Electronic media can collect everything off you these days. Where you go on a companies website, how often you call, how long you call for, what you speak about, how often you open their emails, when you clicked a link, where you came from before you landed on the website, did you click on an advert, the list goes on and on and on. We track so much information and seldom does anyone ask why.

This article aims to discuss what all of the data means, good ways of grouping and comparing data, and how you can ensure you don't get data overload and end up staring at a screen of numbers that means very little to you, or anyone else.

Data collection

First, let's discuss data collection. It's important to note that just like any other system, the principle of Garbage In, Garbage Out applies. If you feed your reporting with poor quality or, even worse, the wrong data, you are sure to end up with meaningless or confusing results.

To collect the right data we need to know the outcomes needed for the different stakeholders of the organisation. For instance, to help customer service, we may need to be recording the number of complaints per product/service and number of resolved versus open issues. This can create a good report but here is the key question – does it provide you with an actionable insight?  If you cannot take any action from the data gathered, there is no point gathering the data. In this case we could be monitoring the complaints per product and if one product has a lot more complaints, we can alter it, take it out of our range or other such action to help reduce complaints.

With website statistics, a similar story occurs. Make sure you are tracking the things you need to. The contact form, how people download your presentation, how many people checkout on your website, how many subscribe to your newsletter. It can all be tracked and you need to make sure what's being tracked is right for everyone in your organisation.

Organising the data

We're now collecting the right data, however we need to ensure that someone is looking at it as well! Giving the right people the right data with which to make decisions and organise actions is a crucial part of your overall data strategy.

First of all, we should already know who our stakeholders are from the data collection interviews in part one. Each of your stakeholders now needs to have the data they needed presented to them appropriately. The best way to demonstrate this is by example. Here we'll take website statistics for an ecommerce site. Let's also assume we have three stakeholders: the acquisition team, the conversion team and the boss.

The acquisition team is only interested in how many people are visiting the website and by how much that is increasing (or not). So we make them a report that shows:

o The number of unique visitors
o How many of those are new visitors (e.g. new acquisitions)
o Which sources these visits came from (email, PPC, referring sites and so on)

Now normally that's probably all they need to see. If we're smart though, we know that not all visitors are of the same quality and the acquisition team needs to know that! So we should also add some outputs of the those visitors, so we could add:

o Bounce rate
o Average time on the site

Now just an example of one of the other stakeholders, the boss. They are all about the money in this case. His other departments take care of the rest. So that report is simple, so for them:

o Visitors
o Revenue
o Average order value
o Which sources these visits came from

That's all! Only organise the data your stakeholders want to see! Otherwise they won't look or won't understand.

Understanding the data

So, after all that, we actually need to understand and create actions off the data. We need to understand what, if anything data is telling us. Using our website example, we can show some ways to understand data.

Say the acquisition team is looking at their report and comparing the different sources of visitors. They may see that one particular traffic source, let's call it websitestrikesback.com has a very high bounce rate (bounce rate means the people that landed on your website and so immediately left, indicating for whatever reason, it wasn't for them). This clearly means that any money being spent acquiring these visitors is not as effective as other sources.

That is all data can tell you. The 'Why' has to be investigated. In this case, a visit to the offending website and to the page where the link to your site should be done. It's possible that this website is:

o Sending visitors to the wrong page of your site
o Has mislabelled your site as something it isn't
o Is the wrong site for your product, e.g. could be too low budget when you're a luxury goods provider.

Which could all cause higher bounce rates. This is where human investigation comes into play and also following that, testing.

Once you think you know the problem the only way to be sure is to test a solution versus your current situation, and see which does better, and then make a change or test a new version instead.

Conclusion

This article has dived from quite a high overview into some simple examples at a detailed level. Hopefully it has given you some insight into a four-step process to making sense of data:

1. Collect the right data for the right people
2. Organise that data for them so they only see what they need
3. Look at, understand and investigate that data
4. Theorise, test and change based on your investigations

This needs several different types of people to help set this up – one person alone does not normally have the detailed level skills as well as business level skills as well.

Final thought – do not make reports for reports sake. Look at all of your reports you look at and read. When was the last time you changed something as a result? If you haven't done so for ages then you might be able to bin it as you're not doing anything with it anyway.