Cross-Device and Household Targeting | TV to Phone Explained
How cross-device and household targeting connect a television ad to the phone in your hand, what identity graphs actually match, and where the method breaks.
Key takeaways
- Most television to phone matching works from a shared household IP address, not from audio.
- Household targeting treats every device on one network as a single addressable unit.
- Identity graphs add logged-in accounts and observed behaviour on top of the IP signal.
- The method is probabilistic, so a share of matches are wrong in ways the report will not show.
What is cross-device targeting?
Cross-device targeting is the practice of reaching the same person, or the same household, across more than one screen by connecting the devices to a shared identifier. It is what allows a brand seen on a living room television to appear later in a phone feed.
The experience feels uncanny. The mechanism is fairly mundane.
How does the television get connected to the phone?
Three methods do most of the work, usually layered.
| Method | What it uses | How reliable |
|---|---|---|
| IP matching | Shared home network address | Common, loose, breaks on shared or public networks |
| Deterministic matching | A logged-in account on both devices | Strong where a login exists on both |
| Probabilistic matching | Observed behaviour patterns, device type, timing | Weakest, used to fill gaps |
IP matching is the backbone. When your streaming stick requests an ad, it reports the network address. Your phone on the same wifi reports the same address. An advertiser buying connected TV can then target devices on that network for a follow-up impression.
Deterministic matching is more accurate and less available. It requires the same account signed in on both devices, which is why platforms with universal logins have an advantage here.
What is household targeting and why does it matter?
Household targeting addresses every device on one network as a single unit. It is the default for connected TV, because a television is genuinely shared and there is no way to know which family member is watching.
That has two consequences worth understanding. Reach figures describe households rather than people, so a campaign claiming reach into 100,000 households did not reach 100,000 individuals. And frequency capping applied at household level means one heavy viewer can absorb the whole cap while others in the home see nothing.
Where does the method break?
More often than the reporting suggests.
- Shared networks. Apartment buildings, offices, cafes and university accommodation put unrelated people behind one address.
- Mobile data. A phone off wifi reports a carrier address and the household link disappears.
- VPNs. The signal points somewhere the user is not.
- Multiple households, one person. A commuter with two homes is two identities.
None of these show up as errors. They show up as matched impressions that were served to the wrong person and counted as success.
How does this affect measurement?
It creates the view-through crediting problem. If a household saw a television ad and somebody in that household later converted, the platform can claim the conversion without any interaction taking place. Widen the lookback window and the number of claimed conversions rises without any change in the campaign.
Google’s guidance on understanding conversion tracking data covers how cross-device conversions are counted and reported. This is why we plan cross-device work on delivery quality and prove the business impact with holdout testing rather than accepting the platform’s own account of what it caused. The same logic applies to standard retargeting, where the audience was already close to converting before any budget was spent following them.
What this means for advertisers
Cross-device and household targeting are genuinely useful for sequencing a message across screens, and genuinely unreliable as a basis for claiming credit. Treat the matching as a way to deliver a coordinated campaign, and treat the attribution it produces as directional.
If you are buying connected TV and reading its conversion report as proof, read our note on CTV attribution next.
FAQs
- How does a brand show up on my phone after I saw its TV ad? The streaming device and your phone usually share a home IP address. An advertiser can target the devices seen on that network, so an ad served to the television can be followed by an ad served to a phone in the same household.
- Is this the same as my phone listening to me? No. The match is made from network and device data, not audio. A shared IP address, a logged-in account, or a device graph built from observed activity is enough to connect a television to a phone without any microphone involved.
- What is household targeting? Targeting every device on one home network as a single unit rather than targeting individuals. It is how connected TV buying works, and it means a four-person household is treated as one addressable target.
- Can I stop cross-device tracking? Device and platform privacy settings, ad identifier resets, and limiting ad tracking all reduce it. A VPN changes the IP signal. None of it removes cross-device matching completely, because several independent methods exist.
- Why do reach figures overstate how many people saw an ad? Household targeting counts networks, not people. A campaign reporting reach into 100,000 households did not reach 100,000 individuals, and one heavy viewer can absorb the whole frequency cap.
PUT THIS KNOWLEDGE TO WORK
NEED MORE HELP?
Talk with our team about applying Cross-Device Targeting to your marketing. Request a complimentary audit to identify practical next steps.
REQUEST A COMPLIMENTARY AUDIT