Useful Personalisation

In my experience as a consumer and as a CRO consultant working with customers, I can understand why Personalisation has created such a buzz in the last few years. I think it’s exciting to see how many businesses are using this as a marketing technique now. But not all Personalisation is created equal, and I want to show you a few examples, some of which I think are extremely effective and some of which maybe not so.

Lets start with an offline example of personalisation: Starbucks.

I’ll admit I’m not a huge Starbucks fan, but the handful of times I’ve been into a branch it still always surprises me when they ask for my name. I keep forgetting that they do this kind of personalisation. And when they ask that question, two thoughts pass through my mind. The first is:

“Why do they need to know!?” Then I quickly remember, “oh yeah, this is what they do so they can call my order.”

But my second thought is usually “quick, think of another name.”

See, I’ve been burned in the past. I don’t have the easiest name, and they never ask how to spell it. So the last time I was in Starbucks, my name wasn’t ‘Liss’, I told them it was Ana.

Now, if I have these concerns in real-life, how common do online users think the same?

We are all already aware of how cautious users are at sharing their data. Getting something as innocuous as an email address or first name is already tricky. Relying on that data to be accurate is naive.

Recently, the European GDPR rules came into play, and like many of you I received a bucketload of privacy policy emails from companies, some of which employed what could be called ‘personalisation’.

They basically mail merged my name into their emails. Except, I hadn't given them my name, only my initials, so how personalised are these e-mails really? Simply putting my name into the e-mail isn’t useful.

Amazon

Amazon are considered Gods in the world of personalisation & recommendations, but controversially, I don’t believe that Amazon are any good at personalising the shopping experience.

It’s often said that Amazon (and some other websites) personalises the My Account section by calling it “Elise’s Account” or “Toms Account”, but this is not personalisation, this is a UX choice and some basic programming.

Anyway, we recently went on holiday. In preparation for this holiday I bought a number of items off of Amazon; sea sickness bands, packing cubes, flight socks and some water shoes as we were going to be spending some time on the beach.

I find that when researching what item to buy, Amazon’s recommendations do come in handy - comparing one item against another and being able to read reviews this is helpful and Amazon enables you to do this very easily with it’s ‘related products’ and social proofing on the product pages.

However, post-purchase the recommendations I received were simply for more of the same items I’d just bought; I only have one pair of feet — why would I need another set of flight socks? Or, another pair of water shoes?

What would have been useful would have been suitcase recommendations, or perhaps swimwear to match my water shoes. And in fact, when you look at the ‘what other customers bought’ panel, many customers who purchased water shoes appear to have also bought snorkelling gear — now that would have been helpful on our trip, but the main bulk of that panel is taken up with other water shoe brands.

Some more example of Amazon not giving useful personalised recommendations is their emphasis on ‘buy it again’; I have a whole panel on my homepage which is dedicated to past purchases and encouraging me to buy an item again.

The problem is that my partner & I not only try to buy in bulk, but we’ve also been trialling a zero-waste lifestyle since the beginning of the year, so many of the items I’m being told to buy again are, by design, reusable.

Also, ‘Inspired by your wish list’ — thanks Amazon, but I’ve already done the research & decided that the stand mixer I want for Christmas is the one that I want. But, since you now know I might be interested in baking, why not recommend me some cookery books or piping kits instead?

My point is that Amazon clearly has the data available, but they’re not sharing what is useful for the user. Their algorithms do not take into account the idea that users sometimes buy multiples to find the best product & return the rubbish.

It’s unfiltered big data and as a result it’s not a personalised experience.

So, having shit all over Amazon’s attempt at personalisation, who do I think does…

…Useful Personalisation?

Well, again, let me first tell you an offline story:

My husband & I have a favourite bar, but we’ve only been there about 4 times. So why is it our favourite?

Because the first time we went in there, one of the waitresses took the time to sit with us, ask us about our favourite types of drink, she checked in regularly and, at the end of the night, recommended us other local bars to try (and, in return, we recommended our favourite nearby restaurants.)

What makes them truly special though, is that this waitress remembers us and our preferences:

• We followed some of her recommendations and shared notes when we returned.
• They know our favourite drink so that when they’d run out of one of the ingredients, they were able to recommend something else.
• And, when we returned after a 2 month hiatus, we were welcomed with a hug and genuine excitement that they’d restocked that ingredient so we could have our favourite Bootlegger drink again.

The ironic thing is that they can’t remember our names — just our preferences but that is what matters.

So, online, who remembers preferences?

An easy one that many people cite is ASOS.

They recall whether on your last visit you were on the mens or women site and redirect you accordingly.

Simple, subtle but it works and it’s useful.

Grocery stores also usually get it right: Tesco uses Clubcard and previous shopping data to provide relevant coupons and offers via both post and e-mail.

And up until February this year, Waitrose allowed you to pick your own favourites which then gave you 20% off those items. They’ve now scrapped that scheme and gone with a machine learned scheme similar to how Tesco works, but either way, food shopping online is often the easiest shopping you’ll do since past purchases are visible and convenient to re-purchase. This is useful.

Graze does personalisation very well — they allow you to provide dietary requirements such as gluten free or vegetarian, and also to rate the available snacks. This ensures that Graze will only ever send you snacks that you’ll actually eat & enjoy.

Their personalised e-mails are sent at relevant times (the day before your snack arrives) and included useful personal information such as the address it’s being sent to and a direct link to the boxes contents.

