Definition
A cross-device journey is one in which the ad interaction and the conversion happen on different devices. The guide to enhanced conversions puts the defining condition in one line: the click and the conversion happen on different devices, and only a signed-in identity can connect them. Without that identity, each device carries its own cookies and its own history, so measurement sees two short journeys rather than one long one.
The term describes the customer's behaviour, not a setting. It exists whether or not any tool can see it; what varies is how much of it each measurement system recovers.
Why paths fragment, and how platforms fill the gap
The attribution models framework lists it among the limits every model shares: models divide credit among tracked touchpoints only, and cross-device journeys fragment into separate paths unless the person is logged in. The tracked channels then absorb value they did not create. A visit typed in directly, or a branded search on the second device, collects the credit for demand an ad created on the first.
Ad platforms close part of the gap from their own side. The comparison of Google Ads and GA4 numbers notes that Google Ads extends what it can join with signed-in matching, enhanced conversions and offline imports, while GA4 does not use that matching. The guide to reconciling Meta with orders records that Meta matches people across devices through their accounts, which a shop's own source records cannot do. Where a platform cannot observe a conversion at all, it may estimate part of the total by modelling, which the guides ask readers to treat as an estimate.
In practice
Each measurement system sees a different share of these journeys, so reconciliation works step by step rather than by asking which number is right.
- Platform against analytics. The Google Ads and GA4 comparison removes cross-device conversions from the Google Ads figure before comparing, because GA4 has no counterpart for them.
- Platform against orders. The Meta reconciliation guide treats the shop's own count of orders from a tagged click as a floor, since it misses purchases completed on another device.
- Content and SEO reporting. The guide to measuring content performance calls every attribution figure a modelled estimate and lists journeys the tool cannot stitch among its known gaps, to be named in the report rather than hidden.
Enhanced conversions help most where journeys cross devices, because the hashed customer details sent at conversion give the platform another way to recognise the person. The gain is small on a site whose customers mostly convert on the device they clicked from.
A furniture retailer's direct traffic is not as direct as it looks
Olvrastine Furniture, an invented online furniture retailer, found that its analytics credited 61% of sofa sales to direct visits and only 9% to paid social. Its ad platform, which matches signed-in users, credited paid social with 23% of the same sales. A survey on the order confirmation page showed why: of 410 sofa buyers who answered, 188 said they had first seen the sofa on a phone and bought on a laptop or tablet, usually between 2 and 6 days later. The retailer stopped reading direct traffic as its strongest channel, added enhanced conversion data to its purchase tag, and began reporting the platform's figure and the analytics figure side by side, with a note explaining the gap.
Related terms
AttributionAssigning conversion credit to touchpoints. is the process that a cross-device journey breaks when only one device is visible. Enhanced ConversionsA Google Ads feature that sends hashed first-party customer data to improve conversion matching. send hashed first-party data at conversion so that more of these journeys can be matched. Modelled ConversionsConversions a platform estimates statistically for activity it could not observe directly, such as unconsented or cross-device journeys. are the platform's estimate for conversions it could not observe, including some that crossed devices. The Lookback WindowThe time after an ad interaction in which a conversion can still be credited to it. limits how long after the first-device interaction the second-device conversion can still be credited.
Common confusions
- Direct traffic means people found the site on their own. Some of it is the second half of a journey that began with an ad on another device.
- The platform is inflating its numbers. Signed-in matching is a structural difference, not an error; the platform and analytics are answering different questions.
- Cross-device is the same as multi-channel. A journey can cross channels on one device, or cross devices in one channel. Only a change of device makes it cross-device.
- Better tagging will recover all of it. Without a shared identity, a tag on each device cannot know the two visitors are one person; matching and modelling recover part of the gap, never all of it.
