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
- Television-to-phone matching can use household IP addresses, device graphs, logged-in data, and modeled signals. It does not require audio surveillance.
- Household targeting can group devices inferred to belong to one household as an addressable unit.
- Identity graphs add logged-in accounts and observed behaviour on top of the IP signal.
- Probabilistic matches are inferences, so aggregate reporting may not expose individual matching errors.
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 common methods may be used separately or together.
| 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 | Inferred rather than confirmed, often used to fill gaps |
IP matching is one common signal. When a streaming device requests an ad, the platform can observe its network address. A phone on the same wifi may present the same address. Depending on the platform, region, consent, and campaign setup, an advertiser buying connected TV may be able to reach other devices associated with that household.
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 groups devices that a platform infers belong to one household. It is common in connected TV because televisions may be shared and person-level identity or attention is not always observable.
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?
The method has several known limitations.
- 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.
These conditions may create missing or incorrect matches that an aggregate campaign report does not expose as individual errors.
How does this affect measurement?
It creates a view-through crediting question. If a CTV impression was recorded for a household and somebody associated with that household later converted, the platform may credit the impression without a click. Widening an applicable lookback window makes more prior impressions eligible for credit, but it does not guarantee that the number of reported conversions will rise.
Google’s guidance on connected TV measurement says household measurement can incorporate IP addresses among other signals. Its guidance on modeled conversions describes additional aggregated signals and modeling. This is why we plan cross-device work around delivery quality and use a suitable incrementality design where scale allows instead of treating platform attribution as causal proof. The same logic applies to standard retargeting, where an audience may already be close to converting before more budget is spent reaching it.
What this means for advertisers
Cross-device and household targeting can sequence a message across screens, but a device match alone does not establish causal credit. Treat the matching as a delivery method and evaluate the resulting attribution under the platform’s stated rules and limitations.
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? A platform may associate the television and phone through a household IP address, a logged-in account, a device graph, or modeled signals. The available methods and targeting rules vary by platform and region.
- Is this the same as my phone listening to me? No audio is required for this type of matching. Network, account, device, and modeled signals can associate devices without using a microphone.
- What is household targeting? Grouping devices that a platform infers belong to one household for targeting or measurement. It is one approach used in connected TV and other cross-device campaigns, with implementation and availability varying by platform and region.
- Can I stop cross-device tracking? Device and platform privacy settings, ad identifier resets, limiting ad tracking, and network choices can reduce available signals. Their effect varies because platforms may use more than one permitted matching method.
- Why is household reach different from people reach? Household reach estimates households, not individual viewers. A report of 100,000 reached households should not be read as 100,000 confirmed people who paid attention, and household-level frequency does not show how impressions were distributed among viewers in each home.
PUT THIS KNOWLEDGE TO WORK
NEED MORE HELP?
Talk with our team about applying Cross-Device Targeting to your marketing.
Get a free marketing audit call