Even the Graze website is easy to navigate with lots of shortcuts to the areas you need to manage your deliveries or to rate the snacks you last had. You have the ability to block in bulk ingredients you dislike such as bananas or spicy snacks.

And all this data is feed back to the business to help them create new snacks in the future ensuring they don’t waste time creating foods that their users won’t like.

From a customer perspective, these are all useful features.

Social Media

The idea of letting users curate their own personalised content is something social media is doing outstandingly well at.

Popular forum website Reddit shows new visitors the most popular posts on the general homepage, but as soon as you create an account and define some interests, your homepage is completely personalised to your preferences.

And Reddit users grow their own communities by recommending subreddits or forums to each other when it’s relevant, for example: someone posting about an awkward work situation in a subreddit to discuss relationship problems might then be redirected by another user to try posting in the legal subreddit if they need further advice.

People are talking to people and passing on relevant information, it’s all personal and very useful.

Instagram too is excellent at curating relevant content for their users based on previous behaviour; their likes, the people they follow & hashtags. Then, the explore tab suggests more users and hashtags to follow and it’s all based on user preferences.

The commonality between all these services that give you next level useful personalisation is that they ask you what you like or need or want.

• Graze asks you to rate your snacks.
• Grocery stores remember what you previously bought.
• Reddit tells you to subscribe to subreddits.
• Instagram gets you to connect to friends and search for hashtags to find the content you like and then it learns from it.

My point is, do not be afraid to ask your users questions in order to give them a better experience. They know what they want and in my experience, users are not afraid to tell you what that is. Remember their answer, not their name and give them what they asked for.

Gods of Personalisation

But If you ask me who the Gods of recommendation and personalisation actually are, I’d say entertainment companies, Spotify and Netflix.

Many of you will already be aware of how effective Netflix & Spotify use your previous behaviour to recommend entertainment to consume.

But Netflix is even using personalised artwork to draw you back into the service and encourage you to consume more: If you have shown interest in Stranger Things, then you’re more likely to be shown images of the Shadow Monster. If you like Robin Williams, then you’re more likely to have a photo of him in character.

Here is a direct quote from Netflix regarding this new tactic

“Let’s imagine how the different preferences for cast members might influence the personalisation of the artwork for the movie Pulp Fiction. A member who watches many movies featuring Uma Thurman would likely respond positively to the artwork for Pulp Fiction that contains Uma. Meanwhile, a fan of John Travolta may be more interested in watching Pulp Fiction if the artwork features John.”

And these tactics worked. After A/B testing the personalised images, Netflix reported seeing a significant lift in their core metrics, although they don’t reveal exact numbers.

So, I’ve covered a lot of brands and talked about their mostly successful attempts at personalisation and recommendations, and there’s loads more I don’t have time to write about such as Etsy (whose recommendations are way better than Amazon), or Google who walk the line of creepy personalisation but whom I couldn’t live without.

Start Personalising Now

But, I’m aware that much of this success is dependant on technology, namely machine learning and AI, and not everyone is at the stage where they have a recommendation engine setup and ready to go, but that doesn’t mean you can’t start collecting the data you need to start personalising.

Use your analytics tools to look at audience behaviour. Group your customers into similar segments by looking for similarities in actions and think about how you can streamline or improve that personas experience. Here are some examples of effective & useful personalisation that you can do today

Recently Viewed

Many websites resurface your recently viewed items upon returning to the website.

Even Amazon do it, although they hide it at the bottom of the page for some reason.

Here both Etsy and Thomas Cook have the recently viewed modules near the top of the home page for easy access. Users can simply pick up where they left off. It’s not intrusive but it is personal and it is useful.

Abandoned Cart or Booking

Abandoned bookings is where a customer has completed part of the booking journey but not converted.

At Thomas Cook we A/B tested the idea of having a slide in panel reminding customers of the holiday they almost booked, and we saw a huge uplift in engagement in the pop out itself. There was also uplift further along our funnel proving that this was a helpful feature for the user.

Subtly reminding return users when they’ve left items in their cart is a form of personalisation that you can do today, and it can be implemented via a website popup or via email.

Campaign Consistency

Another method of personalising your website for your customers is by ensuring that they arrive on suitable landing pages for the campaigns that you’re running.

For example, a Facebook or Twitter campaign highlighting a certain product or offer should lead directly to a landing page about the same thing. You may even want to adjust the messaging depending on which channel they come through.

Here, the V&A are running a campaign on natural fibres in fashion and links from their social media posts go directly to a web page where you can book tickets to the exhibition and find out more.

Personalised Segment Messaging

Finally, if you know something about your customers habits or preferences, use it to your and their advantage.

For example, if your customer always buys size 9 shoes, why not have that size filter pre-selected on your search pages? The same goes for clothing sizes .

Maybe you work in travel and you notice a segment of your audience has been identified as a family, so use family friendly imagery on their homepage or highlight hotels with kids clubs.

Or use IP addresses to personalise content or filters based on location. If the user is in London — set the departure airport to All London Airports.

In summary.

Define your audience segments and find their sticking points. Then think about how to streamline their individual experiences.

Ask your customers what they like, what they want, what they need, and then follow through — show you care.

Start personalising now based on preference, not name. And validate your ideas by A/B testing them.

Both personalisation and recommendations need to be useful to be effective.

You’ll see success if you give your customers something that’s useful for them, and I really hope this article was useful for you too.