How brands are bringing outside data in

Share this page

Written by Nick Ribeiro on 31 March 2016

I recently attended a talk given by a data mining company who work with some pretty big brands. The focus was about how they are tapping into as many permitted sources of open (and closed) data out there, and then crunching it through highly mathematical predictive models.

Data is the ubiquitous genderless royalty, and some of its most intriguing and powerful outputs relate to sentiment analysis - a way of identifying feelings and emotion through gateways such as blogs. It’s a bit like Cameron's 'state of the nation' happy sheet conducted back in 2010, but on a 24/7 global scale.

Here's an example of how this particular dataset is being used.

Traditionally, some supermarkets have been solely reliant on chaos, specifically the chaotic behaviour in our natural world, even more specifically the Great British Weather. This behaviour is linked to the Chaos theory, a field of mathematics which can be found in disciplines such as meteorology and, rather conveniently, sociology.

Now we all know that weather forecasts suck 67.7825% of the time. Meteorological data is terribly unpredictable and unreliable, and yet for years supermarkets have used this finger-in-the-wind methodology in an attempt to crack this equation...

Sausages + Burgers / Buns x Condiments 3

Previous attempts have all come up with the same answer. Waste.

But social chaos now presents a new, far more accurate layer. So for example, by monitoring trends and keywords on Twitter or other platforms, coupled with geotag information, supermarkets now have a richer supplementary source of key data that aids the point of need in both supply and demand. A personalised, adapted response for both supplier and consumer.

What strikes me though is that some organisations are becoming so reliant on this wealth of free/premium mashup data, that they feel it devalues the data they already have i.e. historical trends, year-on-year forecasts, and yep, even the weather forecast. I know the world is unpredictable, but is it so volatile that external now trumps internal? If you are a supermarket or any social brand (let’s face it, who isn't), it probably does.

Collating and mining data into meaningful patterns is out of the reach of most of us at the moment. But xAPI is gaining momentum and, along with other Application Programming Interfaces, the ability to bring employees' open data within closed environments.

We're already seeing LMS providers outsourcing their data to researchers. Canvas is offering anonymised network data to researchers 'as a means of furthering the impact of open education'.

So let's pretend these researchers sell their findings back to the LMS host.  Does this strengthen or diminish xAPI? Will the data be so powerful that it dictates and rewrites business objectives on a rapidly iterative and unprecedented scale?

In the not too distant future, predicting trends before they happen will be within the reach of advertisers and marketeers. Topology is one of the key R&D ingredients to that secret sauce.

The claim is that 'Predictive Police' is the new term and task force that is set to replace its learning predecessor. Are we (and/or ‘the bots’) going to operate like meteorologists, assessing where the sentiment jetstream is heading? 67.7825% likely.

And that's the forecast where you are.


About the author

Nick Ribeiro runs PT3, a micro Learning and Development consultancy helping organisations develop the use of digital technology as part of their learning and performance strategy.  Find out more: and @PT3HQ