Image of a honey badger - social data is the 'honey badger' of personalisation

It’s digital marketers’ ultimate prize: consumers fill up the internet with their ideas, preferences and interests. (This is the big, big data, people!)

Turn that into real-time offers on your website, and jackpot! Angels sing. All the web’s kittens dance. And you’re badder than honey badger overnight.  

 A whole host of technologies have emerged to make this happen. Marketers have no problem finding solutions to personalise their websites.

The core challenge remains: what’s the best way to do personalisation?

Go to the source data

In June last year Econsultancy and Adobe published ‘Personalisation, Trust and Return on Investment’, which concluded:

  • Over 52% of respondents agreed that ‘the ability to personalise content is fundamental to their online strategy’.
  • 41% are ‘committed to providing a personalised web experience’.
  • Only a small proportion of companies are actually able to provide personalised content in real-time. 

The survey explored the sources of data brands use for personalization. Of these, only 6% of brands use social graph data. Significantly more users of social data felt the impact on ROI was high than the equivalent proportion of users of any other source of data (such as purchase history or web behaviour).


Impact on engagement told the same story. 88% said that impact was high when using social data whilst 74% said the same of purchase history and 72% of behaviour on your web properties. 

Econsultancy & Adobe Personalisation Report - Figure 6 

The social data uplift

What’s the difference with social data?

First of all, it’s the best proxy for real life we have, and therefore both a broader and a more accurate expression of intent. Social networks are full of user-generated content – content that people give (for the most part) without gain.

Once you’ve filtered out “likes” for discounts and offers, you have the most holistic source of intent around. It reaches across relationships, interests and roles, and spans trends. 

Purchase data says “I purchased something before” and behavioural data says “I can find a pattern by looking at where you click”. This tells us what specific purpose a customer had in coming to the site once – seldom a great indicator of who a person is, what she loves or what she wants to do next time.

It’s one moment in time. For loyal customers, you can build up a persona in time, but not in real-time (and I’d argue, not to the same quality or accuracy).

If you combine behavioural data with purchase data, you get an improved picture; stores like in the US can do this very well. Some do it too well

If you are really trying to get at a customer’s intent or real interests, you want to identify something important to them, but not transitory. Ideally, it provokes pleasant surprise, serendipity. To give this point some context, an anecdote: Having grown up in Kenya, I have a great affinity with Africa in general and Kenya specifically.

It’s not something most people who aren’t close to me would realise. But if you presented interesting articles about Kenya to me on a news site, for example, or art on a retail site, I would be delighted and most definitely delve deeper. 

When looking at personalisation, one of the most important filters is recency. Ad re-targeting is based on the concept of recency and uses reminders to take you back to a site that you have recently left.

Done badly, this can be distracting and annoying. Done well, it can be effective. Using social data is a much more subtle way of doing this, offering someone products or services that are relevant to a user “right now”, after they have recently talked about it on Facebook or shared it on Twitter. 

Marketers are always looking to a find way to target customers at the right time – something that can’t be solved with behavioural or purchase data alone. It can be done with social data, though. 

The social data discipline

Social data does come with some caveats. For example, we tend to operate in spaces – e.g. my friends’ space (Facebook), my work space (Linkedin) and my family space (a private network).

For a great read on this topic, I’d suggest a book called ‘Grouped’ by Paul Adams or some of Danah Boyd’s earlier work on Facebook. Social networks tend not to overlap much between these spaces (we seldom use Facebook for work) so – in deciding which social data to extract and use for your site personalisation – decide what sort of intent is right for your business. 

Using social data to personalise the web experience is a new discipline - it’s an emerging science at the cross roads of behavioural psychology and data.

Because it’s new, it’s experimental but extremely exciting, and as the Econsultancy & Adobe paper and subsequent infographic reveals, the propensity to affect a good result is by far the greatest from social data compared to other forms of web data available today. 

I’d love to hear your comments and feedback on this topic. 


Published 19 March, 2013 by Jonathan Lakin

Jonathan Lakin is CEO at Intent HQ and a guest blogger on Econsultancy. You can connect with him on LinkedIn or follow on Twitter

2 more posts from this author

You might be interested in

Comments (6)

Save or Cancel

Simon Hoff

Hi Jonathan,
great article on how return is influenced by personalized offers. I was wondering what you think about retargeting in general, if it is just not as good as real time personalization or annoying in general?
And would you in general say the more data the better (Geo data, other campagin data e.g.) or is social king in your view?

over 5 years ago


Jonathan Lakin, CEO at Cariga Ltd.

Simon thanks for your comments - a point of view on retargeting. Done badly, it's just horrible. Done well, the results can be impressive. Worth noting 3 points i) it works because psychologically we respond to reinforcement ii) marketers often miss the fact that they are paying for the same lead many times over with a mix of PPC and re-targeting so the CAC is often understated iii) its helpful in triangulating a users "mission" for a previous visit - but not necessarily for "this visit".

On other data - the more the better, so long as there's a way to extract the insight from the combined data. I tend to think about Social data being the prime driver for intent, and what you're generally calling other data being useful for "mode" (at home, in between, at work - we do things differently in each mode) and "mission" - what I'm trying to do right now.

HTH, Jonathan

over 5 years ago


Simon Hoff

Hi Jonathan, thanks for your answer. I like the distinction between intent and mode, makes a lot of sense!
Best, Simon

over 5 years ago

Anna Lewis

Anna Lewis, Google Analytics Analyst at Koozai

Those were probably the best opening two sentences to a blog post I've ever read!

I've found that you can use Google Analytics to target ads on a more personal level with the remarketing feature (, but it's hard to really personalise the ads through these platforms.

Using social graph data would be so much more powerful but as you say, it's underused right now.

The worry is that if more people use it, general web users will become more attuned to ads being personal to them and I know a lot of people don't like it, so I'm not sure how it would change people's opinions or activity. Hopefully it would all be for the best but it might also increase people's awareness of how to remove tracking!

over 5 years ago


Gareth Fryer

Nice post, certainly an interesting read.

I think personalisation through the social graph does have the potential to create sales uplift, based on increased resonance, relevance to and understanding of that consumer. However, I personally don't think that alone gets you 'badder than the bad ass honey badger'. The real opportunity as I see it, is to overlay the social graph, with the interest graph, because you're then matching social relevance and understanding with genuine product experience interest and passion created by a wider network.

To give you a real world example of this: Bondsy is an excellent example of solving a genuine real world problem by matching the social and interest graph.

What this answers is the fundamental problem with sites like craigslist and gumtree. In that, you always worry if you can trust the person you are selling to. By connecting to your social graph and interest graph, it finds people that are no more than two degrees of separation from you (friends of friends), so there is an audit trail, but it also filters those people who are interested in what you have to sell. Effectively finding connected 'strangers', but it's ok because your friend knows them, and they like what you're selling = #honeybadgerdance

That's when social data, can be come really, really powerful.

over 5 years ago


Jonathan Lakin, CEO at Cariga Ltd.

Gareth - I'm excited to read your comment. The Company I founded in 2010, Global Dawn has spent many millions creating an interest graph from social data to drive personalisation. Our biggest challenge has been deriving "meaning" - and super accurate interests from raw social data - a process that involves cleaning to get rid of false interests (crudely those driven by incentives), figuring out a way to match interests against a human like rather than machine based classification system, and then filtering to extract what's relevant now, not last month. I'm curious about Bondsy, so I'll spend some time looking at it later today.

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.