The shopper who thinks she is being paid to buy something is, in a strict sense, working. Not for herself. For the retailer whose commission funds her rebate, and for the platform that sold that retailer a customer at a lower cost than Google or Meta could quote. Two dollars back on a sixty-dollar order is not a gift. It is a rebate on a marketing budget she helped the retailer avoid spending elsewhere.
This is the model Ebates built, sold to Rakuten for $1 billion in September 2014, and turned into one of the quieter power structures in online retail. Versions of it now run in most large e-commerce markets — Fanli in China, ShopBack across Southeast Asia. And almost nobody who uses it can explain how it works.
The actual mechanic is affiliate marketing. Retailers pay a commission to Rakuten each time a shopper arrives through a Rakuten link and completes a purchase, and Rakuten shares a slice of that commission back with the shopper as cash. FinanceBuzz’s breakdown of the model puts the partner network at more than 3,500 stores and total lifetime cashback paid at $3.6 billion across 21 million members. Rakuten Group, the Japanese parent, reported revenue of more than $15.5 billion in 2024.
Most shoppers never ask who is paying. The common assumption is that the retailer is discounting to bring them in the door, and that is close enough to correct that the platform has never needed to explain further.
The elegance is that everyone in the chain is doing something rational. The retailer would have spent that money on a Google search ad or a Meta placement. The shopper would have made the purchase anyway. The platform sits between the two and takes a cut for the introduction. Rakuten has, in effect, built a private ad network where the ads pay you.
What made the original Ebates pitch to advertisers unusual was the reversal of risk. Search and social ads charge per click or per impression. The retailer pays whether or not the shopper buys anything. Cashback platforms charge per completed sale. Joel Leong, who co-founded the Singapore-based cashback company ShopBack in 2014, put the merchant case directly to Options, The Edge: his company could always beat Google or Facebook on rates, because it charged on a cost-per-sale basis rather than per click. Merchants signed on, he said, because they carried no risk.
That is a substantially better deal than the one Google and Meta offer. It is also why the cashback model quietly ate a slice of digital ad budgets that never made headlines.
The shopper’s side of the deal is more complicated than it appears. The rebate is real. So is the surveillance. To route a purchase through Rakuten, the shopper has to log in through the portal or the browser extension, which means the platform sees what she searched for, what she clicked, what she abandoned, and what she eventually bought. Rakuten also invites members to link a credit card so that purchases in physical stores earn cash back too, which extends the record beyond the browser.
The trade is legible once you look at it. The shopper gets a small percentage back. The platform gets a granular map of her buying life, which is worth considerably more than the rebate over time.
A March 2026 explainer in The Conversation, by University of Newcastle law lecturer Mirella Atherton, made the point without softening it: cashback companies channel traffic to retailers in return for a commission, and they typically collect extensive data on their users’ online behaviour — searches, shopping history and more — while doing it. The rebate exists because the platform’s cut, plus the retailer’s remaining margin, still comes out ahead. If it did not, the model would collapse.
The individual returns are modest and slow to accumulate. The FinanceBuzz writer who explains the model, Danielle Letenyei, has been a member since 2018 and puts her own lifetime earnings at $991.39 — real money, spread across years of purchases she would largely have made anyway. What the platform accumulated over the same period is a profile of her consumption. The thousand dollars is not the price of that profile. It is the reason she keeps building it.
The retailer side is where the model has become genuinely powerful, and where it now competes with the giants. When a merchant weighs a cost-per-sale deal with a cashback platform against a cost-per-click campaign on Google or a placement on Meta, the cashback pitch reads as unusually clean. You pay when the shopper buys. You pay a known rate. You do not fight an algorithm about attribution.
That was always the ambition. Pay the shopper enough to get her to route through the portal, then sell that routed traffic to merchants as higher-intent than anything Facebook could deliver. TechCrunch, reporting the acquisition on the day it was announced, framed the deal as handing Japan’s largest e-commerce firm a fresh way into the growing American market. What Rakuten actually bought was a distribution channel that competed sideways with the largest ad platforms in the world.
The model has limits, and they are visible in the fine print. Rakuten pays out quarterly, not on demand. An account left untouched for twelve consecutive months starts accruing a $5 monthly maintenance fee until it is reactivated or drained to zero. Amazon, once among the retailers reachable through Ebates, ended the partnership in 2020, removing the single largest destination in American e-commerce from the network. And the rebates are generally treated as discounts rather than income for tax purposes, which is convenient for shoppers but also a reminder that what is being returned is a fraction of a fraction.
Travel is where the arithmetic gets most visible, because the base purchases are large. ABC News catalogued the rebates on offer for hotel and airline bookings in 2015 — a few percent on one booking site, six and a half on another, occasionally doubled during promotions — and the logic has not changed since. The rebate makes the booking feel affordable. It also steers the booking toward retailers inside the platform and away from the ones outside it, which is exactly the routing behaviour Rakuten sells to advertisers.
There is a competitor logic underneath all of this that deserves naming. Cashback platforms are one of the few consumer channels where the shopper knows, in dollar terms, exactly what her attention is worth. Two percent back on a hundred-dollar order is two dollars. The credit-card industry runs a similar accounting on a larger scale — CNBC Select’s roundup of premium cards lays out rewards and statement credits tied to named partners, from airline fee credits to monthly dining credits redeemable at specific delivery services.
The mechanism is identical. The scale is different. And the shopper who understands what she is being paid for, in either case, is a different kind of customer than the one who doesn’t.
What the Ebates model actually contributed was not cashback. It was the honest naming of a transaction that Google and Meta preferred to leave abstract. When you click on a search ad, a company is paying to have found you. When you click through a cashback portal, that same company is paying to have found you, and the platform is giving you a receipt for your part in the trade.
The receipt is the thing. Most digital economies work by hiding it.