Thursday, 10 March 2011

Using Data to Drive Online User Experience (a quick look at some of the opportunities)

Cloud Computing, the Semantic Web, Linked Data, Web 2.0, Web 3.0...

...all this coupled with other data you've got e.g. web analytics, BI, MI, Research etc. means that there is probably increasing amounts of data available to you and your E-business team. Unfortunately there probably isn't that much increasing insight. Why this is the case is for another article, perhaps by someone else. For my part I'm just interested in talking a little bit about tapping into this data (how) and doing something with it (what/why) to improve user experiences (UX) and/or to sell a bit more online.

Lets start by chucking all the data sources you can think of at this point into one data pot, and perhaps ignore data sources completely. I'm not just suggesting this to make it easier to write this article. I'm also doing it because that's all it is, really - a big load of data. Until you start to work out what it means that is.

How you get at this data is through technology (services, API's etc.) but from an E-Business perspective that is just a necessary mechanic (albeit a very important mechanic and one that that will require very smart people to be involved). For me though it all starts to get a bit more exciting when we start to explore how we can turn the data into insight or start to put further mechanics in place for using insight to shape an online presence for prospects & customers.

An obvious starting point (but surprisingly little used in E-Business) is to explore your data world without looking at the data first. For example, taking a look at how it is already being put to good use in other channels, and thinking about how you might do the same. Or by taking a further look at your customer and your 'total customer experiences' (e.g. a multi-channel, full life-cycle view). You probably do this all the time in your organisation. If your not already, perhaps this time make sure that instead of looking at the 'what' start to explore the 'why'. If your customers are doing certain things when you ask them to, it is important to know this, but perhaps more important is to know why, so that we can think about how we might apply this to your digital opportunities as well. For example, if in a DM campaign customers are successfully taking up a second credit card when a particular promotional APR is applied across both. Is it the offer that makes the campaign successful or also because it is aimed at people who are more likely to be in a relationship, or who have a good 'referrer' profile? If it is both then you can think about how you might identify these people when they are browsing online and target the same message at them. This will have a few benefits:

1. Maybe obvious - but you will improve the relative performance of that campaign.
2. You should improve your cost per acquisition (CPA) and possibly budget for other campaigns.
3. You'll reduce the noise for people who are more unlikely to be interested / be able to take-up the offer.
4. By virtue of 2. & 3. you will be able to offer those people something that's more relevant to them

From a purely customer or user experience perspective this use of data can be applied in different ways. For example, if your business objective is to only drive calls where they can be defined as 'value adding' (generating some kind of value contribution) then to achieve this perhaps only offer this channel at particular points in a journey e.g. when a difficult purchase decision needs to be made. Or perhaps offering this service to different people at different times, where you know there are likely differences in these types of people e.g. people of xyz demographic exit at this point online but we know they buy this successfully offline. These sorts of solutions can also be put in place until you have properly identified 'why' certain things are happening, and until you can change the experience online to overcome those problems. Until then though, making the sale can still be achieved using the data in this way.

To supplement this, digital channels can provide even further data/insight opportunities and to a degree 'self optimise' using various technologies e.g. Omniture Test & Target. These optimisation tools are often used to help determine optimal outcomes e.g. web pages, that generate the best KPI's. This is good at telling us 'what' works best, but it often doesn't really tell us 'why'. It therefore becomes important to start to look at why. Who is it working for and who isn't it working for? What could be driving this? These are the sorts of questions we answer all the time in User Centred Design (UCD) and through various research studies, and this insight is essential to driving the best outcomes. It then also becomes really important to identify all of the relevant moments in your user experience and start to optimise them all, but also to move the experience forward we perhaps need to make web experiences more semantic and start to offer different people, at different times in different places different things. The potential of infinite combinations of outcomes is scary, and I'm not suggesting this end point is easy, or always necessary to get to, but to a degree it is already happening in sites like Amazon and will be become increasingly common and should be a consideration when thinking about data and online experiences.

These are just some of the ways we can start to apply data to improve user experiences online. And perhaps thinking about it in these terms rather than as a full blown, company wide, fully intelligent use of data (and the 'cloud') from day one, we can just start to use some of the insight data can generate a little bit more often and build our E-data world to further improve online experiences.

No comments:

Post a Comment