What Are My Statistics Telling Me?
Written May 16th, 2007 by KinstlingerRemember when the only things us “webmasters” could really track were monthly hits and visits? When they went up, everybody was happy, except maybe my then-boss who asked me how many of that increased traffic was due to more people “going online”. Our industry has progressed and we have an enviable tool, those of us who maintain and oversee websites – access to an avalanche of statistics. With little trouble, we can learn how many people visit our site, view certain pages, and how they got there. Three mouse clicks show us which of our forms are getting the most responses, how our site visitors are using our internal search, and which of our PPC terms are getting the most traffic.
Of course, the dilemma is created: What do all of these statistics mean and how can I use them?
A few years ago, author Michael Lewis wrote a baseball book called ‘Moneyball’ which detailed how batting average, HR and RBI were limited statistics, and that we should follow things called ‘OBP’ and ‘OPS’ instead. Like those tried and true sports numbers, it’s warm and fuzzy feeling when overall website hits and visits counts rise, but they can’t tell us the true effectiveness of our site(s). Do those increases represent people searching around the site for something they can’t find, or are they new visitors who found my site due to SEO and marketing efforts?
April’s AdWords activity increased. But, is that because people trust my hospital’s reputation and are following up with appointments, or are they just using our website for reference?
I’ve realized that I’ll never be able to effectively use ALL of the numbers at my disposal, but I also know that a statistic is only as meaningful as the thing it’s supposed to represent.
Once you know what you’re after, statistics are great for a few reasons:
- I can learn if they’re doing what I want them to do. Are they taking the bait – finding and filling out our spring health fair RSVP? My cardiology coordinator wants to get the word out about a new doctor and procedure – Here’s what I did; here’s what worked and what didn’t. Let’s review.
- I can learn what they want to do. In October, my monthly report showed me that 10,000 people from Texas viewed our sleep center page. Thirty seconds later, I knew how and why. Looks like people aren’t downloading Dr. Schon’s podcast; guess I better position it better. Is our targeted Ravens page really getting THAT many hits? As their official medical center, I’m not surprised.
- Oh, yeah… I can see if this month’s hits and visits were greater than last month’s. I guess some things are harder to outgrow.
May 16th, 2007 at 7:39 am
For me, there is only one statistic that matters – the number of new patients we get because of the web site. That can’t be tracked by your log files – it has to come from your call center, or tracked by the individual departments. It takes a lot of manual work to get this data, but a positive ROI is the most powerful metric you can present to management.
Back to web stats – we had a similar anomaly in the traffic to our Sleep Center. The single word “child”, was the number one search bringing visitors to this section. It turned out that Google Images listed a stock photo from a page on “Sleep Disorders in Children” as the #1 result for this search. None of those visitors cared about our Sleep Center, they just wanted to grab a picture.
Lesson learned – we now filter out traffic from Google Images from our reports.
May 16th, 2007 at 8:20 am
Interesting. Our sleep center bump was b/c the Dallas TV station linked to our sleep quiz and soon after, the story was picked up by other stations. We soon had people from all over the country taking it. It was never intended as a stand-alone item. We had to change the ‘thank you’ msg to make it more proactive.
May 16th, 2007 at 1:42 pm
There is little question that too many people use web stats as just numbers rather than analytics. When we make any substantive change on the site, we have theories regarding which metrics will change and why. We then monitor those to see if we were right. We also have more complicated amalgams of data that reflect things like conversion rate, physician and facility searches, etc. Much more useful than just visits and visitors (but I still love those numbers, too!).
May 16th, 2007 at 3:49 pm
Here I am , agreeing with Ed
Numbers are great, but let’s get back to our mission – do the numbers indicate they did what we most wanted them to do? Choose to see one of our doctors? Actually use our services?
We have to sort through the volume of possible stats and decide what ones have real value – and consider that the stats that have the most value might not come from your Web Trends pages at all.
Having said that, the lure of having volume figures is almost irresistable. But when you look at minutes spent on your site, that bubble can be quickly burst. How many of you have 0-1 minute stays as your largest volume?
May 17th, 2007 at 6:42 am
With bots, people using the site as a doorway to other portals, etc., short visits are indeed quite common. Which begs another question: Many people assume that the longer the average visit the better. But if the visit length is because people are struggling to find what they are looking for, that’s not a good thing.
May 17th, 2007 at 9:50 am
Are there any statistics we wish we had ?
I love that I can now track website conversions via Google PPC.. just another great offering from Mountain View.