What does personalisation really mean?

‘To reach the ultimate expression of personalisation, algorithms need to do something far smarter based on all the data at their disposal’

Retail is an industry frequently awash with buzz-terms and next big things. Usually these terms pop up, are written about extensively, potentially find some expression in reality and are then replaced before too long with something that promises to be bigger, bolder and better.

Personalisation is a bit of an oddity in this respect as it has been talked up for a long time, but still remains a hot topic. One of the main reasons for this is that its scope is so vast; it covers every aspect of customer engagement and experience and, therefore, is highly complex in execution. It cannot be quickly implemented in a comprehensive and complete manner to become just another old concept that no one has to think about anymore.

Since it is complicated, progress has also been more pronounced in some areas than others and the definition of what actually constitutes personalisation shifts over time. At the start of the 2010s, for example, as software platforms started to enable it at scale for the first time, saying “Dear Frank” instead of “Dear Sir or Madam” in the introduction to mass marketing emails was regarded as personalisation. That’s now very basic stuff.

Personalisation vs customisation

The interpretation of what personalisation means can cause confusion; businesses sometimes think they are doing personalisation when actually another term would be more appropriate.

Let’s take an example. If you go to a retail site and are presented with the same display – layout, product range, greeting and so on – everyone else receives, that’s a standard, generic experience.

Now imagine viewing a product page, perhaps a T-shirt, and finding out you can select it in different colours, select some wording to put on the back and choose from 100 emoticons to put on the breast pocket. That might sound like personalisation, but it’s actually customisation because the customer has to actively make the decisions and selections themselves. To be personalisation, it has to do that work on your behalf.

If, instead, when you return to that site the content you are presented with has been structured in accordance with your perceived interests and preferences, this does represent a form of personalisation. Should you then see something you like and proceed to the checkout, it may “remember” from a previous purchase that you paid with PayPal and collected it from a store in Crawley. If it defaults to these options for you, again this is a form of personalisation.

Letting algorithms make decisions for you

But this is still some way short of being personalisation in its truest sense. It may seem smart, but actually all it is saying is, “This is what you did last time”. With product suggestions, they may be more relevant and accurate, but the algorithm probably made those associations based on other products previously viewed or purchased; this is not really exhibiting a deep understanding of the individual customer.

To reach the ultimate expression of personalisation, algorithms need to do something far smarter based on all the data at their disposal. They would need to make decisions on your behalf and tell you things you don’t know, rather than just remember things.

Instead of the algorithm showing the shopper product ranges that have a higher chance of being relevant, perhaps because other people who have exhibited similar purchasing behaviour went on to browse or purchase them, they would focus on the specific individual. What do we know about you and how can we use that information to give you the most relevant, accurate experience possible?

When these algorithms reach a level of sophistication where they are able to say, “This is actually what you like, you just don’t realise it yet”, then we are talking about personalisation.