Lana Gibson gave a talk at the AgileNZ conference on using analytics data to design content, based on her experiences as Content Performance Lead working on the GOV.UK whole of government website.
InfoQ: Thanks for talking to us Lana – your role at gov.uk sounds interesting, what does a Content Performance Lead do?
Lana Gibson: It’s a unique role, helping the content team to listen to the voice of users through data. The main task was to help the content team use data in a practical way to make content that’s really user focused.
InfoQ: Your talk was using analytics data in terms of the design of content. Aren’t content designers and analytics almost opposites?
Lana: Yes, definitely and one of the big challenges was to actually cross that divide. We had to move away from the traditional data model, where analysts go into a dark room for three weeks and create masterful spreadsheets with hundreds of metrics that never get used to actually improve things.
There are fundamentally different viewpoints between “data people” and “words people”, and to operate effectively in an Agile environment you really need to have collaboration, you need to make what you do understandable and applicable to other people. So it was about transforming that raw user data, taking useful insights from the data world and giving them to the content team in a usable form. And the tools were a big part of that, both commercial tools such as Google Analytics and Htiwise, but also the tools we created in the team such as dashboards and automated reports.
InfoQ: One of the points that you made was in fact that good tools foster collaboration. How do they help? And what does collaboration look like in that regard?
Lana: In the past it felt like tools prevented collaboration. For example when you created a document you’d have to finalise it perfectly, and then send it out to everyone who’d make their individual changes on their individual versions. It couldn’t be revoked, you couldn’t make changes yourself in response to feedback without sending it out again, and it multiplied into numerous versions. It was this crazy world where your work needed to be perfected before you shared it, then you completely lost control once it was out there. Whereas with shared documents everyone can work together and have an evolving dialogue about what needs to happen.
I think that’s a really good analogy for a lot of stuff that goes on in Agile teams - if you’ve got your experts working in silos you’re not going to create good, coherent products. It’s only when people can collaborate effectively that they can make streamlined products. It’s not easy for us to get rid of our egos and understand the different languages that go along with each digital discipline, but it’s really necessary to understand how our contribution fits into the whole.
InfoQ: Another point that you made was data is the voice of your audience. How do we hear that voice?
Lana: Through any kind of data we can get our hands on - analytics, call centre tracking, on-page feedback, email enquiry form, sales data. Review it all to see if it tells you something. In terms of using it effectively, I’ve found that most people have two approaches to data. One is that you can just turn it on and millions of insights just come out. Or the alternative viewpoint is that people think that it’s unusable, that it’s just numbers that don’t reflect the user experience.
These are two opposing views, whereas I think we are somewhere in the middle. Uou need to use it sensibly - work out what you are trying to achieve, and make an assumption about what users are currently doing on your site. If you put yourself in the shoes of your users, start off testing small assumptions about your users and making small tweaks to improve their experience, you’ll naturally branch out and find more insights. Eventually you’ll build up a comprehensive picture of your audience that can inform all of your work.
InfoQ: How do we design these dashboards and tools to actually be useful to people like content designers?
Lana: Start with brown paper! I learned that lesson early on: I finalised diagnostic measures for a project too soon and made a very complex tool right at the start. It was based on what my team thought they needed, but it didn’t actually give them what they needed.
As we were saying earlier a huge part of making data effective is creating shareable, usable tools. So start out with a workshop to find out what your team needs, then get some data in front of them. You have data - you have call centre data, you probably have people commenting on your pages, you might have emails coming in through your site, you have analytics. Make a reasonable guess about what you think will work based on their feedback, put some real data on some brown paper, and put it in front of your team quickly.
Get everyone engaged and find out if it’s useful. Find out how it tracks important things, what it tells about your users and your site, and then iterate based on that. Once you’ve got a useful measuring strategy, bang on about it. Show how it can prove success in stand ups, and make fun things like stickers that remind people to use it. It’s a process to get data useful and used.
InfoQ: It doesn’t just come out of the box ready.
Lana: Unfortunately not, tools like Google Analytics are very smart. There are so many things they can do that it gets overwhelming. It takes a while to emerge from the rabbit warren and understand what will help you to improve your website. But once you understand your users you can deliver measurable improvements and run a much more effective and efficient site.
InfoQ: Lana thank you very much for taking the time to talk to InfoQ.
Lana: Great, thanks Shane.
About the Authors
Lana Gibson helps people improve their websites by turning user data from search, analytics, on-page feedback, call centre and sales into snappy customer insights and clear measurement strategies. She provides insights and tools that help businesses understand their audiences, measure their performance, and achieve their goals.She also speaks at events and blogs, usually about the need to shape digital products around user needs rather than organisation assumptions in order to get them useful and used.