9 reasons why your analytics fail in the newsroom

9 reasons why your analytics fail in the newsroom
6 March 2017 Laurens de Knijff

9 reasons why your analytics fail in the newsroom

Big screens with a fancy dashboard on the wall, they have become a standard item in a modern newsroom. Every major publication seems to have one. It looks cool and gives the editors a feeling of being informed. But in reality nobody is looking at those dashboards and they rarely drive editorial decisions. Why is that? It is because they show the wrong type of data. Journalists need different data than businesspeople. Here are 9 reasons where the traditional analytics approach fails for journalists:

1 Analytics speak
Using correct but technical analytics lingo in dashboards or reports doesn’t facilitate conversation about data. Journalists are not data analysts. They don’t care much for ‘unique cookies per specified main directory for tag2’. Just call it ‘visits’ or ‘views’ or some other name that sticks. (the real meaning of the metric should be explained somewhere else, so it can be looked up by those who want to know more.)

2 Because We Can
It is very tempting to show all data that is available. The more the better, right? Not true. Not all data is actionable, the more you show, the less it will mean to the user. Location-data is shown on most dashboards. A map with constantly moving dots on big cities looks fancy. But is it actionable? Or word clouds of trending search topics that your customers are using. What are you actually going to do with this kind of information? Is it even information? A good rule of thumb for editorial analytics: If you are not able to act on it, it is a bad metric.

3 Showing big numbers
Let’s say that your dashboard shows that your story had 98.345 visitors so far. That may look impressive. But what does that really mean? Is it a good or a bad number? Simply showing numbers or graphs, without any context is pretty useless. A number should always be accompanied by a reference. Some form of judgement that tells you if you should take action or not. Traditional reports show the previous reporting period as a reference (+0,5%) But that is not the kind of context that you need. (see problem nr. 7) Clear targets, goals or baselines is what you should aim for.

4 Another Social Monitoring Tool With a Fancy Name
So your marketing department bought in a special tool for monitoring ‘social’? Sure, it can give you some great insights on how people are talking about your brand as a whole. You’re driving the buzz. It probably has a separate dashboard too. But it won’t give you an omnichannel view per story. Besides: every channel has its own tool, insights, metrics and dashboards. Twitter analytics, Facebook insights, Youtube statistics, App data. Who has the time to monitor them simultaneously? And even if you do, you are not getting the full story. You are monitoring channels, not stories, nor the relation between them.

5 Website Tunnelvision
A custom Google Analytics dashboard with lots of numbers on an big screen can look impressive. But it only shows data from your website. Is twitter stirring? Is Facebook or Instagram going crazy? People are talking about your content on the channel of their choice. The party is at the busstop. You’ve got a big blind spot on your fancy dashboard if you don’t integrate social data somehow with your website data.

6 Reporting periods are arbitrary
If you try to monitor your content by daily or weekly reports, you will miss the full picture of a story. Stories have a certain biorhythm. They may live for several weeks or might pop-up because a related story is catching waves. This pattern is never in sync with standard reporting periods. So if you get a weekly report on monday, the sunday-posting will have a small chance of making it to the top 10. Even if it was very successful. If you rely on weekly reports (i.e on mondays) take in mind that recent stories (i.e. the sunday-postings) will never be on your radar.

7 Totals don’t relate to content, stories do
Traditional reporting is focussed on showing website-totals. Total unique visitors yesterday, total pageviews of news articles today. Totals are the fuel of the business intelligence department. Managers love totals. Usually they are the focus of your analytics solution. But totals aren’t very actionable for the newsroom. You cannot optimize on totals, only on individual items. And for that you need a view per story. A view that gives you the impact of individual stories, combined over all channels where that story is told.

8 The next day… is one day too late
Are you getting analytics reports in your inbox about yesterday or last week? You might need that for your long term strategy. But it won’t help you to optimize your current stories. If you want to keep track of your story you need data about it right now, not tomorrow. Why? Because you want to improve it, write an update, summarize the comments. Because you own the story and therefore the traction it realizes. Because you’re an editor in the fast and furious digital age!

9 The myth of accuracy
Because data traditionally supports long term business goals, it has to be precise and accurate. It has to be verified, double checked, aggregated, and stored in a secure place to become an official currency. But making data ‘official’ has its price. It is hard to access for the outside world and has a considerable delay. So when it has to be presented in real time, on a custom dashboard in the newsroom for instance, it can take considerable development effort to unlock it for other purposes.
And editors are not interested in this type of accuracy. All they need is trends: which story is growing, which one is dying? And they need it fast, rather than 100% accurate. Raw, good enough, data is often more effective in an editorial setting, than the company’s traditional enterprise systems. Free tools like Divolte or Piwick are perfect for getting real time data to a newsroom.