A TV campaign goes live and, within minutes, brand searches increase, website traffic rises and Brand Search campaigns record more activity. GA4 measures these digital behaviours, but can it identify the TV campaign as the source of this momentum? Not necessarily.
This is one of the fundamental limitations of attribution: what is measured and attributed is not always equivalent to what actually influenced the consumer. This distinction is essential when analysing a media mix combining online and offline channels whose roles differ throughout the funnel.
The challenge is therefore to distinguish between the media channels that create or stimulate demand, those that capture it and what analytics tools are actually able to attribute.
GA4 measures observable interactions, not the full extent of media influence
GA4 is a key tool for analysing acquisition, on site behaviour and conversions. Its Data Driven Attribution model distributes conversion credit across different observable touchpoints rather than relying solely on a last click approach.
However, this approach remains dependent on the signals available. A significant share of advertising influence can occur before any identifiable digital interaction takes place.
This is particularly evident with offline media. TV, Out of Home, DOOH, radio and print can build awareness, memorisation and intent without generating a touchpoint that GA4 can directly use. When a consumer exposed to a campaign later searches for the brand on Google or visits its website directly, GA4 observes their digital arrival but does not necessarily have the signal required to connect it to the media exposure that preceded it.
The same phenomenon exists in digital environments. Exposure to video, Social Media or Display advertising can influence a user without generating a click, before they later return through Search, Organic or Direct.
The absence of attribution to a media channel does not therefore necessarily mean the absence of contribution.
From immediate effects to journeys lasting several months
Timing adds another layer of complexity. The effect of a campaign can be almost instantaneous. A TV spot, for example, may generate a spike in brand searches or homepage traffic during or immediately after it airs.
Conversely, for products or services requiring greater consideration, the conversion may occur several weeks or even several months after the first exposures.
The consumer journey can therefore develop across different timeframes, media channels and devices. The more fragmented it becomes, the more difficult it is to precisely connect all the influences involved to an individual conversion.
The main biases to consider when analysing performance
Beyond offline media, several phenomena reduce the level of visibility available in analytics tools:
Measurement bias | In practice | Impact on analysis |
|---|---|---|
Exposure without a click or digital interaction | A user may be exposed to a TV, Out of Home, video, Social Media or Display campaign without generating a directly trackable interaction. | The media channel may contribute to awareness and intent without clearly appearing in the attributed journey. |
Cross device behaviour and loss of identifiers | A user discovers a brand on a smartphone and later continues their journey or converts on another device. | When available signals cannot reconcile these interactions, part of the journey may be lost or modelled. |
Consent and modelling | Consent related restrictions reduce the amount of directly observable data. | Reports may combine observed and modelled data, with varying levels of visibility depending on the journey. |
These limitations do not make GA4 data less useful. They simply define the context within which it needs to be interpreted.
Creating demand is not the same as capturing it
This distinction is particularly important when analysing Search.
When a user actively searches for a brand, their level of intent is already high. Search, SEO or Direct traffic often intervene at the stage where existing demand is being captured. But that demand has not necessarily appeared spontaneously. It may have been created or strengthened earlier by TV, Out of Home, video, Social Media or Display advertising.
The journey may therefore look like this:
TV / Out of Home / video → exposure and memorisation → brand search → Search → conversion
In GA4, Search will be particularly visible because it intervenes at a point where intent is high and the interaction can be measured. The media channel that helped create that intent may, however, be absent from the attributed journey.
Awareness media contribute to creating and nurturing demand, while demand capture channels make it possible to turn that demand into a visit and, potentially, a conversion.
Comparing these channels exclusively on the basis of the conversions attributed to them therefore means comparing media channels that perform different functions within the funnel.
A concrete example: when attribution can shape interpretation
Let us take a deliberately simplified example. A brand launches a YouTube campaign designed to increase awareness.
Channel | Before campaign | During campaign | Change |
|---|---|---|---|
Search | 80 | 110 conversions | +30 |
Direct | 20 | 60 conversions | +40 |
Attributed YouTube | – | 15 conversions | +15 |
Total | 100 | 185 | +85 |
An interpretation focused exclusively on attribution could lead to the conclusion that YouTube played only a marginal role, with just 15 attributed conversions, while Search and Direct appear significantly more effective.
Une lecture média pose une question supplémentaire : la campagne a-t-elle contribué à l’augmentation de la demande ensuite captée par Search et Direct ?
The simultaneous increase across these channels is an interesting signal, but it is not proof of causality. Seasonality, promotions, SEO, competition or other campaigns may also influence results. These variations must therefore be compared with campaign periods and other indicators before drawing conclusions.
This is precisely the difference between attribution and incrementality. Attribution distributes credit across observed interactions, whereas incrementality seeks to determine what the campaign actually generated in addition to what would have happened without it.
At Mediamix, we analyse the role of media beyond attribution
This perspective has a direct impact on the way we evaluate campaigns: media channels cannot be compared solely on the basis of the conversions GA4 attributes to them.
An awareness channel and a Brand Search campaign do not operate at the same stage of the funnel and do not serve the same objective. Their performance therefore needs to be interpreted according to their role within the media mix and the context in which they are activated.
In our analyses, we put GA4 data into perspective alongside other available signals: changes in Direct traffic, brand searches, Brand Search performance, media platform data and variations observed during campaign periods. The objective is not to artificially attribute these changes to a specific media channel, but to identify consistent signals that enrich the analysis.
To go further, other measurement methodologies can complement attribution, including Brand Lift studies, incrementality testing and Marketing Mix Modeling, or MMM. MMM makes it possible to statistically analyse the contribution of different marketing channels to overall performance without relying on the reconstruction of individual user journeys in GA4.
An increase in Brand Search during an awareness campaign, for example, should not automatically be interpreted as an independent improvement in Search performance. It may also reflect stronger demand generated earlier in the journey. Conversely, a media channel with few directly attributed conversions should not be considered ineffective on that basis alone.
This is why we favour a holistic view of performance in which attribution is one indicator among others rather than a comprehensive measure of the contribution of each media channel.
From attribution to measuring media contribution
GA4 remains an essential component of the measurement ecosystem, but attribution alone cannot fully reflect the contribution of an entire media mix.
The question “Which channel obtained the conversion?” must therefore be complemented by a second one: “Which media channels helped create the attention, intent and demand that made this conversion possible?”
This is the perspective we favour at Mediamix: interpreting GA4 data within its media context, considering the role of each channel throughout the funnel and combining the available signals rather than comparing channels solely on the basis of attributed conversions.
The objective is not to artificially attribute every conversion to the entire media mix, but to develop a sufficiently comprehensive understanding to make better media investment decisions.