own inventory. That software became the seed of Teikametrics, founded in 2015. The bet was simple and slightly heretical for its moment: give independent sellers the same data-driven, profit-first tooling that only the largest enterprises could afford to build in-house. Democratize the math. Then get out of the way.

It is a tidy origin story, and like most tidy origin stories it leaves out the years of unglamorous work in between - the fitness startup Traineo, the e-commerce systems built for Newscorp, L'Oreal, and the New York Marathon. The throughline is consistent, though: take a messy commercial problem and turn it into something a computer can optimize.

One platform, three marketplaces, zero SQL

The flagship is Flywheel, the marketplace optimization platform. Flywheel 2.0, launched in 2022, added cross-marketplace keyword harvesting - it finds the search terms quietly winning on one channel and deploys them everywhere a seller's products appear. The pitch is not "more ads." The pitch is fewer wasted ones.

Then the company did what every ambitious software company did in this decade, which is add the letters A and I to everything - except here it actually meant something. In August 2025 came GenAI Smart Pages, which rewrites product titles, bullets, and descriptions using a seller's own performance data instead of the generic copywriting sludge most AI tools produce. In December 2025 came ARI - Artificial Retail Intelligence - a patent-pending engine that lets a seller run Amazon, Walmart, and TikTok Shop from one seat, tap Amazon Marketing Cloud audience insights without writing a line of SQL, and launch full-funnel campaigns several times faster.

For the customer who would rather not touch any of it, there are managed services - actual humans who run the account for you. It is a refreshingly honest acknowledgment that software does not solve everything, and that some brands just want to hand someone the keys. The technology stack underneath is the unfashionable kind that actually ships: Kubernetes, Terraform, OpenSearch, Datadog, and a wall of AWS analytics, all in service of a product whose entire job is to make a complicated thing feel like one button.

What makes the platform genuinely useful, rather than merely clever, is that it is built around a metric most marketplace tools politely avoid: profit. It is easy to grow revenue on Amazon - just lower your price and buy more ads until the numbers look impressive and the bank account does not. Teikametrics optimizes toward the figure founders actually keep, which is a harder problem and a more honest one. With it, a small brand can run the kind of disciplined, cross-channel campaign that used to require a full analytics department, a media buyer, and a standing 4 p.m. meeting nobody enjoyed.

Numbers that have to earn their place

Skepticism is the correct default with any company that says the word "AI" this often, so here are the figures that matter. Teikametrics reports it has optimized billions of dollars in marketplace GMV across thousands of sellers. It counts roughly 2,500 customers and around $45 million in annual revenue. The named brands include Clarks, Razer, Power Practical, Zipline Ski, and a clutch of Mark Cuban's companies - the kind of logos that do not lend their names to software that does not work.

Data, AI, and actual human expertise

The stated mission is to help sellers and brand owners maximize their potential on the most valuable marketplaces by bringing data, AI, and expertise together in one system. Strip the press-release polish off it and the ambition is clear enough: level a playing field that has always been tilted toward whoever could afford the biggest analytics team.

The algorithm isn't slowing down

TikTok Shop turned discovery into entertainment. Retail media is now a margin business that rivals advertising's golden age. Generative AI is rewriting product pages faster than any human team could approve them. Every one of these shifts makes the marketplace more complex, more automated, and more hostile to the seller flying blind. The competitors - Pacvue, Perpetua, Skai, Helium 10 - all see the same horizon. The race is over who turns the chaos into something a brand can actually steer.

So return to that 2 a.m. dashboard. The brand manager still cannot see why the product stalled - but now something else can. The bid adjusts before they wake up. The keyword that quietly won on Walmart is already live on Amazon. The listing has rewritten itself around what is actually converting. The mystery did not disappear. It just stopped being the seller's job to solve alone. That is the unglamorous, deeply useful thing Teikametrics built: not magic, just the marketplace, finally readable.