Nobody clicks a TV ad. So how do you tell whether connected TV is working? Look at what people do after they see it.
The problem with clicks
Connected TV (streaming) ads run on a television, where there is nothing to click. Click-based attribution gives them no credit at all, even when they work. Platforms such as MNTN report “verified visits”, matching households that saw the ad to visits to your website. That is useful, but it is the platform grading its own homework.
Brand-search lift
People who see a TV ad and are curious usually search for your brand. That leaves a trail in data you already own: branded searches in Google Search Console, brand keyword impressions in Google Ads, and Bing Webmaster Tools as a cross-check.
The method is simple:
- List your brand terms, including common misspellings.
- Mark the flights: the days the TV ads ran.
- Set a baseline from the weeks before each flight, by day of the week, leaving out other flights.
- Compare branded searches during and just after each flight with the baseline.
- Use a control. Non-brand searches on the same site show whether everything rose (a season, a site change) or only your brand did.
If branded searches rise during flights and the control doesn’t, TV is very likely the reason.
What it doesn’t tell you
Brand-search lift shows interest, not sales, and it misses people who come straight to your site or buy in a store. Treat it as one honest signal alongside the platform’s own reports and your sales.
In Metric Sensei
The TV impact report in Attribution does all five steps: brand terms and flights saved per client, or flights read from MNTN with daily impressions, and searches from Search Console, Google Ads and Bing. The AI assistant can run it too. See Attribution.