10 Best Or Worst Ways To Visualise Web Analytics Data

Web analytics data is worth nothing if it can’t be used in the correct way. In order for it to be used, it has to be understood. So how do you make data easy to understand? With visualisation of course! But visualizing data isn’t as easy as you might think. You have to get it right for both the user and the data.

As Ian Laurie, founder of Portent, said in his 2012 SES talk about Data That Persuades, “Your career isn’t just about being right, your career is about being understood.” You obviously have to get the data right, but what’s the point in giving the right number if no one understands it?

You don’t need to quiz them on what they like (although this may sometimes be beneficial – especially if it’s a one off analysis) but test what you think works best with the data and see how much explanation they need in order to understand it, then simply adjust for the next set of data. Feedback Is Essential!

I’d highly recommend going through Ian’s slides on the link above (they have explanatory notes) to see how he describes the best ways to get your data to persuade the recipient in the way that you want – and firstly to persuade them to even read it!

On top of this, earlier this year, I attended a Big Data and Marketing Conference at ESCP Europe and heard Judy Bayer from Teradata emphasise that analysts should be story tellers. This resonated with me as it is through stories that your data gains context and purpose, making it easier to understand and gain insights from.

So having covered why you need to think carefully about your visualisation, here are 10 common web analytics visualisations. You will have to make up your own mind about whether or not they would suit the story that your data needs to tell, as every data set and situation is different, but hopefully they will give you some ideas and inspiration.

From a couple of numbers in boxes, to spreadsheets with hundreds of rows and columns, tables are there to show us the numbers, the whole numbers and nothing but the numbers:

What I like to do to add meaning to the data, is benchmark it and show whether it is good or bad:

By adding colours with conditional formatting, suddenly you are showing what the trends are and making it easier to make decisions. This is much simpler for readers of the data rather than expecting them to remember which numbers are good and which are bad as well as picking out highlights or concerns from a standard column of data.

To help you improve your tables, Annie Cushing has written a guide on conditional formatting which can help you make your data tell a story quickly and easily.

Bar/column charts are not very commonly used, which is probably for good reason. A simple one shows a clear picture; however, the format does not lend itself for truly understanding the data, how it’s got there and where it’s going.

Here’s an example where the simple column chart tells the story well enough. It’s the algorithm weather report on MozCast, graphing how much the Google algorithm for organic search results has changed compared to the previous day, based on a scale they have created:

Now, they could have used a line graph for this, as each bar represents the next day, but using columns actually shows each day by itself clearer than it would in a line graph. In addition to the graph, MozCast also uses a picture chart to help tell the story – see the examples under Pictures, below, to see how they add meaning to the column chart.

This means the column chart is used as a backup to show more data on a scale. Simple, yet effective for spotting anomalies because your brain can quickly spot which bars are higher/lower than others.