Piles of data – data coming out of our ears. Google Analytics telling us who was visiting our website, where they were coming from, what country they were in and what age they were and whether they were on an ipad (increasingly, yes). Facebook insights telling us who was looking at a post, who was clicking a post, who was sharing a post with family and friends. Union Metrics drawing us little jelly-fish graphs of the journeys of our Tumblr posts. Our booking system telling us what people booked and when and whether they’d opened emails from us and where they lived…
And almost all of it accessed almost exclusively by the communications department.
Early on I wanted to make sense of this data – I wanted to take it all and find stories and themes. I wanted to use them myself, for improving our newsletters and facebook posts and the layout of our website. But increasingly it became clear that the data told stories that went beyond the communications department – twitter posts raving about a particular evening class in aerial rope were great to share and tell people how much fun it was to train with us – but weren’t they also of use to the department that run the classes and the teachers that teach them? The YouTube comments and discussions of our students’ performances, questions about technique or the choice of music, shouldn’t we be sharing these with the students who created and performed the pieces? When someone told us on facebook how much they loved our Teambuilding workshops but that they found the tightwire shoes a bit smelly – the tightwire-shoes-purchasing-team needed to know that (there isn’t really a tight-wire-shoe-purchasing-team, even in a circus school, it’s probably just the technical department). When record numbers of people were following and asking questions during our online Higher Education Open Day – the HE team needed to know this, they needed to prepare for higher application numbers and more audition spaces.
Presenting data to colleagues
My first set of experiments were around these areas – how we could share ‘our’ data with the wider organisation. Dividing the type of data into qualitative (personal stories, images etc) and quantitive (numbers, graphs). I talked to my mentor, Devon, and we came up with some ideas for sharing:
1. A qualitative experiment: Looking for a way to edit qualitative data down to relevant and manageable sections. I created a Storify that focussed exclusively on our recreational programmes – images and comments, mostly from Twitter and Instagram (see right). Separating this out from the rest of our social media interaction meant we could see how our recreational users talked differently about their training from other users. I shared this with the department who runs the recreational programme. It was reassuring for them to see the positive (people are positive on these channels, they won’t send an email unless they have a complaint but they’ll share a photos of themselves pulling off a perfect gazelle on the static trapeze and talk about how much they love their training). But, while the department loved it, they weren’t sure where to take it next. In conclusion: this experiment worked, but only slightly. We’ve found an easy way to collage and share user-generated content (and have used the same method since, for other areas of work) but in time-strapped organisations who has time to congratulate themselves on all the people loving their experiences when you’ve got five people complaining in your inbox?
2. A quantitative experiment: how could the data we already generate could help us make decisions. I looked at booking patterns for a certain group of recreational classes to decide how we can improve the booking process and try to get some of the sales online. Devon suggested a really simple, easy structure for presenting this kind of data – one page: a question, a graph, an explanation of said graph and a suggested conclusion. Numbers might seem clear and self-explanatory but introduce even a single level of complexity – for example everyone books on a Monday because we actually have a booking restriction in place that means that’s the first time they can book – and leaving the data to speak for i
tself becomes impossible. In conclusion: Again, the experiment worked to an extent – some booking patterns were easy see from the graphs and between us, with knowledge of the vagaries of our systems, we could gather some useful information which helped start conversations. However, my colleagues who I presented the data too felt it wasn’t enough – it answered certain questions but we were left with other ones. Some things we couldn’t know but some things called for more experiments…