What will the future hold for ad targeting?

September 20, 2020
SCMP Insights
It’s becoming likely that there will be two major choices for data activation in the future.
  1. Deterministic, first-party data collected by the owner of a direct-to-consumer relationship (i.e. a publisher or brand).
  2. Probabilistic data collected by a third party and aggregated into anonymous groups.
In this article, I want to discuss the new breed of third-party probabilistic data sets.

Probabilistic data
Definition: Probabilistic data is information that is based on relational patterns and the likelihood of a certain outcome. A common example of probabilistic data at use is in weather forecasting, where a value is based on past conditions and probability.

This is useful in a privacy-first world as, in media terms, this is likely to be data that is acquired en masse by scraping data (such as article context) from various sites and then categorised into standardised segmentation before applying some modelling to expand the base.

This practice does not require you to target an individual, but a group of people that have, or are likely to show, this attribute (for example, an interest in sports). It’s also easy and cheap to do and can be pretty niche in context.

On the face of it, this is a great tool for media buyers. However, it does have some issues.

Firstly, the quality of data may be unclear (when was the data collected? How much was the data modelled? Does a user reading one article qualify as having an interest?). It does not allow a brand to establish a relationship with a specific user, measure changes in awareness and attitude, retarget or monitor sales attribution.

Issues on the horizon
Looking to the future, this kind of data, and the providers associated with it, may be in trouble for two major reasons.
  1. Most of this data is scraped from publisher sites (sometimes with explicit permission, but sometimes not). There is a growing sentiment that publishers are not fairly compensated for this as they rarely see a significant or measurable payment. In fact, it may lead to the continued commoditisation of their data as brands may activate data from premium publishers for run-of-network OMP (open marketplace) buying. A quick LinkedIn poll shows 82 per cent of media professionals believe publishers should stop third parties scraping their contextual data and the tools to do so are readily available.
  2. Growing governmental, browser and OEM (original equipment manufacturer) regulation mean users will be easily able to select what kind of data is collected and by whom – making trusted first-party relationships even more important.
It’s my opinion that this will put third-party data providers in a difficult position with heavily restricted access to quality data.

In the UK, we saw a few location-based data providers close shop when GDPR (the European Union’s General Data Protection Regulation) initially came in as it became apparent that users would not give them permission to be tracked. The upcoming changes are far more comprehensive and I foresee many more data-based third parties going the same way.