How to do personalisation without doing personalisation

Related Articles


We’ve achieved the highest possible accreditation from Investors in the Environment!


We are RocketMill: This is Summer Knights


How to do personalisation without doing personalisation


Today, I’m going to be talking about how to do personalisation without actually doing personalisation. It seems like a really, really hot topic at the moment. Hands up if one of your clients has asked you about personalisation. Yeah, enthusiastic waving at the back.

What is personalisation?

I think it’s important to understand what we’re talking about when we talk about personalisation.

The vast majority of personalisation online is, basically, taking your data – usually data that you’ve handed over knowingly, so that could be your email address, it could be a user name, it could be by giving something access to your Facebook account – a website then taking that data and matching it with their own data. Running your information through their database and seeing what matches, and then out the other end comes a personalised experience.

Examples of personalisation

That can come in lots of different forms. The way you see The Guardian is probably different from the way I see The Guardian because I’ve personalised my homepage to take out sports, because I don’t care about sports, so it’s not in the homepage, but I’m sure lots of you do, and you probably see that. That’s an example of a, on the face of it, quite basic but quite meaningful personalised experience.

You could also take that one level deeper and think about your online banking has a personalised experience because you’re logging in, and it’s giving you your information. Here’s what my Coinbase account looks like at the moment. You can see that sad graph of all my pretend money disappearing over the last month. This is personalised to me because it’s my data.

Marketing and personalisation

When we talk about personalisation, it’s usually done as a marketing thing, but actually, this is personalised as well. Personalisation is usually a logged-in experience because you’re handing over data, and then those companies are matching it with their own data to show you what you want to see.

Increasingly, as we know, there are lots of services out there online that are slurping up lots of data about users and tracking you around the web – in a GDPR-compliant, anonymised way, putting you into different buckets of users doing different behaviours. I believe these are the four most commonly encountered third-party cookies online.

There’s more and more data being collected out there, and it’s letting people do some interesting things. Just to give you an idea of the scale of this, our good friends at The Daily Mail, never one to miss an opportunity to make their website worse, have over 19,000 third-party tracking cookies on their website, according to a study by Digiday. They’re tracking 19,000 different bits of personalised data, which is fairly ludicrous if you think about it.

Google Analytics audiences

You can kind of get an idea of the kind of information that’s being slurped up if you visit the Audiences tab in Google Analytics, but there’s lots of interesting data in there. You can see, in this instance, this website has 22,000 users who are TV lovers and also 22,000 visitors that are value shoppers. These are anonymised buckets of users, but we know what their behaviour is like online because these services are tracking them and grouping them up for us.

Personalisation is becoming a ‘logged-out’ experience

What this is allowing people to do is use services like these – Dynamic Yield, RichRelevance, Episerver – to actually make personalisation a logged-out experience, so when you arrive at a website, it’s looking at your cookies and going: “Ah, you’re in the value shoppers group,” so maybe showing you some products that have slightly lower prices.

Personalisation is moving from being something that you only get when you log in to something that you see by default, having not knowingly handed any information over to a website, and this is the personalisation that we’re often getting asked about. How do I make my website more relevant to people? How can I serve different messages to different groups of users? These are the types of platforms that we can use to do that.

However, these platforms are expensive, they can be quite difficult to implement, and they often need a quite sophisticated level of expertise to get the best out of them. Where you have your ‘one-to-many’, or no personalisation, versus your ‘one-to-one’, logged-in, “here’- my-specific-information”, you have increasing complexity as you come this way. I can make a WordPress website, probably in about 10 minutes, that was at that end of the scale. I definitely couldn’t make a website that was at this end of the scale.

On the ‘one-to-many’ scale of personalisation, or no personalisation, if you’re not using one of these personalisation platforms, your website copy is probably going to be at that end. You only have one homepage. You can’t personalise that for anybody. What I’m going to talk about today is how we can maybe bring things a little bit this way. As you can probably guess, because I’m the head of content, the way we do that is with editorial.

‘One-to-some’ personalisation

You may only have one homepage, where you have to write one piece of copy that targets everybody. With editorial, you can publish as much stuff as you like, and you can make that as personalised as you like. What we have kind of come to call this level of personalisation is ‘one-to-some’. We’re not targeting people on a one-to-one basis, but we are targeting groups of people based on what we know about them and based on what we know their struggles are, what they want to read, all that sort of stuff.

This basically boils down to one specific thing that you need to get right, and that is: knowing your audience. You have to know what your audience’s pain points are. You have to know what makes them get up in the morning in order to correctly personalise your editorial content to them. This isn’t actually as difficult as you might think because there are loads of great, absolutely fantastic data sources out there that we can lean on when we’re trying to learn about our audiences.

Using data sources

I was just talking about Ofcom earlier. They publish these fantastic studies about media consumption and what social networks people use that is absolutely incredibly valuable. Think with Google. Obviously, Google want us to use their services more often and throw more money at them, so they’re providing us loads and loads of market research data that we can use to do this kind of thing. Obviously, ONS publish data sets on just about anything you could imagine.

A little tip on this point when you’re trying to find your data sources. When you’re searching for whatever it is you might want to search about, if you append that with a ‘filetype:xls’ modifier at the end there, so that’s telling Google to only search for spreadsheets, what that will help you surface is the original research that sits behind the press releases, and the articles, and the annoying PDFs that you might otherwise find.

Example: ‘one-to-some’ personalisation

Just giving you an example of how this works in the wild, some work we’ve actually done for a client, this is an article that we have published for OneFamily. It, obviously, continues down this way. How do we come to this article? We knew that their audience was over 50s, so we slurped up as much knowledge as we could about this over-50s audience. What are the issues facing them? What are the questions that they’re asking? This was done through all of those tools I previously mentioned and a whole bunch more as well, including the keyword research that the technical SEO team did.

One of the issues that we picked up for the over-50s audience is, for those with children between 18 and 24, 50% of those children live at home. This audience of over-50s, for 50% of them, this article is personalised because their children live at home, and they’re probably wondering should they charge them rent? Because we looked in Keyword Planner and all those sorts of tools and got an idea of search volume, we knew that, actually, a large portion of them were asking this question. So we said: “Yeah, we definitely need to produce some editorial about that.”

I’m really happy to say, actually, that again, due to some really excellent guidance from the technical SEO team, we keep talking about voice search, this is actually one of the best kind of results we’ve had from voice search, so let me … I’ll give you a quick demo here. “Okay Google, should you charge your kids rent?”

[Google: According to OneFamily, how much rent to charge your child…]

Yada, yada. It goes on. Tick technical box, that’s great, but it was the insights that drove that that led us to that result.

What we also did, which was a lot of fun, was that we promoted this article on Facebook to a specific audience of over-50s who had children, so 50% of that audience are going to have those children living at home, and it had kind of predictable results. It got a lot of engagement. I’m a particular fan of Niki Kendall there, because you don’t want your kids to be an ‘entitled arse’. The interesting thing that we took away from this was, that all of the people responding to this were mums. No dads were involved in this conversation. If we were to redo this again or do something similar, we might narrow that targeting down to just women.

Why you should start ‘one-to-some’ personalisation now

What I’m saying is, when it comes to personalisation, I would encourage you to not be an entitled arse. You don’t need to go right off the deep end and buy all of these really expensive personalisation platforms. You can, and if you’re ready for it, you absolutely should, but there are ways to get in the middle to this ‘one-to-some’ personalisation without investing any money, without much technical expertise, and without needing the data analysis skills to get there. Thank you.