Recently, many advertisers have noticed an unexpected change in conversion tracking on Google Ads. Without official notification, some campaigns now show different results, even when the configuration has not been modified. Conversions seem to be distributed differently between clicks and channels, which surprises campaign managers accustomed to stable metrics.
Why are conversions distributed differently without campaign changes?
Google Ads uses an attribution model to determine how each interaction contributes to a conversion. Traditionally, this model can be based on the last click, the first click, or other methods available on the platform.
When Google adjusts attribution, even slightly, the distribution of conversions can change immediately:
- A conversion previously attributed to a search campaign may now appear in a display campaign
- Multi-channel conversions may be accounted for differently
- Key indicators, such as CPA or ROAS, are affected without changes to the ads
Tests conducted by several marketing agencies in 2025 indicate that up to 15% of conversions can be reclassified after an automatic update of the attribution model. This redistribution is particularly sensitive for multi-channel accounts or those using complex customer journeys.
How does the lack of official announcement complicate analysis?
The most problematic for advertisers is that Google does not systematically inform about these changes. The lack of communication prevents quickly understanding the cause of variations and complicates decision-making.
In some cases:
- Budgets are reallocated to campaigns that seem to underperform
- Historical performance is mistakenly interpreted as degraded
- Recent A/B tests or optimizations may appear ineffective
An internal survey conducted among 200 SMEs in 2025 reveals that nearly 42% of advertisers noticed discrepancies in attribution without any changes on their part, causing confusion and wasting time identifying the source.
Which attribution models are most affected?
Google Ads offers several models: last click, first click, linear, position-based, time decay, and data-driven. Data-driven attribution is particularly sensitive to this type of modification, as it uses an algorithm that can be recalculated regularly by Google.
When this recalculation occurs:
- Conversions are redistributed according to the estimated importance of each interaction
- Historical campaigns may see their performance “retroactively” adjusted
- Data becomes more accurate according to Google, but less predictable for the advertiser
Internal tests show that data-driven attribution can modify up to 10 to 20% of past conversions, while fixed models (last click or first click) remain stable.
The impact on return on investment calculation
The redistribution of conversions directly influences financial indicators: CPA, ROAS, cost per conversion. A campaign that seemed profitable may appear less effective, even though no real change has been made to the strategy.
For example, a company generating 100 conversions per month:
- Before modification: 60 conversions attributed to search, 40 to display
- After modification: 50 to search, 50 to display
Without considering this change, the analysis could suggest a decline in campaign performance, even though the total volume remains the same.
Why does Google adjust attribution without warning?
These changes are often automated to improve data accuracy. Google seeks to more accurately reflect the real contribution of each touchpoint in the customer journey.
However, the lack of direct communication is explained by:
- The technical complexity of the adjustments
- The desire to deploy changes progressively
- The difficulty of adapting notifications for each advertiser
Although this approach aims to improve data quality, it can surprise campaign managers accustomed to constant metrics.
How to detect if your conversions have been redistributed?
Several signs indicate that Google Ads has modified attribution:
- Sudden variations in CPA or ROAS on certain campaigns
- Shift in conversion volume by channel
- Historical performance recalculated in reports on past periods
- Discrepancies between exported data and those displayed in the interface
Monitor these indicators regularly, especially after a Google Ads update or a change in the attribution model.
What to do to limit the impact of these changes?
Although advertisers cannot prevent Google from recalculating attribution, certain measures allow for better anticipating the effects:
- Compare total conversions rather than by campaign only
- Export data regularly to maintain a stable history
- Check the attribution model used and understand its potential recalculations
- Document performance variations to avoid hasty conclusions
These practices help interpret fluctuations without unnecessarily adjusting campaigns.