Whether TikTok Shop steals sales from your own site, what the incrementality data actually shows, and how to test it on your own brand in 30 days.
TikTok Shop usually adds sales rather than moving them. Haus analyzed hundreds of TikTok incrementality experiments and found TikTok delivers 1.9x more value when brands measure omnichannel impact instead of DTC lift alone, with 62% of the lift going to new customers. The cannibalization brands think they see is mostly an attribution artifact, because a site dashboard cannot see a sale that closed inside the TikTok app.
Does TikTok Shop cannibalize your direct-to-consumer sales? This is the first question most DTC operators ask before they open a shop, and the honest answer depends on evidence rather than dashboards.
Cannibalization is a specific claim, not a feeling. It means a customer who was already going to buy from you at full price on your own site instead bought the same item on TikTok Shop, so you paid a referral fee and a creator commission for revenue you already had.
For that to be true, three things have to hold at once:
Most brands never check the second and third conditions. They see site revenue dip in a month when TikTok Shop revenue climbed, and they call it cannibalization. That is a correlation with at least four other explanations, including seasonality, an ad account change, and a paid-social CPM swing.
Incrementality testing is the only method that answers this cleanly, because it holds a control group out of exposure and measures the difference. Attribution models cannot do it, since they only see the conversions they were able to tag.
Haus, an incrementality measurement platform, published an analysis in December 2025 covering hundreds of TikTok incrementality experiments across 2.5 years of data, with an average experiment length of 21 days. Three findings matter for this question:
That last number is the direct answer to the cannibalization worry. If most of the lift comes from people who were not already your customers, the channel is doing acquisition, not reshuffling.
One honest caveat: this analysis covers TikTok advertising broadly, not TikTok Shop creator affiliate programs specifically. It is the best public incrementality dataset on the platform, and the direction is consistent with what TikTok Shop operators report, but it is not a controlled study of affiliate GMV.
Three measurement problems produce a false cannibalization signal, and all three are structural rather than fixable with a better UTM convention.
The sale closes where you cannot see it. A shopper watches a creator video, taps the product card, and checks out inside TikTok. Your site analytics never records a session. From the site's point of view, that customer simply stopped showing up.
The delayed effect lands outside the reporting window. Haus found brands observed an additional 68% lift to their primary KPI during the post-treatment window, the period after the experiment ended. Creator content builds demand that converts later, often on a different surface. A 7-day attribution window systematically undercounts it.
Your channels report in separate systems. Shopify reports site revenue, Seller Central reports Amazon, Seller Center reports TikTok Shop. Nothing joins them, so the only view anyone has is three lines that move independently, which invites a causal story that the data does not support.
Hubfluence connects your Amazon data to your TikTok Shop program with Halo Effect Tracking, so creator activity gets measured against the channels it lifts instead of only the one it closes in. Book a call and we will read it against your numbers.
You do not need a measurement vendor to get a first read. You need a clean before-and-after on total revenue, not channel revenue.
The cleanest version of this is a geo holdout: run creator content in some regions and not others, then compare total revenue between them. That takes more discipline, and it is what an incrementality vendor sells, but the logic is available to anyone with regional sales data.
It happens, and pretending otherwise is not useful. Three situations where a TikTok Shop sale is likely a moved sale rather than a new one:
The fix for the first one is straightforward: send owned audiences to your owned store, and let TikTok Shop serve the traffic that TikTok generates.
The commercial version of this question is not academic. If your finance team believes TikTok Shop cannibalizes, the creator budget gets cut at the first soft month, and you lose the channel before it compounds.
The defense is measurement that spans channels. A creator program that only reports TikTok Shop GMV can never answer whether it lifted Amazon or retail, so it always looks more expensive than it is. That is the same structural blind spot that makes paid social look worse than it performs.
For agencies, this is the single most common reason a TikTok Shop retainer gets questioned in month four. The client compares site revenue to last year and asks what the creator spend bought. Being able to show total revenue, new-customer share, and cross-channel movement changes that conversation from a defense into a report.
There is also a planning consequence. If 62% of the lift is new customers and a third of the value shows up on Amazon, then your TikTok Shop creator calendar should be timed against Amazon inventory and retail promotions, not run in isolation. Brands that coordinate the two get paid twice for the same content.
Hubfluence tracks creator activity against the channels it actually lifts, including Amazon through Halo Effect Tracking, so a creator program is judged on total revenue movement rather than the one checkout it happened to close in. If you want to see whether your TikTok Shop revenue is additive or reshuffled, book a strategy call and we will read it against your own numbers.
From outreach to GMV reporting, Hubfluence runs every part of your creator campaigns for agencies and enterprise brands. Set it up once, scale it across every brand you manage.
Creator Discovery
Campaign Management
Social Intelligence
Reporting & Analytics
