An unnamed company was pleased to be getting 90% of their web leads from search engines. Until Matt Bailey and his company SiteLogic dug deeper into the data and found that they were getting exactly 0% of their sales from search engines. All their sales during that particular time period were from a link that had been bought ages ago and cost $20 a year. This is why analytics – meaningful analytics – is important. You need to know where your money’s coming from.
Matt Bailey of SiteLogic Speaks on Advanced Analytics for SEMPO Atlanta
In his presentation on Advanced Analytics to SEMPO Atlanta last Friday, Matt listed three obstacles to advanced analytics:
- The Dashboard. Most analytics software packages have them, and most dashboards were designed by “the geeks in the basement.” Just because someone, somewhere decided to slap a metric into a dashboard doesn’t mean it should be an important metric to you.
- Velleity. Matt defined this as “the intent to do something, but not enough desire to actually do it.” (That’s a term that could apply to so many things, but I digress….)
- Hamster Wheel Analytics. The person in charge of analytics copies and pastes some data from a dashboard into a spreadsheet and forwards it to someone else. The someone else asks why a number changed since last month, so the hamster goes back to the wheel and spends the rest of the week running, trying to figure out what happened. It’s far better to spend your time and energy on moving the needle all the time.
“Caveman Analytics” are defined as reporting on the number of visitors, the time spent on site, the top 10 pages viewed, page views or even (God forbid!) “hits.” These are pointless metrics. Reporting needs to focus on one goal only – making more money.
Data – Information – Knowledge – Wisdom
The traditional knowledge pyramid comes into play here:
Knowledge Pyramid
So you know you had 34,000 pageviews last month. That’s Data – it doesn’t mean anything in a vacuum. Information is data+context, such as “the average page views per visitors was 8.” Knowledge is information+more context, for example, “visitors who entered the site on X page and stayed for more than Y amount of time converted at a rate of Z%.”
The top of the pyramid, Wisdom, requires an analyst. It’s often been said that companies should spend less money on analytics technology and more money on analysts, so they can use the data to move the needle and make more money.
Question-asking is the single greatest tool humans have. Asking questions is the way to wisdom. A good analyst is a rebel and a trouble-maker. He or she questions everything and is disruptive to the status quo.
The Star Trek Red Shirt Phenomenon
Aggregate data assumes everyone does the same thing on your website. Averages can be misleading. Matt used his Star Trek Red Shirt Phenomenon to demonstrate this in an entertaining way (that link goes to Matt’s blog post with more details).
In aggregate, Star Trek TV show crew members had a 13.7% mortality rate. Those crew members wearing red shirts (as opposed to blue or yellow), however, had a 73% death rate. Yipes! But the more you ask questions and dig into the data, the more interesting it gets.
The first question is, why do these red-shirted crew members die? Inter-planetary travel was a risk – 57.5% of the time a red shirt beamed down onto a planet with Captain Kirk, he or she was killed. Alien battles on the Starship Enterprise were also dangerous – during 42.5% of these, a red shirt died.
So let’s ask another question – is there anything we can do to prevent red-shirted crew members from meeting their fate? Again, the data holds the answer when properly segmented and analyzed. When Captain Kirk got involved with an alien woman in an episode, the red shirt mortality rate dropped to 12%. This probably indicates a “friendly alien” situation. So here, we can move the needle by ensuring that Captain Kirk meets a new conquest!
The moral of the story, obviously, is to segment your data and then segment it some more.
Segmenting & User Behavior
The 3 Cs of analytics are Context, Contrast and Comparison. Segment the data to understand the context, and therefore, the intent of the searcher (what are they looking for? What should you show them?). Then compare and contrast that segment with other segments.
People think that analytics is about numbers, but it’s really not. It’s about people and user behavior. A good analyst is a psychologist more than a numbers person. They find intent, determine expectancy, observe reactions, and analyze behavior.
SiteLogic’s client – a tourism site for the Black Hills/Badlands of South Dakota – had a high bounce rate and a miniscule conversion rate. The numbers told several stories, including the fact that a large percentage of visitors were looking for maps of the area. The site’s one map page was not particularly search engine-friendly, and it was buried in the site. So Matt’s company created a slew of (optimized) maps in a new, easily accessible section of the site. The conversion rate rose from 0.01% to 5.88%.
Another example: a SiteLogic client ranked #3 for the highly competitive search term “vacations” (plural). From this, they averaged 600 visitors and 0 conversions per month. They also ranked #34 for “vacation” (singular). From this, they averaged 1800 visitors and a 2% conversion rate. Clearly, the singular “vacation” was much more valuable, even in position #34! So they “unoptimized” the page that was ranking for “vacations” (plural) and let that search term go, targeting the singular version instead.
Sometimes you’ve got the right ranking but the wrong page. A high-ranking page may not be converting as well as a different page with a higher conversion rate. So reoptimizing here would be beneficial as well.
You can even measure the value of your inbound links with analytics. Segment by type of link (contextual, ad, blog, keyword-rich, social, news, etc.) and then look at time spent on site, page views per visitor and conversion rate. You may find that contextual links drive fewer but better visitors.
Usability
An advanced analyst understands usability. And sorry, designers, we’re not talking about the pretty details. Surveys show that layout, typography, font size and color schemes are the four factors that affect (a) readability, and (b) credibility.
Keeping these factors consistent (across your site and even across the web) will boost conversion rates. Every degree you get away from using blue underlined text links, for example, you’re asking your users to relearn what links are.
Matt recommended that analysts have three computer monitors on their desks – one with an analytics program open, one with a spreadsheet to work in, and one with the website onscreen (analysts need to look at the site!). With these tools, a good analyst can question, dig, segment, analyze, and use their wisdom to improve the metrics and make more money for the organization.
Follow Matt on Twitter at @sitelogic.
Tomorrow, we’ll cover Adam Proehl’s presentation on Advanced Social Media. Click here or on the “sempoatl1010″ tag below to see all the posts in this series (a new one will be added at 6 am EDT daily from October 4 – 8).
