A Totally Worthless Metric?
Written September 12th, 2007 by Neal LinkonI recently went to the web site for a company that manages one of the college loans we’ve taken out for my son. I spent 20 minutes trying to get the answer to what I thought was a simple question, only to give up and send them an email (don’t get me started on how long it took to get a worthless answer – I wound up using the phone to get the info I needed). But after my unsuccessful visit, I’m sure the webmaster at that company bragged that the average length of my visit went up.
Meanwhile, we recently streamlined some common search forms on our site at Aurora Health Care. The month we did that, our average length of visit went down. I’m not 100% sure it was because of that streamlining, but the average length of time on those search pages went down by a lot, so I’m thinking it had an impact.
So why does anybody brag about the average time on their site?
If you could assume that every visit was a customer or potential customer soaking up the brand and all the wonderful information you have to offer, then average time on site would be extremely valuable. But in the case of my 20 minutes on that college loan site, it was filled with frustration and not-so-muffled obscenities directed at the company.
I’d like to believe the streamlined search feature on our site made some patient’s visit to a little nicer and easier, even if it meant less time on the site.
You can even take out the two extremes from the data – the guy who got a phone call and forgot that he was on your site for the entire evening and the many bots and others that spend 1 second on your site – and I still question the value or meaning of the metric.
I’d argue that most people using the web these days aren’t surfing; they are seeking specific information or attempting to complete a certain transaction. That means the less time somebody is on your site the better.
I know in my case, the site offered a number of interesting features, graphics, etc. I may have actually learned something or found out how to save money if I had ran down any of those rabbit holes. But I wanted the answer to a question. I’m guessing most of our visitors are the same way.
The exception may be those individuals researching a health issue. Perhaps, the average time on the site for people who start or wind up on the health information content is more meaningful. But even then, if all I want is to learn more about that procedure Mom is having on Friday, do I really need 10 minutes on the site to find and read it? I sure hope not.
September 12th, 2007 at 8:29 am
Nice post, Neal. I absolutely agree.
The biggest problem with LoV, as you stated, is that there’s very little you can actually infer from it. A longer average could either mean they’re interested or frustrated. A short time could either mean they couldn’t find what they wanted or they found it quickly. We’re then left with an almost meaningful statistic: medium length averages that we would expect.. but perhaps we see them as meaningful only because we’re looking for that particular range.
I would be more interested in how the LoV changes over the year. But, again, there’s very little we can infer without directly quizzing or, at the very least, observing the visitor.
What I would like to see, if there is such a thing out there, is something like Google Analytics’ “Site Overlay” whereby the statistics are laid out over your website. Then if we could change the statistic to the average length of time between coming to the page and clicking on a link. We could perhaps infer better whether the visitors are able to find the link they want quickly or whether they’re reading the content and clicking on links along the way. (For instance, a shorter length at the top and longer length at the bottom would infer that to be the case.)
Anyone know if there’s a tool like that out there?
September 13th, 2007 at 4:25 pm
Neal, I agree with you, too.
All too many of the ‘foofy’ web marketers would be quick to equate longer length of visit with ‘stickiness.’
I’m pessimistic… I’d be wondering if they were having trouble finding what they were looking for or experiencing slow page loads.
How can I tell the difference?
I guess I’d have to be able to see their navigation path through the site at the same time… how many pages did they browse? What kind of content were they browsing? What was their connection speed? How many times did they hit the back button? and so on…
What would I do with that information?
1.) I could track conversion rates a little better… how many times did I “get the click” like I hoped to?
2.) Are there any apparent navigation problems?
3.) Did someone just come from a competitors site?
4.) Did someone immediately leave to go to a competitor’s site?
Supposedly, CoreMetrics allows you to do this. It drops cookies on the user’s computer so that the analytics can keep track of the visitor’s path. We don’t use CoreMetrics but are considering it. Anybody else use CoreMetrics for any of this?
http://www.coremetrics.com/solutions/web_analytics.php