I do not share the sentiments of those that say digital is merely a marketing ‘channel’, but I do believe that the noise about digital transformation has given rise to some thorny misconceptions. One is that personalisation is somehow the end-game of marketing.
The idea that, one day, the data and algorithms behind every personalised message we see will be unerringly accurate is poppycock (caveat: unless the so-called ‘neural lace’ becomes a reality and privacy is totally outlawed).
But more to the point, the idea that an accurately personalised experience is always an amazing one is poppycock, too, for a couple of reasons. The first is a matter of transparency.
As Jane Ruffino writes in a tweet, as part of a thread about how UX writers have to influence transparency: “If a service brags about ‘personalisation’ it’s our job to be clear where the data comes from. If it’s ‘gamification’ (ugh) are we letting people in on where our agenda is? Are we limiting choices to reduce ‘cognitive load’ but making it seem like we’re comprehensive?”
This limiting of choices could manifest as a fairly harmless ‘filter bubble’ effect, where users become isolated from the options they don’t realise they don’t have; or, more dangerously, as the mirroring of human bias or prejudice present in an algorithm’s training data. There are also plenty of sensitivities in between – see urban legends of parents finding out their daughter is pregnant through Target coupons sent to her in the post, or those people who have been retargeted for engagement rings with their partner looking over their shoulder.
You may find the customer’s idea of personalisation is more akin to excellent customer service, rather than a post-purchase email offering a similar-looking jumper on sale.
Even if you leave these transparency concerns aside, the experience that personalisation describes is often fairly mundane. Most of the time we’re talking about automation of follow-up messaging or fairly banal product recommendations.
Of course, these features may be gold for marketers, increasing engagement or average order value, but taken on their own they do not amaze the customer because, frankly, they are fairly impersonal. Collaborative filtering, for example, just seems obvious to users – yes I’ll have some batteries to go with that kid’s toy I’ve added to the basket, and yes I’ll have the bottom half of that shell suit to match the top half I’m browsing.
I’m not even that sure that customers always want personalisation. I had a trawl through some research: lots of survey reports expound that a majority of customers are frustrated by impersonal experiences, but often personalisation isn’t defined in the survey, and you may find that what the customer has in mind is more akin to excellent customer service, rather than a post-purchase email offering a similar-looking jumper on sale.
I found editorial that talks of customers “yearning” for “hyper-personalisation”, and a study by Segment that states 44% of consumers say they will likely become repeat buyers after a personalised shopping experience. Again, these consumers are probably thinking of more than just simple ecommerce product recommendations.
The Netflix effect
We are continually told that customers expect personalisation because of Netflix or Spotify. But this is just a platitude. It is pointless to compare a library of 30 million songs with, say, the content on a small charity website.
The personalisation on Spotify is needed to fulfil customer need – the user simply feels it is a fluid experience, they can achieve what they want to (find music they know, discover new music). A great experience on a small charity website doesn’t have to be personalised to the same degree – a good CRM and some sensible email segmentation and automation will go a long way to making the user feel loved.
Let’s get to the point here. Personalisation is, more often than not, deeply unsexy.
Personalisation is a name in an email greeting, a product recommended because people that behave like you have bought it, an ecommerce home page that defaults to men’s products because your cookie says you looked at men’s shirts last time you visited, a post-purchase email telling you how to clean the filter on your new hoover, the Domino’s website recognising your location and defaulting to your local branch, Facebook ad copy tweaked to appeal to your predicted personality type, a website landing page tailored to the search term you typed into Google, recommendations influenced by your average order value.
It’s handy to think of it all as segmentation, even if you have enough big data analysis capability to effectively market to segments of one.
Though it’s boring, it is also effective, whether the consumer knows it is happening or not. Econsultancy’s 2018 Optimisation Report (in partnership with RedEye Optimisation) found that four in five (80%) client-side respondents reported an uplift since implementing personalisation.
If you are a marketer struggling with personalisation, it may be helpful to consider a model I found on the Experian website titled ‘six stages of sophistication’. They were categorised as follows: static (no personalisation); identity (eg name, gender, age); insight driven (eg engagement status, propensity to purchase); enriched insight (using more variables still, and potentially partner data); single customer view (real-time using identity, behaviour and circumstances); and predictive optimised (ie next best actions, enabled by machine learning).
My advice, if I’m qualified to give it, would be to refer back to these six stages and avoid getting carried away with personalisation.
And here’s a thoughtful note to end on – Kelly Molson shared an email on Twitter that she had received from Bloom & Wild, the mail-order flowers company. The email read: “I wanted to get in touch as I know that Mother’s Day can be a very sensitive time for some of us. So if you don’t want us to send you any Mother’s Day reminders this year, we won’t. Just let us know by opting out here. Then we’ll do the rest. And don’t worry, if you opt out we’ll still keep you updated with everything else, like normal.”
Life is imperfect. Sometimes, taking a step back and asking the customer what they want is the best option. Data can’t always tell you everything.