When is Personalisation Not Personalised?

By James Doman | 08 Feb 2012

Are you embracing true e-commerce personalisation or merely paying lip service to the concept? James Doman explores the heightened potential of embracing genuine personalisation in online retail.

Working for an online retailer can be hard: not only do you have to deal with products, stock, SEO, PPC, ROI and, finally, customers. But you’re also probably fielding off calls every day from eager salespeople, all touting the next big thing.

The global success of ecommerce has spawned hundreds, if not thousands, of “add-ons” from social media widgets to powerful analytics. But this year, you’ll be hearing a lot about one particular technology: personalisation.

This article is about real personalisation. Not wishy-washy “people who bought this also bought…” stuff – because that’s not personalisation. This is about real, powerful, effective personalisation that doesn’t just recommend products – it creates relevant experiences.


I learnt to write effectively by starting with an inspirational quote that sets the scene, so here’s my choice:

We have 6.2 million customers, we should have 6.2 million stores.”
Jeff Bezos, CEO, Amazon Inc. – 1999.

Yes, you read that right. Thirteen years ago, Jeff Bezos was sitting in his office in Seattle, thinking about how they could make the perfect store for every single customer they ever had.

Wikipedia defines personalisation as “…using technology to accommodate the differences between individuals.” And whilst that’s a much generalised definition, it works. But let’s go deeper.

Here’s my definition:

Personalisation technology enables the dynamic insertion, adaptation or suggestion of content that is relevant to the individual user, based on the user’s implicit behaviour and preferences, and their explicitly given details.

Essentially: personalisation should be personal to each individual visitor.

Personalisation can be either the adaptation of page content, or just the suggestion of other relevant content. It can use both explicit details, such as gender and age, and implicit preferences such as “I prefer blue”, or “I’m more likely to buy Levi products”, mined by making sense of the user’s behaviour.

The important bit

This is a plea to all “personalisation providers” in the world: personalisation isn’t the crowd – please understand that!

Whilst useful in its own way, “People who bought X also bought Y” isn’t personalisation, so please don’t call it that. Call it what it is: simple product recommendations.

“Recommendations” are different from personalisation, in that they look solely at the crowd. Here’s a real example from the UK.

John Lewis is a renowned department store, and recently implemented “personalisation” technology. After browsing around in the Men’s clothing category, I decided to look at a suit jacket. Halfway down the page, they suggest some products. Take a look.

I bet you’re creatively imagining what on earth anybody could do with a lacy thong, tights and a meat thermometer. It’s a prime example of where “recommendation” technology just doesn’t cut it. At the very least, the site should have known I’m a man. It could have gone one step better, and understood I was looking for formal clothing – why not suggest me some trousers, a suit package, white shirts, a tie or a bowtie? In fact, suggest me anything more relevant than a meat thermometer and hosiery.

It feels like they’ve just said “Oh, we need product recommendations – find something, plug it in and let’s forget about it.” There hasn’t been any thought behind what they actually want the block to do. Should it upsell more expensive products? Or should it cross-sell relevant items?

Of course, there could be some clever, Jedi-style mind tricks going on here. On second thoughts: maybe not.

Personalisation is…

I’m not going to go into detail about how personalisation works – that’s another post. But I just want to demonstrate how personalisation, particularly on the product page, can be done well.

Beacon Lighting actually thought about how they wanted to implement product suggestions on their product pages. Firstly, they wanted to make sure they showed the right globe that fitted the product in view. Secondly, they engaged true personalisation technology and strategy for the  “You may also like” section.

The “logic” behind the suggestions firstly looks for the most popular cross-sell, making sure it’s from same range. It will look for a popular upsell: a similar product (same material, size, etc.) that is more expensive. Finally, it will fall back to other popular, similar products, regardless of the price point.

As you can see in the screenshot, it provides much more relevant suggestions – making for a more effective shopping experience. It’s guessing the intent. The visitor has clicked on a lamp with a wooden base – there’s no point showing them something irrelevant.

So there’s an example of the move away from “recommendations” towards true, real personalisation. And the opportunities are limitless.

Personalisation could be…

There is so much scope for personalisation from just powering suggested product blocks on a page, to driving search results, homepage images and even adapting the whole look and feel of a site. Personalisation will become a huge part of the internet as more companies experiment and innovate.

Here are a few ways in which personalisation can make a difference.


Ladies, imagine it’s your birthday soon: you want to treat yourself to a new outfit. You’re shopping online at your favourite clothing retailer, and everything is going well – until you search for a generic term like “t-shirt”. The first page of search results returns only men’s t-shirts in lime green.

New technologies available today enable retailers to do Google-style personalised search – not only can it recognise you’re a woman, but it knows that you prefer the colour purple and that you’re a size 10 – subsequently bringing purple t-shirts that are in-stock in your size to the top of the search page, without needing to refine through faceted navigation.


Why are you showing your most popular products on your homepage, when you could be massively increasing the chances of a conversion by putting the most relevant products in front of every new or returning visitor?

Here’s a comic strip that demonstrates how relevant your homepage could be.


Putting the right products in front of the right people at the right time isn’t hard. It’s not just suggested products – it could be images and banners, you could adapt the text or suggest a relevant blog post. You can turn your site into a personalised, relevant selling environment, optimised to encourage purchases.


Many retailers are at least delivering some level of segmentation in their email marketing – but are they really relevant to the individual? Retailers no longer need to “Spray n’ Pray” – they can send relevant emails with personalised content and product suggestions to every person. Transactional emails, promotional emails and reactionary (e.g. cart abandonment) emails can all be relevant to the individual.

To summarise…

Product recommendation technology just isn’t the same as personalisation technology. Unless your site is understanding that a visitor prefers a certain colour or likes a particular brand, and acts upon the data in real time, you’re not personalising – you’re just recommending.

Personalisation is about the individual, their behaviour and their details. However, there is a certain amount of crossover with recommendations – I think the best way to personalise is to filter crowd recommendations by personal preferences – but again, that’s another post.



4 thoughts on “When is Personalisation Not Personalised?”

  1. What a hilarious range of suggestions but I can see their crazy logic: You’re a man therefore you love barbecuing and here’s a meat thermometer. You’re a man and you love to see your woman in nice lingerie. It’s not what you would expect, but hey, it got your attention!

  2. James; If you look at the strategy, it is personalised; you are just running across female categories to demonstrate the recently viewed strategy. Struggling to understand why you would do that? Personalisation, as you say, is enticing the customer into preferences in real time but with an assurance that your context of now is the key. I would enjoy watching your browsing history as it clearly demonstrates your behaviour for this article.

  3. James Doman says:

    Personalisation Expert, as I said in my post, there could be some clever mind tricks going on. But I doubt it.

    Visiting the same product page, direct through Google from a clean computer in Private Browsing now gives me the same thong, the same deep v-neck pyjama top, a throw and a table/floor lamp. But it seems like the collaborative filtering is working better, as I am actually getting a pair of trousers and a shirt now!

    Go on, give it a try: http://www.johnlewis.com/146488/Product.aspx

    And now, I’m going to look at two other men’s suit jackets. I’ve refreshed the page… and nothing. Exactly the same suggestions. Maybe I’m doing something wrong? What is your definition of “personalisation”?

    Surely product-page product recommendations should be upsell – our research found that they perform 20 times better than cross-sells. So recommendations should be restricted to the most popular similar more expensive items.

  4. Amit says:

    too good ,opens my mind with a new ideas thanx

Leave a Reply

Your email address will not be published. Required fields are marked *