Personalisation in retail is often seen as the latest development in online marketing but the practice itself is as old as the concept of retail.  

From the time of the earliest shopkeepers, good retailers would recognise their customer and tailor their pitch according to what they knew about them. 

A clear example of this sort of personalisation is a story from the history of Lock and Co Hatters of St James. In the 18th Century, so the story goes, a trusted customer could simply descend from their carriage outside the door of the shop, shout ‘hat’ and then leave.  

The shop’s workmen would then refer to their records for details of the customer’s hat size and style preferences before making the hat and delivering it to their residence.

This highly nuanced personalisation (if not to the same degree) formed the basis of all good retail for hundreds of years and it was only with the rise of supermarkets and chain stores that retail began to get impersonal.

The web, when it arrived as an engine of commerce, picked up where the faceless mega-chains left off. Early online retail made little if any effort to personalise their offerings, assuming instead that all customers were more or less the same.  

Now, however, personalisation is very much back on the agenda. Falling margins driven by intense competition and rising customer acquisition costs mean that every conversion counts. If a personalised experience is what it takes to get you to buy, that is what retailers must do.

That said, not all personalisation is equal.  There have been two distinct phases of best practice to date and we believe that there is at least one still to come.

1: The era of Amazon

Like so much in online retail, web personalisation was first introduced at scale by Amazon. Their ‘customers who bought this also bought that’, introduced a degree of personalised recommendation to the sales cycle based on your previous shopping habits.  

However, it’s hard to call this ‘true’ personalisation. It’s more like simple recommendation which, whilst similar, falls some way short of the dream of an online experience built to your true preferences.

Moreover, this sort of recommendation requires a relatively advanced level of interaction before it can function to its full extent. Most importantly, it requires you to have expressed a strong retail decision before it can recommend – you basically have to buy something before it can suggest what else you might want to buy.  

This is fine but the people that most retailers want to convert are not the 3% of people who have converted before, but the 97% of people who never do. Targeting prior customers is good practice, but its driving additional revenue from a small existing base rather than growing your overall market share.

2: Big data personalisation

The next stage of personalisation, and one which we are seeing now, is more akin to the ‘real thing’. Big data approaches to personalisation harvest huge quantities of information about individual user journeys and use this to serve bespoke content based on an understanding of patterns of behaviour.  

For example, analysis of site journeys might identify that people who visit two category pages without putting anything in their basket are susceptible to pricing offers.  

It might therefore serve a bespoke discount offer as a message layer to those users when they arrive at the second category page, hopefully driving conversions.  

Alternatively, analysis might reveal that those arriving from specific search keywords will tend to be interested in a specific product area or attribute, enabling the retailer to serve a tailored home page to those users.

This personalisation meets the barrier of ‘real’ personalisation, as it delivers a bespoke experience based on an individual user’s behaviour.  However, it does require a fundamental change in the way you manage and view your customer level data.  

Rather than managing data about attribution, on-site activity, purchase history and the rest across multiple different software silos, for this approach you need to gain a single, easily interrogated view of the customer and then use this to build an understanding about aggregate patterns of behaviour.  

Only with this insight can you then hope to deliver truly insightful personalisation campaigns.

3: Minority report

The final stage of personalisation is akin to that displayed in the Tom Cruise sci-fi film Minority Report, where personalised real world advertising and retail experiences are delivered by holographic projections.  

The personalised website of the future, for this isn’t an experience anyone can yet deliver, will recognise the individual visitor and will then build a radically personalised experience based on a deep understanding of their preferences.  

This might be to display a bespoke product range, particular sizing, colour or pricing preferences or even an entirely bespoke site design based on their visual and navigation preferences.

Clearly there are some barriers to the development of this sort of personalisation. First and foremost, this sort of customer insight would require a new agreement between retailers and their customers about data collection and usage.  

Customers would need to both feel confident in retailer’s ability to keep their data safe and also in the value to them in sharing detailed personal preference data.

Moreover, this sort of personalisation has a real impact on the way that brands present themselves. It’s going to be hard to maintain a single look and feel for a brand’s website when every page is potentially bespoke to the individual user.  

Clearly brands will have to make the call as to whether they value the sale more than the minutiae of their brand identity.

These barriers aside, this type of personalisation must be the Holy Grail for all online retailers. Hundreds of years of commerce have proven the efficacy of personalised experiences and this is surely the closest that the web can get to delivering the sort of personalisation and convenience that St James’ hatters did all those years ago.

Ian McCaig

Published 11 March, 2013 by Ian McCaig

Ian McCaig is Founder at Qubit and a contributor to Econsultancy.

29 more posts from this author

You might be interested in

Comments (11)

Save or Cancel
Dan Huddart

Dan Huddart, Head of Analytics & Web Development at RSA Group

I'd suggest personalisation will become the holy grail for most online companies, not just retail, although the competitiveness of retail makes it a bigger factor at the moment than for other industries.

I'm seeing great results in finance, telecoms, and utilities already. Customers will soon expect it.

over 5 years ago


Depesh Mandalia, CEO & Founder at SM Commerce

(the first link 404s...)

I love the first example. Of how we've gone for small niche, personalised retail experiences over the centuries in the real world to functional digital experiences in the last decade. Easy to forget online is still in a comparative infancy stage.

Its the combination of social and personalisation which intrigues me most, since the web is impersonal by nature (one person on one device) how can a business make you feel valued in the digital world? How can a website know you have a family, or you're planning a vaction soon? Does personalisation only start when you visit their website? What about other touchpoints?

