What’s the best way to ensure ads are viewable and placed alongside suitable content? Zohar Plan reveals how brands in Latin America are using data in new ways to target consumers more precisely.
Online advertising has received a lot of bad press recently. Unanswered questions about viewability and ad fraud are making brands worry where their money is going, while many remain concerned about their ads appearing next to inappropriate content.
These issues are as relevant in Latin America as anywhere else, and MediaCom and the marketers we work for are all asking the same questions: Are my ads actually being shown on screen? Are my ads being shown to real people on real sites? And, are my ads harming my reputation?
In our region, it’s not always easy to find the answers because, while we can measure Viewability, we don’t yet have the option of buying media based on
this metric. That makes it all the more important we learn from the data we
are collecting to help our clients target consumers more precisely and effectively.
Part of the challenge in LatAm is that buying audiences on Demand-side platforms DSPs – which manage multiple ad exchanges and data exchanges in one place) is unreliable. The data they sell is usually built on pre-existing clusters of consumers, which are often out of date or imprecise.
When data is the new oil you don’t want to work with anything that’s had its effectiveness diluted. So if you want to reach the right consumers, you need to think creatively.
When data is the new oil you don’t want to work with anything that’s had its effectiveness diluted
New targeting strategies
All marketers, particularly those at FMCG brands, need to know two things: where their consumers are on the brand affinity and value spectrum, and where to find them. In LatAm, brands are still using on- and offline data to get these answers.
But too many brands lack actionable digital insights because they have limited cookie access. As a result, their programmatic buying strategies are financially inefficient and performance is non-transparent.
To resolve these issues, brands need to start using new targeting strategies to reach consumers at scale, based on a clear understanding of who they are and where they are likely to be. They need to stop buying pre-built digital audiences and start becoming media and platform agnostic.
Too many brands lack actionable digital insights because they have limited cookie access
To take more control of their first-, second- and third-party data, brands can use an owned Data Management Platform (DMP). They can then work with a technology partner to integrate their DSPs and ensure they buy the right media to reach the right consumers.
At MediaCom, we’ve been using this approach for Coca-Cola since 2015, working to shift its digital spend to programmatic. By the end of 2016, we ensured that 17% of Coca-Cola’s digital budget was assigned to programmatic buying and that there were 33 active campaigns in its DSP. Through smart buying, the effective Cost Per Mille (eCPM) went down by 74%-82%, while the Cost per Completed View was 2%-70% lower in the best
performing quarter versus non-programmatic investments.
Adopting a cyclical approach
Traditionally, brands have adopted a ‘one-way’ approach to DSPs. They work in a way we call ‘from back to forward’. They look at the brief (the ‘back’) to understand the kinds of consumers they should be looking for. These profiles are based on data from offline sources, such as panels and focus groups. Then they move ‘forward’, using the DSP to find pre-built audience clusters that match these descriptions as closely as possible.
But brands would be better off adopting cyclical ‘two-way’ approach, working ‘from back to forward’ then ‘forward to back’ again. In other words, using the brief to inform the buying strategy (as per normal), then analysing the performance data it collects to challenge and refine the brief – and its suggested audience targets – for next time.
To begin working in this way, brands should:
1. Use their first-party data as the underlying basis for their audience segmentation. Once consumers have been seen at least two or three times during a set period, they can then be classified and segmented with mathematical certainty.
2. Unify their first- and second-party data, by dividing interactions by quarters according to their performance (specifically, visits, views and content consumption). They can use the variables which were assigned a statistical value to build a matrix of low and high-affinity clusters.
3. Complete the process by applying third-party data to build audience ‘profiles’ based on who interacted with the brand, and who over-indexed against average internet behaviour.
We used this approach to help Coca-Cola challenge its understanding of its consumers and identify secondary audiences beyond its original brief. As part of the process, we built more than 24 ad-hoc brand clusters over 12 brands, based on insights from 65 million users in Mexico. We are now using this data to change Coke’s digital creative depending on who we’re talking to.
In future, this approach may inform more creative processes, which just about underlines a key lesson from LatAm: While our region may still be developing our data capabilities, we are already introducing best practice frameworks that can be reapplied all around the world.
First published by Latinspots here.
Get insights in your inbox
Interested in the future of media and marketing?
Get actionable advice, insights and opinion from MediaCom experts every month