The disciplines of both web analytics and usability have experienced marked changes, driven by client demand to reflect the increased complexity of online user experiences.  Current economic pressures are accelerating this trend further, forcing ever greater emphasis on measuring the effectiveness of online spend. 

And yet there is now a new consideration that promises to shake things up further.  Direct marketers’ offline disciplines of customer analysis and propensity modelling have traditionally been outside the remit of web analytics and usability, but no more.  Bridging the divide between anonymous users and identifiable people promises to make the challenge tougher, but the rewards much greater.

This is, for some, a natural progression; we’ve always been focused on applying direct marketing principles to all channels, approaching this in three steps:

· Step One – Analytics.  Often a ‘train of thought’ approach is needed to go beyond traditional flat reporting; acknowledgement that an answer usually sparks another five questions!  But a heritage in data and analytics helps in recognising the need to drill down beyond the standard metrics and KPIs of current web analytics into the measurements of why and how.  This goes from the basics – what is worth measuring from clicks v visits v page views v duration v sales – to more advanced behavioural and conversion measurements.  Some quick wins can be identified and actioned from this such as the implementation of triggered emails that retarget users who started but abandoned the buy process. 

· Step Two – Behavioural Segmentation.  With flexible analytics, and access where appropriate to third party data (such as surveys or geodemographics), brands can take a step on from traditional usability and start to create behavioural segmentations that create a real picture of a user, and most pertinently not just ‘the average user’.  The full richness of variation in customers and prospects can be discovered and harnessed using the sort of analytical and modelling skills that are the staple of offline direct marketing.  The result of this is that a site need not be designed with all users (or even the lowest common denominator measures) in mind.  Instead a case can be made for delivering appropriate content to small sub sets of users sharing similar behavioural characteristics. 

· Step Three – Synaptic Marketing.  This is the term used to describe the synthesis of two processes – generally traditional and digital media.  Here the output of both the above steps can be applied across channels in near real time.  It is a living and dynamic proposition that moves away from batches to live environments and, as such, is far removed from simply merging datasets to achieve ‘integrated marketing’.

Above: Ben Langdon, DMG

The principle behind synaptic marketing is the collection of cross channel customer data, mining it, creating intelligence from it and applying that intelligence to marketing communications.  Crucial to this is understanding an individual users’ site behaviour alongside their offline behaviour, to create a profile that then drives the messages and the channels through which they're delivered. 

The result is a deeper knowledge than is generally gathered from online interactions alone.  It can even supplement some of the excellent work being undertaken in triggered environments.  For instance, we run a number of programmes that trigger an outbound call to a site visitor who was driven to the site from an email.  We know when they arrive, where they’ve come from and what we can do to help them.  Synaptic marketing seeks to go beyond the intelligence behind a single real time event and to look at the overall picture of the user to drive a wider range of events.  

In an online environment this means that, when a user visits a site, it should trigger the question ‘what should this user be shown?’  By assessing the intelligence held about that visitor based upon web analytics and behavioural segmentation, the system returns the appropriate content for display online, but maybe also via other channels such as mobile and email, as appropriate. 

These channels are increasingly important as we all realise that purchasing decisions are never made in isolation. The days of online and offline being independent disciplines are long gone because every exposure that a consumer has to a marketing message can contribute to a final buying decision – wherever that decision is actioned.  This is the key to a synaptic approach.  If a brand wants to influence that decision then it must control those messages and recognise their impact.  This is the big challenge of today’s environment and it demands more than just a single ‘hit’ view of a campaign impact.

The most complex area of deployment remains the web.  For too long the promise of real time targeted messages has remained unfulfilled.  Online marketers would scoff at offliners sending large mailing campaigns that take months to plan when they are able to instigate online campaigns in a tiny proportion of the time.  Offline DM'ers however would point to the fact that those online campaigns are targeted at often bland non personal profiles; the perceived USP of the web was simply not being realised. 

Web personalisation seeks to redress this and it has of course been around for a while.  Yet one of the inhibitors of enterprise wide web personalisation has been the large scale data handling requirement, especially where this includes offline customer service history (retailers being the prime example) because it entails identifying and merging vast volumes of data from disparate silo’s.  The way round this, via the process of synaptic marketing, is to take a data-lite approach based on analysing the key necessary data to meet identified objectives – so there are no enterprise wide systems integration requirements.

The practicalities of all of this are key; creating a picture of a user and delivering appropriate content to that user are two very different things and an in depth understanding of real time communications (from Content Management Systems to email and mobile deployment platforms) is a prerequisite to delivering on this vision.  Creating a complete end-to-end solution relies on it; a collection of API’s will not deliver.

Web personalisation has in the past been about limited multivariate testing in short bursts, but the new principle delivers an analytical environment in which rules can be set up, tested and deployed on an ongoing basis.  This means that, with focus, clients can establish a wide ranging set of hypotheses to be tested with few limits – thereby giving the business far greater flexibility.  This environment is the client’s laboratory for testing new ideas as well as an engine for delivering them.  

Of course, all of this starts with web analytics.  It’s likely that in the current economic environment, clients will be shining an ever greater spotlight on their spend and, in online terms, they’ll be looking at how best to measure this.  They may start by asking why they should spend on a web analytics function at all, especially when they have free products from major names such as Google and (shortly) Yahoo.  The answer is simple - how far down does one need to dig to discover these data nuggets? 

It’s all ultimately about intelligence - using data to drive messages at a group or individual level based upon all that we know about that person at that time.  It all starts with web analytics because there is so much gold to be found here.  But to uncover this treasure an evolved and joined up approach is required.