It's all to play for!

over 5 years ago


Oxaz Ixonirivo

I would love to believe in this vision, but where is the evidence for your apparent premise that you can meaningfully and effectively target customers that you know very little about? For many of the 97% of non-converters that you claim are a massive opportunity, you will only see for a handful of pageviews at most: obviously not enough for them to shout 'hat' and be offered their ideal hat. Perhaps enough for smaller, marginal improvements, but again: where is the evidence?

It also looks a bit strange to depict Amazon as being one step behind and not delivering 'true personalisation', when in fact they seem to be (ten years ago as well as today) driving innovation in this space.

Is there not also a case to be made that too much personalisation may lead to impoverished user experience? I am only talking about the 3rd 'minority report' step here. I would be freaked out to see a completely different website from my son just on the basis of some algorithm deciding we have radically different preferences. Why does he get cheaper prices on some products? Why is his landing page different from mine?

Such a 'deep' understanding of my preferences could also come at the expense of serendipity. I love to be exposed to products and content that challenge my preferences; maybe it will not make me convert more, but perhaps it may help the website in ways harder to measure, e.g. I may talk to my mum about that product that I think she may like.

The industry's quixotic focus on personalisation, I feel, is also at risk of coming at the expense of just offering a better experience to everyone. A retailer designing a user experience for a small segment that is anticipated to make up 5% of their user base will not spend as much effort and money and will not test and optimize it as rigorously as if the change was meant to rollout for everyone.

Pehaps the author is right to believe that personalisation is the key to the future of ecommerce, but this piece would benefit greatly from a more balanced, pragmatic view on the subject. This is an empirical question begging for a serious empirical inquiry rather than for wishful thinking.

over 5 years ago


Tom Sheepshanks, -

@Depesh. I agree with your point about combining social & personalisation being the place to be. For me (and I'm bias due to who I work for) it's the ability to personalise using social data. There's a vast amount of untapped information sitting within social networks that can be used to personalise. How can a website know you have a family? Because you've explicitly said so on Facebook. Not only that but you can infer their interests, the interests of their friends, and begin to dynamically adapt their experience of a website using that information.

over 5 years ago


Martin Lowther

"For example, analysis of site journeys might identify that people who visit two category pages without putting anything in their basket are susceptible to pricing offers.

It might therefore serve a bespoke discount offer as a message layer to those users when they arrive at the second category page, hopefully driving conversions."

Lets for a moment imagine that such an analysis was done on 100,000 customers of a retailer. From 90,000 odd non-converting users, let say 25000 are users identified by such an analysis. Now, who is willing to give some substantial discount to those users? Why would the company take a hit? Will amazon do it? Well no, given their profit margins they would not. Any other retailer? chances are less. Unless you are high profit margin business willing to invest so much in technology for such analysis. And frankly, only god knows how many of those 25000 would ever convert, and even if they did what has the retailer achieved? What if this was a one time conversion. To many unknowns to put anyone's money on.

over 5 years ago


Martin Lowther

And now my comments are getting removed! Paid article and paid commenters!

over 5 years ago


Oxaz Ixonirivo


I also commented twice, expressing slightly dissenting views but (I thought) making decently well-argumented points.

Twice, my comments got deleted, with no warning or explanation.

For shame, Econsultancy. For shame.

over 5 years ago

Graham Charlton

Graham Charlton, Editor in Chief at ClickZ Global

@Martin @Oxaz This blog has an Akismet spam filter which marked your comments as spam, sorry about that. I've now published your original comments.

We receive hundreds of spam comments every day so I'm afraid the filter is necessary, though as we have seen here, it isn't perfect.

Btw - we don't do paid articles or paid comments and we don't censor comments.

over 5 years ago

Carlton Jefferis

Carlton Jefferis, CEO & Founder at Gettus!

Hi Graham. Off the topic of the post but re the spam filter: I wonder if you could place some brief advisory text somewhere in the comment submission area so that people understand the existence of the filter? This could be further enhanced by providing a user with a direct way of signalling to Econsultancy that their legitimate comment has been caught by Akismet.

And how about lowering the spam filter criteria a tad and instead providing a way for the community to mark comments as spam? Something like a 3 strikes and you're out system. I'm sure we're an active bunch with a thorough dislike of spam!

I realise none of these suggestions is perfect but I also know from experience how frustrating it is to have taken time to write a comment on this blog for it to then be blocked for no apparent reason. It's a dreadful user experience.

I doubt you have time to sift through the hundreds of comments marked as spam each day, so providing a way for commenters to help you identify the valuable ones more quickly would help avoid the cynical (though completely understandable) comments seen above.

over 5 years ago

Graham Charlton

Graham Charlton, Editor in Chief at ClickZ Global

@Carlton That's not a bad idea.

The thing is, the workings of the spam filter are something of a mystery, and I don't think it's as simple as lowering it a little.

I'd love to remove the filter and let the community police it, but the sheer volume of spam makes that unrealistic. While newer posts may be policed, older ones would fill up with spam.

I understand the frustration of anyone who has left a comment only for it to disappear and that some people will think they are being censored. I want to encourage more comments so we will look into alternative ways of managing spam.

Btw - if anyone has this problem, email

over 5 years ago



Do you have any video of that? I'd want to find out some additional information.

over 5 years ago

Save or Cancel

Enjoying this article?

Get more just like this, delivered to your inbox.

Keep up to date with the latest analysis, inspiration and learning from the Econsultancy blog with our free Digital Pulse newsletter. You will receive a hand-picked digest of the latest and greatest articles, as well as snippets of new market data, best practice guides and trends research.