Streaming TV and Amazon Search Lift: The Connection Worth Measuring

Why branded search is the right thing to watch

A streaming ad doesn't get clicked. What it can do is put a brand name in front of someone who, days later, opens the Amazon app and types that name into the search bar — an action Amazon can see and measure directly, on the same platform where the ad ran. That makes branded search fundamentally different from most other post-exposure signals: it's not modeled or inferred, it's a person's own deliberate action, tied to a specific brand, on a platform that already has the tools to measure it against a pre-campaign baseline.

This is also why branded search matters more here than it would for a channel that already produces a click. A search ad's own click is already the strongest available intent signal; adding branded search analysis on top of it teaches little that wasn't already visible. For a channel with no click at all, branded search is close to the best trackable proxy available for what the ad actually changed in a viewer's behavior.

It's also a signal that requires no extra pixel, no clean-room setup and no third-party tool to access — it's inside the same Amazon Ads reporting environment every advertiser already uses for sponsored ads. That accessibility is part of why it's such a poor sign when a streaming report skips it: the barrier to checking is genuinely low, which makes not checking a habit problem more than a tooling problem.

The evidence, from Amazon's own case study and our own accounts

Amazon's published case study on kitchen-appliance brand Fellow is the clearest documented example available: after expanding beyond online video into streaming TV — including FAST inventory through Private Auction deals, run across Prime Video's ad-supported inventory, Fire TV Channels, Twitch and third-party publishers — branded searches rose 500%, and 90% of streaming-influenced purchasers were new to the brand, a 75% higher new-to-brand rate than online video alone had produced.

Our own accounts show the same underlying mechanism, even without a single case study this dramatic to point to. In our own accounts, DSP prospecting audiences run around two-thirds new-to-brand, and the visible effect shows up first as growth in branded search on Amazon rather than in the DSP's own attributed-sales report. The practical consequence is real: existing sponsored-ads campaigns often get cheaper as more of the incoming traffic already knows the brand — a second-order effect that shows up on branded ACOS, not on the streaming line item's own dashboard, which is exactly why a brand reading only its DSP report will systematically undervalue what prospecting is actually doing.

A worked example

Pull branded search volume for the four weeks before a streaming flight starts as a baseline, then track it weekly through the flight and for at least two weeks after. On a $15,000 monthly Prime Video and Fire TV test, a meaningful branded search lift — even a fraction of Fellow's reported 500% — alongside a healthy completion rate is a stronger signal that the campaign is working than a same-week DSP ROAS figure, which was never built to capture this kind of influence. If branded ACOS on your existing sponsored-ads campaigns also compresses over the following month, that's the second-order confirmation — cheaper branded traffic because more of it already recognizes the brand.

What to do when search doesn't move

Give it longer than the first week or two before concluding the campaign isn't working. On one account we watched, organic order share rose from 13.5% before video to 27.9% after, and was still climbing four weeks after spend had actually stopped — one account, not a guarantee, but a reason not to judge lift on the first week or two alone. If search genuinely shows no movement after a full month at a real budget, check completion rate before anything else: weak completion means the creative isn't actually being watched, which would explain flat branded search far better than a targeting problem would. If completion rate is strong and search still hasn't moved, that's a legitimate signal worth taking seriously — not every campaign produces a Fellow-level result, and a flat search trend after real spend and real attention is meaningful information, not something to explain away.

The common mistake

The mistake is measuring a streaming campaign exclusively inside the DSP dashboard and never pulling a branded search report at all — the single most accessible, most directly attributable signal this format produces sits one report away and gets skipped because it isn't the default view. reMKTR runs 109 live Amazon DSP advertiser seats and pulls branded search alongside completion rate as a standard part of every one of those streaming reports, specifically because it's the metric that turns "the DSP dashboard looks unimpressive" into an actual answer about whether the campaign is working.