Actioning media attribution, or why only half your marketing works

Media attribution is both a business and an analytical problem that sits right in the centre of the modern marketing paradigm of data, technology, and creative. Often overlooked in this discussion is the who and how of actioning the insights generated.

Last week at an event on programmatic, attribution was the most talked about topic behind fraud and viewability, and attribution even has a role to play in identifying those. Datalicious has spoken about media attribution at length and it’s exciting to see that word being used more regularly in discussions around the future of marketing.

Through the huge growth in personal and mobile technology use, the customer journey has only grown more complex, with touchpoints happening through many different marketing channels. This has created opportunities for marketers to use attribution in order to track customer journey behaviour in ways never possible in the offline world.

People are engaging with brands in a growing number of ways to choose, research and purchase products; and a greater proportion of these micro-moments are being captured. In spite of this, marketers are still struggling to answer the enduring question: How do I effectively allocate my marketing budget?

Media attribution is how we help clients answer that question — to find the best and worst performing channels and shift the budget to optimise marketing effectiveness.

Business decisions need to be based on sound reasoning and assumptions before you start to change things. Inferior, irrelevant or inaccurate models could actually end up leading to spend decisions that create more waste and inefficiency. The models used historically to assign credit to the touch points in the purchase path — first click, last click, even allocation, time-decay and the rest are no longer acceptable to the majority of marketers.

A recent research paper we ran with Econsultancy, The State of Marketing Attribution in Asia Pacific,found that custom modelling is believed to be the most effective form of attribution, the next best being Data Driven Attribution, an algorithmic model offered as part of Google Analytics Premium.

Once the entire team is agreed on which model to progress with, actioning attribution insights is still a challenge. Whilst the models enable them to see the potential benefits of changing things, see the benefits of changing things, the reality of actioning is usually more complicated.

Among the marketers that are actioning attribution insights, research shows that the biggest direction of change has been an increased spend in digital channels.

We usually see is display advertising being severely undervalued and so once this imbalance in reporting is corrected, there is a stronger business case to shift budget to that channel.

But building insights isn’t enough — marketers need to think about the practical steps that need to be taken to foster an actioning culture within the organisation. It requires engaging with the relevant stakeholders across the organisation, which includes technical staff that maintain ongoing implementation, marketing managers, media and creative agencies, as well as digital and campaign teams. It’s important that all the stakeholders go on the journey together.

These engagements can be run:

  • daily — maintaining input of attribution data into DSPs and ad servers for real-time buying
  • weekly — review findings and optimisations at a micro level and adjust campaigns as required
  • monthly — review strategy including new data and supporting requirements
  • quarterly — review and set budgets by media channel based on actual ROI

Of course every business and marketing office is different and the challenges faced in actioning attribution insights are unique. However, the survey mentioned above does highlight some trends that might be a familiar to anyone involved with a marketing attribution project. Another common challenge we see is disparate platforms/data silos. This is due in part to the legacy of large vendors acquiring marketing technologies, joining them all together and selling the result as a marketing stack. You can describe these types of solutions as that. The problem is that the technologies were never designed from the ground up to work together, so the data sits in silos. With the growth in ‘walled gardens’ and more news that adtech vendors will be closing off their ad servers to third parties (Google, Facebook), having a solution that can bring in third party data is only going to become more important in solving this issue.

With the knowledge gained from media attribution, we can bastardise the old adage and say, “half my marketing doesn’t work, only now I know which half”. However, if the insight isn’t actioned then knowing doesn’t do you much good.

Originally published at on December 4, 2015.