skip to main content

The Determinists versus The Probabilists: Why consumers are moving targets for brands

Created at January 26th, 2015

Consumer identity and privacy issues are increasingly dominating media and technology news headlines. This is largely being driven by the wide scale adoption of mobile as the de facto method of communication in today’s world. However, although extremely effective for ‘right time/right place’ advertising, handheld devices produce many challenges for marketers: not least the fact that the much maligned cookie (the default method of identification for ad targeting on desktops) does not work on them.

This has resulted in technology vendors investing megabucks to try and find new and innovative ways of recognising consumers who are ‘on the go’. Two methods of identification have recently arisen:

The Determinists– identify me based on the information I use to log into a publisher website. Facebook has no problem knowing I’m Paul Hatley whichever device I’m using or wherever I am during the day.

Similarly, when I download an app like Uber, I provide personal details and confirm my location. As part of this value exchange, I am giving away all of the behavioural signals that are picked up from my use of that app to receive a better user experience, and ultimately, more relevant advertising.

The Probabilists– will try to identify me based on my browsing behaviour and the locations I appear at throughout the day. Powerful algorithms predict I’m the same Paul Hatley on my mobile, home and work devices, based on any common traits I exhibit. Sometimes referred to as ‘device fingerprinting’, these techniques infer my details and location by a set of coincidences. I will then be aggregated, segmented and monetised in much the same way as my cookies are currently through ad exchanges.

Why are probabilistic techniques becoming so important? The answer is simple – my value is far higher if my location is known (or inferred). Where I am right now suggests where I might go next, releasing a treasure trove of opportunities for contextual and time sensitive brand advertising. This is compounded by the ‘deterministic’ brigade (Facebook, Google, Twitter, ebay, Apple, Amazon et al) – the acknowledged titans of consumer identification – at least within their own walled gardens – my ID is unique to each publisher site but is not shared with the wider digital ecosystem.

A key challenge for probabilistic algorithms is that they need location data to work effectively: GPS coordinates, check-ins, IP addresses, beacons and Wi-Fi recognition all help with this, but none is perfect in their own right. Real insight can only be gained by blending mobile data with the rich demographic, attitudinal and behavioural information contained in offline data. This is creating a fascinating collision of ‘traditional’ data elements related to where I live (static, permanent, slow) with those on my device related to where I go (dynamic, transient, fast). Consumers are now moving targets; fixing specific (known) data points to them requires significant skill and expertise.

Many observers fear that, in the absence of clarified law, these new forms of identification are taking us the way of the 3rd party cookie – a recent opinion of the EU privacy working party said exactly that. As marketers it is up to us to be vigilant as to how this data is collected and disseminated and ensure aggregation and de-identification occurs in a privacy compliant way. How we tackle these issues now will set the data privacy reform agenda over the coming years and ensure that our industry is working transparently and fairly in the eyes of the consumer.

Paul Hatley – Data Strategy Expert, Acxiom