group-blog-1

Blog

Inbound Marketing & Sales Development Inspiration

Web Analytics – Separating The Good Data From The Useless

Posted May 2, 2011
6 minute read

This is the beginning of a series of blog posts that focuses on going back to the basics of digital marketing. In this post, I want to highlight some of the web analytics do's and don'ts when it comes to metrics. There are dozens and dozens of different metrics available to us, enabling us to evaluate the performance of a website. Some of these metrics aren't nearly as valuable as we think they are while others that hold the keys to success are often ignored completely.

It is my hope in this blog post that I can begin to help shine some light on where we should focus when evaluating how good a website actually is and where we should stop paying so much attention. Beyond that, I hope to correct some common misnomers when it comes to web analytics data. Let's dive in…

Hits: 'Hits' are one of the most widely used (misused) terms when it comes to web metrics. In most instances, hits are used as a way to measure a site's popularity. Not only is it unwise to measure a site success based solely off total volume, but the term hits is not analogous to a visit*. Most web analytics tools don't even report hits as a metric. You're best off removing this term from your vocabulary.

*a "hit" occurs when a file is requested from a server - this includes web pages, images, self hosted video, etc

Total visits:
Everyone wants to know how many people are coming to their website. But at the end of the day, how much does this really tell you when it comes to generating meaningful value to your business. I could have 100 visitors come to my site one day and 10,000 visitors come to my site the next day. But if those 10,000 visitors are not part of my target market, how meaningful are they? It's a delicate balance between quantity and quality, and we can't get so obsessed with increasing the quantity of visitors while overlooking the quality of the visitor and the quality of their experience on our website. Getting more traffic to your website is easy. But getting more meaningful, targeted traffic to your website that accomplishes your websites goals is something completely different. Make sure you're shooting for the latter and then find ways to increase THAT volume.

Average time on site:
It's easy to make a big deal out of this metric, however it does fall under the umbrella of metrics we see over emphasized. It's interesting information, but it tends to be used in aggregate. In other words "my average time on site is X minutes and Y seconds, so I'm happy (or sad)." While measuring average time on site for all traffic is a mostly hopeless endeavor, you can quickly make use of this metric by segmenting it by different traffic sources. For example, how does average time on your website differ by visitors who come from a search engine or social media or through an affiliate program? If your search engine traffic spends 80% less time on site, start looking at other metrics (bounce, goal completion, etc) that support this source of traffic as being less effective.

Another simple way to make use of this metric is to establish meaning or an outcome associated with more time on site. For example, in the video on your home page the sales pitch occurs at the 1 min. and 5 second mark. Your twitter traffic time on page doesn't exceed 40 seconds. You now know your message is failing to reach that segment of visitors.

Pages per visit:
Pages per visit is like average time on site in that we think the higher it is the more engaged our visitors are. But like time on site, we need to dig a little bit deeper to understand the meaning behind this metric and what it means in terms of activity. For example, most websites have pages of all different purposes. Some pages encourage the visitors to take certain actions - which have a meaningful impact on your business. While some pages are a lot less interesting and probably do very little in terms of selling the visitor on your brand, service or product. Knowing how many visitors look at the key pages on your site is more telling than the average number of total pages any given visitor views when they come to your site.

Total page views:
Unless you're selling advertisement space and the key driver of ad revenue is impressions, don't get too obsessed over the total number of pages on your site. Like the metrics listed above, it falls far short of translating directly into value.

-----

Now that I've given you some tips on metrics that aren't quite as lovable as you first believed, I'd like to share a few that should capture our collective attention...

Goal value by traffic source:
Every website should have more than one objective. It's usually easy to identify the primary objective (make a purchase, submit an RFQ, etc), but the secondary objectives of your site carry meaning (and value!) as well. More on establishing website goals in this blog post. When we can identify the different interactions we want a site visitor to take and we can associate an economic value with that action, we can tell which campaigns (social media, SEO, PPC) are generating the most value.

Loyalty:
If you run a nonprofit and the goal of your website is to get membership sign-ups, how do you measure success? Number of visitors that sign up as a member? Take it a step further. You may get 2000 sign-ups in the month, but how many are coming back to your site? What good is there in just signing up if they're not taking any action beyond that? So we look at the number of people who are signed up and then look for how many of those members continue to come back to the site and take part in our message/campaign/program/whatever.

If you're an e-commerce site trying to get up off the ground, are you most interested in the one-time shoppers or the visitors that become customers and brand advocates for life? Of course, you want the latter. And one of the best ways to know if you're building loyalty is to see how many of these visitors continue to come back after they've made their first purchase.

Campaign ROI:
This one seems obvious, I know. But when you're investing in traffic building campaigns such as pay per click, we are very capable of being able to import our click charge data and place it right alongside the revenue generated by those visitors. At a glance, we are able to see which keywords in our campaign generate a positive return on investment and which ones do not. There is no clearer roadmap for where to focus your efforts in a campaign than this metric.

Bounce rate:
While bounce rate is not an end all, be all metric, it is perhaps one of the standalone top-level metrics we can use to determine if something is working or not. Bounce rate of 80%? Either you're attracting the wrong audience or you're showing them the wrong stuff on your site. Bounce rate of 25%? Your visitors like what they see (at least, initially). The one exception to this is if a visitor can accomplish everything you want them to on the first page. For example, if you write a blog, a visitor will land on your site, read a post, then leave the site. Bounce? Technically. Failure? No way. While this would technically count as a bounce because they only viewed one page, they accomplished the goal of your blog - the consumption of your content!

-----

It is important to note that most web analytics metrics are meaningless when they're not put in the context of another variable. One of the greatest sins of web analysis is only looking at the dashboard and trying to draw a meaningful conclusion; that is, looking at your data in aggregate, one metric at a time. This is why I encourage you to get your eyes away from total visits and average time on site for all visitors and instead, measure outcomes separated by traffic segment. This gets us away from looking at our site traffic as one giant blob of people and instead separates them into nicely organized groups; separated by a characteristic and/or their ability to accomplish an action. This is where the real magic begins to happen and where we start finding actionable data and insight.

Topics Web Analytics

Agree, disagree, or just have something to add?

Leave a comment below.