In our three-part blog series, we are focusing on the ongoing and upcoming data privacy updates that our industry is facing and sharing insights on how marketers can prepare. For the first and second blogs of this series, we discussed how marketers can adapt to these changes and leverage first-party data to gain valuable user insights. In this last blog, we are exploring how marketers can measure digital advertising effectiveness more accurately, despite the ongoing restrictions to third-party tracking.
For well over a decade, the method marketers have focused on for judging digital advertising effectiveness has been digital attribution — these are the set of variables used by advertisers to analyse the buying behaviour of online customers and ultimately, credit a conversion – or some part of it – to a specific marketing activity.
Caveats for biased models and missing data have long been barriers marketers have had to consider, but now we also have new, more complicated challenges to overcome. The third-party identifiers that track ad views have degraded and will soon be unavailable, so the future of digital attribution is unclear.
However, one thing is for sure – it’s no longer enough to judge effectiveness with digital attribution alone.
Responding to industry changes
New approaches to judging marketing effectiveness will be key in responding to how our industry is being shaken by developments in the tech and data we use. Previously we’ve focused on analysing data that tracks conversions against ad interactions. Now, we need to complement this approach with others that can prove the incremental conversion caused by our campaigns being active.
There are a number of reasons why marketers will want to adopt additional methods for judging digital advertising effectiveness, including:
- Channel management
- Regular performance reporting
- Creative and marketing strategy decisions
- Ad hoc reporting and impact assessment
- Budget planning
- Strategy
Essentially, we will need a range of overlapping techniques for collecting performance data and judging advertising effectiveness. The solution will be a balanced view of performance where two different measurements intersect. However, as we select the different methods we use to measure effectiveness, we must keep the following in mind:
- No single method can meet all of them
- Every single method has flaws and disadvantages
Judging advertising effectiveness with multiple methods
To meet these use cases, we’ll need a range of overlapping techniques for judging advertising effectiveness.
Digital attribution – Ad platform attribution data is important for driving automated bidding and optimisations. It’s vital for day-to-day channel management, and the data is very readily available for continuous reporting.
Controlled experiments – Automation has made manual testing less prominent in campaign optimisation, but a continuous programme of testing is now vital for calibrating and complementing digital attribution.
Causal effect analysis – Croud uses modelling to prove the impact the introduction of a new channel, campaign or tactic has on a key performance indicator (KPI), including where a split-test isn’t possible. Studies are run quickly and repeated regularly.
Marketing mix modelling – Marketing mix modelling attributes the performance to individual components of the marketing mix, including competitor activity and other market factors. These are big studies that inform strategy and media planning.
Enabling innovation and growth
Using the example of causal effect analysis, we can prove the incremental impact of upper-funnel media, like YouTube ads, on brand search, by comparing the ‘observed data’ with a modelled expectation.

We can also use media investment studies and marketing mix modelling to show channel performance and explore how to balance budgets across paid social, TV and display. If we put these factors together we have the necessary media planning insight or ‘evidence’ in order to grow.
Moving forward
Digital attribution has become unreliable, and it’s no longer enough to use it on its own for judging digital advertising effectiveness. Because of recent changes to the industry, we need a range of methodologies to meet all performance insight requirements. Measuring with more than one technique is the only way to avoid error and bias.
Each business has a different roadmap to overcoming their unique set of challenges. If you want to learn more about data solutions or would like to speak with someone on our team, please get in touch.