Measure It

For a lot of companies, this is one of the most perplexing and challenging aspects of operating a web site. How often do things need to be updated? What parts of the web site should be changed? Is the site accomplishing its goals? Are people using the site in the way it was intended?

Finding answers to these questions is critical to a successful online project, and analysis of web metrics is one of the key tools to assist in finding answers. Peter Drucker‘s phrase sums it up nicely “What gets measured gets managed”.

My career in broadcasting introduced me to metrics early on. Radio and TV stations sell ads based on audience size. They use independent measures exclusively as a means of setting their rates, and they spend a significant amount of money to acquire and analyze those numbers.

My first job was at a tiny radio station in Dauphin, Manitoba. They couldn’t afford to buy into the radio ratings system but the station manager, the late Hugh Dunlop, had his own way of measuring. When reporting on a traffic accident (a big story in a small prairie town), he’d always end his report saying the cars involved in the crash had been towed to so-and-so’s service station. Then Hugh would drive over to see how many people showed up to gawk at the wrecks.

Hardly scientific, but better than no measurement at all. I call this “Dunloping” it.

Contests and write-in campaigns are other ways of “Dunloping” the numbers – the numbers are highly subjective but they’re better than nothing.

Fortunately, web sites generate a plethora of useful statistics. Yet it’s surprising how often we encounter companies that are making minimal use of their web measures, and are in fact more often “Dunloping” it than using more reliable information available to them. So, one of my first recommendations is that they embrace the mantra: If You Can’t Measure It – Don’t Do It.

In almost all cases, embracing this notion is easy because the data is usually being gathered. A little analysis, even simple analysis, will go a long way to helping you make informed choices, leading to more successful online projects.

What Do the Numbers Mean?

Sometimes a client will say to us “We have x-thousand page views a month. Is that good?” Metrics have to be considered in context. Because web technology provides us with a multitude of different measurement tools, answers to those key questions hardly ever come from one set of measures.

The importance of various measures is directly related to a site’s goals. And a combination of measures provides a more complete picture than relying on just one set of numbers.
For a community site, the amount of time viewers spend on the site will be a key measure, as will the number of user-generated posts. For a site that depends on banner ad sales, page views will be far more important.

The important thing is to start with a few key measures, analyzing them on a regular basis. Measurement analysis can get quite esoteric, and if you have the resources to do so, by all means delve deep. For the vast majority of small to medium businesses, a few key measures will provide important data to help determine the direction of the web site.

If you can’t afford analytic software, get the data, put it into Excel and make a few key charts and graphs. Do it monthly, and use the same format so you can compare results.

I like to group 2 or 3 measures together into what I refer to as ‘key indicators’. Take for example, an e-mail newsletter. Grouping list size, open rate, and click through rate together provides a significant snapshot of the campaign’s strength.

While growing the size of an email list is obviously desirable, a small list that has a high open rate (people actually read the newsletter after it has been sent) and a high click through rate (people who click on links in the newsletter) may actually serve the goals better than a larger list with a low click-through rate. In the example below, key indicators are graphed together to provide a quick view of the health of the list.

A set of key website indicators might include page views, bandwidth, and files served. This provides a snapshot of how much people are actually using the web site.

On the other hand, trying to determine which section of a web site to focus more attention on requires analysis of a different data set.

The goal here is to get away from decisions based on anecdotal or emotional data. You have a myriad choices to make as an online content creator, so use the numbers as a means of guiding those choices.

Trending is another important data point. Not all new projects show instant success, so looking at changes over time is important. Analysis helps you make decisions, assess your successes and fix mistakes.

Just as in TV, radio and print, measuring the success of your online efforts is key to your ongoing strategy. The web gives us an abundant set of detailed data, and the companies who are most successful online are aggressive in their use of that data. They’re also strict with themselves, building on the things that people actually view online, and ditching the things that aren’t generating any good usage stats. To be successful online you have to measure your data, and also be prepared to LEARN from that data. Be rigorous, measure what you’re doing, and you’ll be able to build a meaningful project for your online audience

In Summary

• If you have a web site, you have measurable results.
• Devise a way of reporting the data and look at it regularly.
• If your data isn’t pretty, make simple graphs and charts.
• You don’t need a lot of data, but you do need to use what you have.
• If you have to, Dunlop It.
• Ask ‘why is this happening’. Look for linkages in the data.
• Add some qualitative data when you can