Retail partners and finance teams usually share the same immediate concern: if you expand onto Amazon, who loses? The anxiety is understandable. Amazon is where shoppers go to compare options, while your direct-to-consumer site validates your brand. However, overlap does not automatically equal cannibalization and you can model the actual impact long before committing budget.
Why every retail partner asks this first
Shoppers no longer take one path. Our Digital Flagship research puts the average at nearly six touchpoints before a purchase, and the brand’s website is almost always one of them. Channel overlap is already a reality. What truly counts is whether each channel fulfills a distinct function. While retail partners focus on pricing and shelf presence, your D2C team prioritizes customer retention. Rather than debating these competing concerns, address them both by modeling the overlap.
Build the forecast from your D2C data, then pressure test it
Start with what you own. D2C branded search volume, repeat-purchase rate and email-driven orders show how much demand your brand already creates. That demand is the most exposed to leakage, because a shopper who searches your name has already been paid for once. If she then buys through a branded Amazon ad, you pay a second time for the same sale (we cover this in Part 1 of our 2027 planning series). Build a base case where a modest share of that demand moves to Amazon, then a worst case where double that moves. If the plan survives the worst case, the questions stop.
New-to-brand rate and channel migration: the two numbers that matter
New-to-brand (NTB) rate shows how much Amazon revenue comes from shoppers who never bought from you. Channel migration shows how much comes from existing customers changing where they buy. High NTB with low migration is growth. The reverse is cannibalization. Amazon Marketing Cloud can separate what your media drove from what would have converted anyway.
Branded spend is where the two get confused. OCTOBUDDY steered branded searches toward its strongest organic listings, cutting branded ad spend by more than 30%, holding 50% branded impression share and doubling media profitability. Same demand, less paying twice for it. (Read the case)
How to prove incrementality when you don’t run promotions
No promotions means no sales spike to point to, so use baselines. Track branded search and D2C traffic before and after launch, compare against category-matched periods, and measure the halo from off-Amazon media on Amazon rank and conversion. Incrementality shows up as total brand revenue rising while D2C and retail hold steady, not in any single channel’s report. Our closed-loop attribution work is built for exactly this view.
Assortment and bundles that differ from retail
The cleanest defense is not competing with yourself. Scrub Daddy built Amazon-exclusive bundles and configurations for how people shop there, and grew Amazon sales 24% and average order volume 23% without undercutting its retail partners. (Read the case)
Keep what Amazon can’t carry on your own site: loyalty, subscriptions, storytelling. Polaroid rebuilt its flagship to reclaim its audience from the marketplace with community, education and membership, things a product page can’t offer. (Read the case)
What a quiet first-year launch looks like
A good launch is gradual. NTB share builds first, branded overlap stays small because you manage it, and retail sell-through holds. Watch four numbers monthly: D2C branded sessions, repeat rate, Amazon NTB rate and retail sell-through. Agree the thresholds with your retail partner before launch, so a surprise in month three is a conversation, not a crisis.
Want to pressure test your own numbers? Talk to our team about building the model with you.
