Case 1: many customers, one ID, and the mystery of the missing sales
The client was a fast-growing ecommerce company facing a tricky problem: the number of purchase events in GA4 reports was far lower than the actual number of orders recorded in the company’s back end, sometimes by a factor of several. Not only could they not assess the performance of their marketing campaigns accurately, but the team’s trust in GA4 data had hit rock bottom. The root of the problem was an error that is very common during User-ID implementation, and has serious consequences.
Root cause analysis
After a thorough investigation, we found that the client’s engineering team had made a critical mistake when implementing User-ID tracking: they were sending the same static user_id value for several different logged-in members.
The solution and how it works
To give the marketing team relatively accurate data straight away, we recommended a stopgap that proved very effective: temporarily switching Reporting identity from “Blended” to “Device-based”.
This works because of how the device ID mode is defined: it completely ignores any user_id sent in and relies only on the cookie unique to each browser (the Client ID) to count users and events.
Because each real customer placed their order on their own device or browser, each had a unique Client ID. After switching to this mode, GA4 was no longer misled by the incorrect User-ID and could attribute each order to a separate device in a way much closer to reality. Shortly after the setting was changed, the number of purchase events in GA4 reports lined up almost perfectly with the number of orders in the back end.
Case 2: user data imported, but the reports stayed blank
The second client’s situation was the exact opposite. They had invested heavily and successfully implemented accurate User-ID tracking for their membership system, aiming to analyse the cross-device behaviour of high-value members in depth. However, after they imported user segment data and the officially stated processing time had passed, the segment data still didn’t appear in reports.
Root cause analysis
The crux of the problem was very simple, yet easy to overlook: after completing the complex technical deployment of User-ID, the client had forgotten to update GA4’s Reporting identity setting, which was still set to “Device-based”.
The solution and how it works
The solution was very simple: change Reporting identity from “Device-based” to “Blended”.
This simple switch was like pressing an “enable” button. GA4’s reporting system received a new instruction and started prioritising user_id to combine and identify users, linking the same member’s behaviour in the mobile app and on the desktop website. When the client looked at the reports again, members’ cross-device journey data was finally displayed correctly, and they could start uncovering the business insights they had long been waiting for.
Strategy guide: how do you choose the right Reporting identity for your business?
Taking both cases together, we can draw one core conclusion: there is no absolutely good or bad Reporting identity setting, only one that fits or doesn’t. This section offers a clear decision framework to help you make the wisest choice based on your business needs and data maturity.
- Scenario 1: my website has a member login system, and cross-device user analysis matters to us.
- Recommended setting: Blended or Observed.
- Analysis: these two options make the most of the User-ID you’ve implemented, giving the most accurate and complete picture of user profiles and behaviour paths. This is the foundation of truly person-based analysis. However, it depends on the User-ID implementation being completely correct. Otherwise, the data disaster of Case 1 will repeat itself.
- Scenario 2: my website has no membership system, or is mainly a content or brand website.
- Recommended setting: Device-based.
- Analysis: without a User-ID to use, this is the simplest, most direct baseline setting. Although it overestimates the number of unique users, it is the simplest setting. Note that Google signals has not been part of Reporting identity since February 2024, so choosing Device-based does not by itself avoid data thresholds. For websites with lower traffic or simpler analysis needs, this is usually the best choice.
Conclusion: getting data to tell the truth starts with the right settings
Reporting identity is a core setting that affects the accuracy and depth of GA4 data. It isn’t a set-and-forget option. It should be treated as a strategic tool that is adjusted according to your business model, technical implementation and analysis goals.
Data itself doesn’t lie, but the wrong configuration can make it send misleading messages. Take action now.
In the digital age, GA4 isn’t just a tool for large enterprises. Want to be part of this wave of data? Get in touch.
