Explore visualisation | PDF version
(Please switch off the compatibility view in IE browsers to view the webpage correctly)
For the 2012 MRG conference, Google and Ipsos MediaCT teamed up to look at the effect that three screen ownership (PC/Laptop, Smartphone and Tablet) has on online and offline behaviour. We specifically looked at the role that tablets played, with the aim of understanding whether tablets created new online behaviour or cannibalised existing online habits.
In order to fully investigate the behaviours of 3 screen users, we collected data in three different ways:
Qualitative work – We took one device away from 15 three screen users and asked them to write a blog about their experience without this device for 5 days. This stage enabled us to explore the importance and the role that each device plays.
Mobile Diary – We asked 243 people to log/diarise every time they used/accessed a media device over the course of a day. We also asked them to log/diarise where they were and what they did on each device. After respondents completed the diary we followed up with an online questionnaire. This allowed us to gain a further understanding of up to 4 of the activities they logged/diarised. From this stage we were able gain an understanding of the reasons and motivations for using each device, and whether or not each activity was unique to that device or could be replaced using another device.
Passive Measurement – This involved 53 people downloading an app or meter onto each of their 3 devices. This allowed us to measure when they used each device, for how long and what they did.
For the first two stages of this project the outputs we received were fairly straight forward to analyse and present. However, for the passive measurement stage we were faced with a vast amount of data that needed to be analysed and presented in a clear, simple way. The most complicated element was to show when and for how long people use each of the devices throughout the day. After several internal discussions we decided upon the data visualisation below, which shows at which point of the day people use each device and how this differs depending on whether it is the weekend or a weekday.
What can be seen is that three screen users are online nearly all day, sometimes on more than one device. When looking specifically at each device; mobile use is fairly constant throughout the day, with short time ‘snacking’ use, while tablet is more frequent in the evening and users spend more time per session than mobile. PC use is less frequent overall, but for longer periods of time, suggesting more focused online use.
by Imran Abdul-Hakeem
and Ian Jarvis
Our internal client from Ipsos MediaCT (Media, Content and Technology) came to us with heaps of data from a behavioural research study carried out amongst users of three screens - PC/Laptop, mobile and tablet. The study was for the multinational "Search, Ads and Apps" corporation - Google.
They wanted a visualisation from us to better understand the online use of, and relationship between, the three screens. The data came from 53 respondents who installed passive meters on all three devices to measure online browsing and app usage, recording URL level events, time spent online and device used for the event. So, essentially what we got was aggregated data, such as average number of minutes per day on each device.
The first thing we did was to sit down with the client and ask them to explain to us the aggregated data, so we could understand and map the whole dataset. In principle, the initial analysis led us to the idea of using the data to identify interesting examples of individuals who represent the wider sample trends, which would give us the total length of time used for the device per session and note overlaps in their usage.
Initially we had been experimenting with how to set this up, using Adobe Illustrator as a starting point, but once that proved to be too complex and time consuming, we tested a few things in Excel and found we could save a huge amount of time working this way. The data was sorted and then plotted as a multi-series donut chart. Having colour coded the source cells to reflect their activity, we used a simple macro to colour code each segment within the donut. When you consider we were plotting a total of 72,000 minutes across the respondents, this macro saved us at least three day’s worth of work.
The charts were then taken into Adobe Illustrator, the appropriate colours were assigned to each device and patterns started emerging. We are sure that viewers will learn something, or understand something better, after looking at this visualisation.