Less flower power and more brain power – why the 70s science of analytics is coming back full circle
If there's a buzz word in the modern marketing world it's 'data science'. Whilst for many, the phrase feels novel or innovative, analytics has been around for decades. And, what’s more, it’s beginning to come full circle.
This era is widely regarded as the birth of the analytics ‘industry’, where the more progressive IT programmers saw that the code they were writing could be applied to more than just mainframe maintenance and development.
The 90s started to help democratise analytics. User-friendly toolsets opened it up to a community that didn’t need a degree in Computer Science and where less technically minded Insight Analysts could exist alongside the more technical Data Analysts.
The noughties SCV (Single Customer View) boom led to more complex analytical techniques being implemented and the arrival of the Data Scientist.
The 10s saw a massive boom in open source tools, democratising analytics and opening it up again to a much wider community.
In the present day, we’re beginning to go full circle – reverting to the more specialist/technical applications of analytics, fuelled in part by the noise around Artificial Intelligence (AI) and Machine Learning (ML).
Whilst it’s easy to get lost in the PR and excitement about AI and ML, our client experiences at Indicia Worldwide suggests that interest vs adoption don’t seem to be happening at the same rate. Why is that? Here’s a few thoughts…
Whilst database technologies are readily available to build and implement AI/ML solutions, several businesses are still getting to grips with optimising the infrastructure that they already have.
We still see that the number one challenge for most businesses is gathering their data into one place, in a way that supports the wider strategic aims of the business and in a way that allows them to derive value from it.
Put simply, many businesses just don’t have the technical infrastructure to deploy AI/ML solutions.
What’s more, considering the current economic climate and the potential impact of Coronavirus means that businesses are really focused on making the most of what they already have.
Traditional analytical techniques such as regression and clustering are still delivering value and, as the saying goes, if it ain’t broke, why fix it?
The implementation of AI/ML doesn’t just require an optimised data structure and the right toolset, it needs people with the right skills to use them properly.
Going back to my earlier point, we are no longer in an era where almost anyone could turn their hand to analytics and businesses could invest in more readily available, lower skilled (and lower cost) teams to deliver what was needed.
Nowadays, the Data Scientist required to develop and implement advanced analytical techniques doesn’t come cheap and businesses may need to re-think their recruitment policy if they’re to fully make the most of it.
Now I’m only talking about AI/ML in the context of marketing communications here but, the sad reality is that some businesses just aren’t built to really capitalise on the improved efficiencies or depth of insight that AI and ML can deliver.
Machine Learning techniques are best suited to large, dynamic datasets where you’re able to really harness the power of these techniques to drive outcomes at scale.
Unfortunately, many businesses simply don’t have enough customers or sell enough products to have a dataset that lends itself to Machine Learning.
Ultimately, this often means it needs a sledgehammer to crack a nut and lower costs, whereas more pragmatic solutions can deliver a very similar result.
So, you might be reading this thinking I’m anti AI/ML. This couldn’t be farther from the truth.
At Indicia Worldwide, we’ve built several AI/ML solutions for our clients but equally, we’ve seen several instances where businesses feel it’s an area they need to be in because ‘everyone else is’.
Building AI/ML for the sake of building AI/ML into your business is the wrong approach.
There is no doubt that AI/ML represent the future of advanced analytics, but equally, it still requires human intervention to guide and shape the future alongside these techniques.