Combining insights from different data sources to give a single, reliable view of each person’s habits and preferences is the biggest conundrum of ‘big data’ – for marketers, at least. The missing link between online and offline data, between behaviour and attitudes, between individual transactions and general trends, is in part the inability to match up personal information from one source with anonymised or aggregated data from another.
At an ISBA conference last week, the European Commission’s director general for communications, content and technology Robert Madelin revealed data pseudonyms as one possible solution that is being discussed by data regulators. He outlined a system whereby individual IP address and cookies could be assigned a pseudonym that can be tracked across different devices.
Madelin gave a particular example of the implications: “We would be able to actually track who we’re targeting ads at but then link back to whether people go and buy things from the companies paying for the ads.”
But there isn’t yet any indication of whether regulators such as the UK’s Information Commissioner’s Office will deem this to comply with current legislation, so marketers shouldn’t get too excited yet.
There are ways of creating detailed individual profiles already, using a single ‘linkage key’ such as an email address to bring together everything a company can already know about its owner within the current rules. But the problem for any given brand is that it is harder to amass any relevant information about how a person moves around the internet, which ads they click on, what they buy and so on if they aren’t signed into that brand’s own service.
This means brands have started to believe their websites have to become ‘destinations’ where consumers spend a lot of time, which in reality is something that only very few can achieve.
For any brand, linking up different data sets in a useful way is probably one of the most difficult yet rewarding things marketers could currently be focusing on. The best way to do it will depend on the services, products, technology platforms and partners of that particular organisation, but even without the existence of data pseudonyms there are plenty of case studies already to be found showing how you can move in the right direction, not least within this very website.