Bunk Off — Debunking ‘Machine Learning’ Etc, And What It Really Does For Marketers…

What’s the truth behind advanced data science used in marketing?

That’s right, the buzzwords have finally become reality. At first it was “big data”. Then it became “statistical modelling,” before turning to “algorithms.” Afterwards “machine learning” made an appearance before we finally move into “artificial intelligence” as the finalé.

The buzzwords sound AMAZING in presentations, you really do sound like a smart ass. Well done you. But anybody on the receiving end of such presentations will tell you that they’ve already heard it all before. And most of us still have no idea what it means in practice!

Does any of this stuff actually work?! What does it even do? Well the short but irritating answer is: yes it works, if you ignore the hollywood-inspired visuals and do it right.

So to debunk the over-glamourised perception of all this jargon, we must first have a glossary!

Big Data — “is high volume, high velocity, and/or high variety information assets that require new forms of processing to enable enhanced decision making, insight discovery and process optimization.”

Translation — a huge store of data (about something) that can be used quickly and has many related variables… I.e. pretty much anything digital now… I.e. most modern data.

Statistical Modelling — “is using mathematical models to embody assumptions in order to describe a set of probability distributions, some of which are assumed to adequately approximate the distribution from which a particular data set is sampled.”

Translation — an equation that predicts the likelihood of something happening based on the sample data you have to hand.

Algorithm— “is an effective method that can be expressed within a finite amount of space and time and in a well-defined formal language for calculating a function.”

Translation — a defined set of repeatable steps used to automate an operation, usually on a larger scale than a human is capable of manually.

Machine Learning— “is a method of data analysis that automates analytical model building. Using algorithms that iteratively learn from data, machine learning allows computers to find hidden insights without being explicitly programmed where to look.”

Translation — this one’s not too bad of a description. To simplify further, it’s an algorithm where the end goal is still set, but the steps to get there are no longer defined. (The algorithm has to deduce the pattern and the path of least resistance itself as involving a person in the process would slow everything down immensely.)

Artificial Intelligence — “is a flexible rational agent with cognitive function that perceives its environment and takes actions that maximize its chance of success at some goal.”

Translation — the Terminator vs Bicentennial Man?

Now that we’ve dumbed the subject down a bit for the 98.74% of us who aren’t data scientists to some degree, there are a few easily understood benefits from these data science tools.

I can predict the likelihood of something being true, I can automate much of the legwork, even while not knowing how to get to my end goal, and I can do this at potentially huge scale and speed.

These are perks marketers can really do with using day to day! In fact, most of these tools are already commerically wrapped up in “Predictive Analytics.”

Analytics are great. I love analytics. I’ve squealed with joy before when someone uncovered how a supermarket was missing out on profits because they weren’t selling enough black bin bags. Our company is even in effect a predictive analytics company.

But there is a hugely important point that MUST always be considered when using predictive tools — the quality of the data.

Predictions and projections can only be made based on what you input into the system. You really cannot polish a… you know what.

This is why we focused on the best type of data for marketers… intent data.

And the intent data that is best for predicting what people will buy next is purchase data. If I know what you’ve been buying, how, why and when, then the odds of me succesfully predicting what you’ll buy next improve dramatically.

Marketers need to use such predictions to personalise their marketing. It’s time to stop blanketing your entire audience with the same message. It’s time to start showing me what might genuinely be helpful or fun or make me feel better. Don’t advertise, recommend.

To read more, try: “Shift Happens. Audience Data Is Changing Marketing.”