The personalisation problem

Personalisation is tried and tested: tailoring content to individual users’ needs clearly increases conversion rates and makes complex interactions useful and straightforward.

There’s a reason why 35% of Amazon’s revenue comes from the site’s beefy recommendation engine. It’s serious business, when you get down to it. It’s also simple, which is why every digital business is getting their hands dirty.

But there’s a big problem with this trend: the simple tools that allow personalisation mean businesses can use them, but not necessarily well. The result is a mish-mash of businesses trying and failing to tailor content that just ends up pushing more people away.

So how can businesses avoid falling into that trap? Before even thinking about personalising content, businesses need to understand three core tenants of personalisation.

The first is this:

Individuality is at its most enjoyable when it’s done collectively.

Do you remember the Coke campaign a few years ago that was all about people’s names being printed on bottles and cans? Coke manufactured millions of cans and bottles with peoples’ names on them, and encouraged people to share one with a mate if they found their name.

Slightly corny, but it worked — there were 250 million bottles and cans sold in Australia, a country of just 23 million people. And the company increased its share by 4% in the category.

The reason this campaign worked is because it did two things: identified us as individuals, and then reminded us that we’re part of a greater community. “Buy the world a Coke”, and all that.

See, any attempt to personalise content needs an understanding of the Vyasa Paradox at its core: everyone is the same, and that’s what makes us different. Bad attempts at personalisation often look at the differences, without thinking about how we all form a narrative.

That usually happens when marketing is at the core of personalisation, trying to solve a problem that doesn’t exist. You end up just pushing personalisation on people to drive a sale, and then following up with something generic like targeted remarketing — which pushes more than half the people away anyway.

Which brings us to our next rule:

You need to personalise products and prices — not just content

Personalising content is simple. Any decent web CMS now enables you to pick segments of users to personalise content based on location or even more complex arrangements — it’s not hard. We call that “rule-based personalisation”.

Rule-based personalisation has a bigger chance of going wrong. I recently booked a flight to Japan through Booking.com, and even weeks later after I’ve arrived home, the site is still asking me if I’m planning that trip to Japan. Rule-based personalisation may work on paper, but it doesn’t connect me to a greater narrative or even acknowledge my personal circumstances. There are too many ways to slip through the cracks, as my experience shows.

The answer? Start personalising products and services instead.

Do you remember when Radiohead released their album, “In Rainbows” back in 2007? They were the first band to ask people to pay whatever they felt like — and in terms of digital income for the band, this record made them more money than all the other Radiohead albums put together. Again, it was part of a larger movement — Radiohead made their customers feel like they were changing the world, because they were.

Or think about Spotify. The genius of Spotify isn’t that it recommends songs, it’s that it gets better over time. The product literally tailors itself to the user. (And importantly, always allows the user to opt out of a song if they want — another place for the product to learn.)

Let’s call that “predictive personalisation”. This is a harder strategy, but it’s definitely more tangible and valuable. It also requires a delicate balance: most people say that privacy is becoming worse due to technology, but they still want products and services that know them intimately.

So how do you walk that line? It’s simple: you tell customers that you’re using personalisation and be up-front about it. Just like Spotify, and just like Radiohead. The research shows that customers who are told that personalisation is being used on them end up more satisfied over time. But if they’re lied to — they feel betrayed.

Which leads us to our third big lesson:

When it comes to personalisation, don’t assume you have the audience’s attention

We have a retail client that recorded a strange anomaly recently — we found that on Tuesday nights, customers would have a far higher conversion rate than they did during the rest of the week. After some investigation, we found that the buying behaviour started on Sunday, when customs would look around, then on Monday they might add items to a wishlist, and then on Tuesday they would buy.

But we weren’t doing anything different with regards to marketing activity. So what was going on?

As it turns out, the TV show “Married at First Sight” aired during those times, and the Tuesday show was the finale of the week — when people would buy from the site. They were obviously browsing and buying during the ad breaks.

Now, a traditional marketing approach might have pushed us into something like buying more ads. Instead, we opted for a personalised approach with a touch of a bigger picture.

On the site, we advertised some sales based on the outcome of the contestants on the show — if something unexpected would happen, we would launch some sales. It comes back to the idea of talking to an individual — recognising their interests — but also being part of a broader group.

These are just a few simple examples, but hopefully they underline the broader theme: personalisation isn’t just about tailoring some content and calling it a day. We’re moving from a transaction-based economy to a relational economy, and that’s what we should be using personalisation before.

And when you achieve that — your product will simply sell itself.