By Gina Horton, Digital Account Manager
227 Facebook Likes is, apparently, the average data set that artificial intelligence needs to understand your personality more accurately than your friends and family. A number I heard referenced multiple times during my time at Ad Week 2017.
These findings aren’t new. Cambridge University released them in 2015 as an “important milestone” on the path towards more social human-computer interactions. Even bolder is the revelation that based on 10 likes the platform can understand you better than a work colleague.
Is it arrogant to state we can understand people better than their closest friends and family based on a data set?
Let’s take a look at my Facebook Likes for example:
Charlie bit me? As you can see I probably haven’t liked a page since the dark ages of university. Based on the above I sincerely hope that my Facebook Likes do not summarise my personality, for everyone’s sake.
Now I know 2015 is a lifetime away considering the proliferation of digital channels and research, and even Facebook now recognises what little value a Like holds (and has altered its algorithms accordingly).
However, claims that data can understand us better than ourselves are still prominent. As a digital buyer, I am inundated with new ways to target audiences based on big data sets. But to what extent does this data help me (and most importantly my clients)?
During the Big Data Backlash session at Ad Week, the conversation turned to how big data helps advertising. We often become excited by the new ways we can utilise data and target audiences, however to lead solely by data can sometimes prohibit creativity.
A great example of this was the 2016 John Lewis campaign. If this had been solely led by data, it could have emerged that statistically dogs do not jump on trampolines. But then we wouldn’t have had Buster…
Another issue to consider is that of privacy. People are worried, and rightly so, about their privacy and control. Yes, we can target audiences based on their app downloads, where they have been, what they have bought, the list goes on. But in some instances we need to consider that just because we can, should we? And if we do, how do we respect people’s privacy?
I’ve recently purchased clothing from Mango, and since doing so have been retargeted everywhere. This free-for-all approach is less than appealing to me as a consumer. Yes, Mango, I like your clothes, but please stop stalking me.
I’ll tell you what would be handy, if Mango kept that data and targeted me next time it is apparent that I am in the market for a new dress with relevant suggestions. This would mean they’re using my data to actually aid me somewhat in my search, and not bombarding me with every item on their website at every turn.
We are seeing more and more evidence that personalisation is what consumers want, with 86% saying it plays a role in purchase decisions and 48% claiming they purchase more when marketers leverage their interests and behaviours.
Adding a human touch to the data can ensure we are not just counting, but in fact understanding how to use the data to best serve the audience. It is planners and buyers responsibility to interpret data, understand our consumers and know when to communicate with them…but also when to stop.
As seen in my opening stat, data can often be over-read, misread even in some cases. It can also bring us to the small insights rather than the uniting ideologies of our audiences. By adding a human touch to data we can avoid arrogant claims, and provide our audience with personalised messages which are not over assuming.
If you want to see a perfect example of data being used smartly, check out our work for The Economist:
Now read these: