Phia's cookie-stuffing fight puts its affiliate model on trial

Bloomberg found the shopping extension could overwrite referral codes; TechCrunch says Impact.com has suspended Phia.

By ยท Published

Why it matters

Phia's speed came from founder-led distribution and affiliate economics. The cookie-stuffing allegation tests whether its reported commerce traction reflects real influence or checkout-time attribution capture.

A browser cookie icon visibly 'stuffing' or forcibly manipulating a shopping cart icon, with fragmented referral codes appearing to be overwritten or tangled. (Risograph two-color print, coarse grain, with fluorescent pink and navy ink laye

Phoebe Gates and Sophia Kianni built Phia around a simple promise: shoppers get a smarter way to find deals, and Phia earns a commission when it helps drive a purchase. A July 10 TechCrunch brief says that model is now under pressure after a Bloomberg investigation found the shopping startup could claim affiliate credit for purchases it did not generate.

The allegation is specific. According to TechCrunch's account of Bloomberg's reporting, Phia's extension could open a new browser tab in the background when a user was already shopping at an online retailer, even when that user arrived directly or through another affiliate such as Wirecutter. During checkout, Phia allegedly overrode the existing referral code and inserted its own, allowing Phia to take credit for a sale another party drove, or that no affiliate drove at all.

TechCrunch reported that the fallout has already reached Impact.com, the affiliate and creator marketing platform. Impact.com suspended Phia after the Bloomberg report, according to TechCrunch. Phia told Bloomberg, through a spokesperson, that it had made the necessary changes to fix the issue; Bloomberg's follow-up check found the behavior had been resolved.

That fix may stop the technical behavior Bloomberg tested. It does not settle the business question behind the incident: how much of Phia's reported commerce performance came from genuine purchase influence, and how much came from attribution mechanics inside the browser.

A young company built on trust and tracking

Gates and Kianni have made Phia one of the most visible consumer shopping startups of the current AI cycle. TechCrunch reported on January 27 that Phia had raised $35 million. TechCrunch's July 10 story says Phia has raised more than $40 million in total and names Kim Kardashian and Hailey Bieber among its investors. The round was marketed around a broad consumer thesis: make shopping feel easier, cheaper, and more personalized at the moment shoppers are already making purchase decisions.

The founder story is central to why Phia traveled so quickly. Gates, the daughter of Bill Gates and Melinda French Gates, brought a built-in public profile to a category where consumer distribution is usually the hardest problem. Kianni is a climate activist. Their pitch joined those two threads around a shopping assistant designed to save money and time while reflecting values many shoppers care about.

The product has since expanded into a broader shopping assistant. Phia's App Store listing describes the app as an AI shopping agent that compares prices, auto-applies coupons, and offers rewards. Those are company-supplied claims in a marketplace listing, and they should be read as Phia's own marketing copy rather than independently audited traction.

Those permissions and features are common for shopping extensions and apps that need to read pages, compare products, and apply coupons. They also create the technical surface that makes attribution disputes possible. Browser extensions live at the point where user intent, merchant checkout pages, coupon flows, and affiliate networks meet. When the extension is trusted, it can make shopping easier. When attribution is mishandled, the same position can redirect revenue away from the publisher, creator, or partner that actually sent the customer.

The affiliate money problem

Affiliate marketing is supposed to be a performance channel: a merchant pays only when a partner influences a sale. In practice, the system relies on tracking codes, cookies, last-click rules, and platform enforcement. A shopping extension sitting inside the browser can arrive late in that chain, close to checkout, and still become the credited partner if it causes its own referral code to be written last.

That is why the Phia allegation matters beyond one startup. The disputed behavior goes directly to the definition of incremental revenue. If Phia helps a user find a cheaper option, applies a real coupon, or sends a shopper to a merchant they otherwise would not have visited, an affiliate commission is easier to defend. If Phia appears after the shopper has already chosen the retailer and product, then inserts itself into the attribution path, the commission looks like a tax on a sale the merchant or another affiliate had already earned.

Industry literature has long treated cookie stuffing as a form of affiliate fraud that can shift credit away from legitimate publishers and impose costs on retailers. That framework maps closely to the behavior TechCrunch says Bloomberg, an independent consultant, and a competitor found in Phia's product.

There is also recent consumer precedent. TechCrunch points to Honey, the PayPal-owned shopping extension that remains the subject of an ongoing class action lawsuit over similar attribution allegations. The comparison is uncomfortable for every shopping assistant that depends on coupon application, rewards, or checkout-time prompts: the closer a product gets to the transaction, the more valuable its attribution rights become, and the more scrutiny its tracking behavior deserves.

Impact.com's role matters because affiliate networks are the enforcement layer between merchants and publishers. TechCrunch reports that Impact.com suspended Phia, which would move the problem from online criticism into the commercial plumbing Phia needs to monetize partner sales.

What Phia still has to prove

Phia's advantage has always been distribution. Gates and Kianni turned personal brands, celebrity networks, press attention, and a clear consumer pitch into a venture-backed shopping product at unusual speed.

The current dispute puts a different metric in front of the business: trust. Merchants need to believe Phia is sending incremental customers. Affiliate partners need to believe Phia will not overwrite their work. Users need to believe the extension is acting as a shopping assistant rather than as an invisible attribution broker.

The cleanest path for Phia is also the hardest for any young consumer startup: show the rules. Phia can say it fixed the issue, and Bloomberg's follow-up test supports that the observed behavior stopped. The bigger commercial repair requires clarity about when Phia claims commission, when it stands down, how it handles existing referral codes, and whether retailers will receive refunds or adjustments for disputed commissions. TechCrunch's story does not establish how long the behavior was active, how many transactions were affected, how much commission Phia received, or which merchants bore the cost.

For Gates and Kianni, the allegation lands at a sensitive moment. Phia raised growth capital in January to hire machine learning talent and push toward what Gates called a more personalized, end-to-end shopping experience. The company is still early enough that a product correction can become an operating lesson rather than a permanent identity. The cost is that Phia now has to earn trust in the least forgiving part of online commerce: the attribution ledger.

Reader comments

Conversation for this story loads after sign-in.