Invest1 publisher2 min readPublished
Amazon multiplies its shoppable Prime Video catalogue sixfold in a year
Shop the Show now covers more than 8,000 Prime Video titles and the new Shop the Scene runs on 600 of them, while the conversion evidence is a YouGov survey in which 0.4% of US adults bought after a placement.
The Investor · Invest desk

What happened
- Amazon said on Thursday that a new Prime Video feature, Shop the Scene, lets viewers pull up items inspired by the scene they are watching with minimal interruption to playback.
- Shop the Show, launched last year for merchandise such as bobbleheads and LEGO sets, is going from 1,300 titles to more than 8,000.
- A new shop tab inside X-Ray lets viewers open product listings with the remote, browse what is on screen, and complete the purchase on a phone.
- Peacock launched a Must ShopTV feature in 2023 and Disney+ has tested shoppable pages reached by a QR code on the television screen.
Compiled by The InvestorSomething wrong?How this is made
Why it matters
- capability Because the matching happens after a show is finished, in the same way as Amazon's virtual product placement work, coverage can widen title by title without anyone buying a placement into a script.
- exposure Amazon's model decides what a viewer is offered at the moment of intent, and Amazon says the match may not always be exact, so the brand on screen can lose the sale to a lookalike listing.
- constraint The sale still completes on a second device, so Amazon's conversion depends on a viewer picking up a phone mid-episode while the television stays a browsing surface.
- decision A brand weighing whether to fund tagged placements has to price them off a third-party survey, because the purchase figures come from YouGov.
Amazon added roughly 6,700 titles to Shop the Show in about a year, leaving a catalogue about 6.2 times the size of the one it started with [13]. Amazon did not describe new production deals or reshoots. Shop the Scene matches on-screen items using AI, and Amazon says the match may not always be exact [5]. It is the same after-the-fact method as the virtual product placement technology Amazon has run for years, which uses machine learning to insert advertisers' brands into films and shows that have already been made [8].
Scene-level matching covers 7.5% of that shoppable catalogue [14], and only for users in the US [2]. "We are making it easier than ever for Prime Video customers to shop what they see on screen," Michelle Rothman, vice president of Prime Video shopping, said in Thursday's announcement [4].
The purchase evidence comes from outside Amazon. A YouGov survey earlier this year found that just over half of US adults consider product placement an effective form of advertising, against 14% who called it ineffective, Fortune reported [9]. In the same survey, 4% said they took action after seeing a brand in content they watched, and of those, 22% searched for the product online and 10% made a purchase [10]. Multiply through and 0.4% of the adults surveyed bought something, while 0.88% searched [15]. Applied to the ad-supported Prime Video reach Amazon disclosed late last year, 0.4% works out at about 1.26 million buyers [16], and the two numbers do not describe the same population: the survey asked US adults, the reach figure is worldwide [7].
Maybe tagging is cheap, scene-level matching climbs toward the full catalogue, and Amazon gains a second demand channel feeding its retail business. Or coverage sticks near where it is, because clothing rights and match quality bind harder than compute does. Or the money never comes from the shopper and arrives instead as advertiser payments for the placement itself, which is what the virtual placement technology was built to sell [8].
I lean to the third. A 0.4% purchase rate drawn from a survey [15] is a thin foundation for a retail channel and a serviceable one for selling attention to brands, particularly when the seller also owns the matching model that decides which product a viewer is shown [5].
What would move me off that is a first-party figure: an attach rate from Amazon, or a brand reporting sales lift on a tagged title, well above the survey's rate. Amazon did not immediately respond to Fortune's request for comment [12].
What to watch
- Whether scene-level coverage closes on the full 8,000-title catalogue over the next year or stalls near 600. The answer separates a compute constraint from a rights one.
- Any first-party disclosure from Amazon of an attach rate, shoppable-video revenue, or advertiser pricing for virtual product placement inventory.
- Whether Peacock or Disney+ revive their shoppable experiments now that Amazon is routing the purchase into its own shopping app.