Don’t you just love an argument? We often throw paper cups at each other and argue about the finer points of online business. What am I talking about? Well this post by Matt Belkin of Omniture suggests that “visit” should be used over the “Unique visitor” as the one metric to give to your executives.
However this post by Howard Kaplan from Future Now critiques this theory in no uncertain terms.
Ahar har…. a scrap! says I.
So as not to sit up in the crows nest pruning my wings I decided to look at both opinions and give you my opinion, cos that’s one of the rules of this ship, to give an opinion.
So without further ado, lets look at Matts reasoning first, since he posted first.
- Visits are more accurate than Unique Visitors.
- Every Visit represents an opportunity to persuade or convert a visitor to a customer.
- Measuring visits is based on fairly established industry standards
And now lets look at a summary of Howards arguments
- Measured Visits is no more ¢¬Åaccurate¢¬ at measuring reach than is Unique Visitors. In fact, they¢¬¢re far less so. How can visits possibly tell us about Reach, in anything other a relative sense? Reach is about people, not sessions.
- Every session (visit) does not represent an opportunity to persuade or convert a visitor to a customer. What every session does represent is an opportunity to persuade a visitor to decide to take an action.
- To cut a long story short, Howard replied; So what? What value is the industry standard? you should measure based on internal benchmarks not what everyone else is doing.
So who is right?
The answer is “it depends”. 😉
From the e-retail perspective, Howard, without a doubt. From the lead generation perspective I would again fight Howard’s corner with my cutlass.
The goal for each session here is to help the user take an action rather than make a purchase. Expecting everyone to buy or sign up in a single session is not practical as I know Matt would agree.
I think Matt’s point here is valid if he used average visits per visitor as guage of how long on average a visitor takes to become a customer but he didn’t argue that.
Measuring reach. My gut is to again side with Howard because his reasoning is good. It is not about sessions, it’s about how many people you er… reach. Are we redefining what reach is in a marketing sense to cater for the web analytics vendors? Or are we trying our best to gain as accurate as possible the numbers of people visiting our sites?
No-one is arguing that UV numbers are 100% accurate. But I have to point you to things like TV panel measurement (widely adopted – often with just 2000 people on the panel) to say that UV numbers are far more accurate in terms of trending than most other media.
Measuring brand effectiveness. This is where I kind of side with both guys. Note that reach and branding are not the same thing. Branding is how well the visitor associates your brand with your product or service. Reach is simply how many people you reach.
Ok you could argue (and I’m sure Howard will) that we’re still talking about branding in the mind of the visitor. However isn’t the number of sessions each visitor makes then a good measure of how effective that brand is at recalling the visitor? Visits on its own is still too simplistic, but certainly visits per website visitor and repeat visits it could be argued as effective KPI’s for branding campaigns.
Media websites selling advertising. One of their main selling points is the number of sessions, because in each session there is the opportunity to serve adverts. UV numbers are still important, but I side with Matt here. A good example is how the IAB in Finland had to react to the fact that one media site got 6 million unique visitors in a month (due to some of the cookie problems Matt discusses). There are only 5 million people in Finland, so the advertisers lost confidence in the numbers. Number of sessions is a fairer method to sell advertising than UV in this case.
This brings me to the next point about industry standards. Both Howard and Matt are right. Firstly Howard is spot on there are no external benchmarks which are valid enough to make business decisions upon. There can’t be. How can BT for instance compare itself to a telecomms benchmark that includes rivals like AOL, TeliaSonera and Japan Telecom? Answer? it can’t. Too many variables effect each business model to conclude that an average benchmark is relevant.
Secondly Matt is right in that when you talk about accuracy in terms of absolute numbers then sessions are more accurate. This is useful from the advertisers perspective when determining where to spend her money. For instance in the Nordics we have Resoner.nu for the advertising community.
My conclusion is, that in retail, lead generation or any website where your final macro action is to convert the visitor to a customer then I think Howards reasoning is correct. Media websites, websites selling advertising or websites measuring brand campaigns on the other hand might do better to look at visit based metrics, rather than UV based ones.
In both cases I would say that KPI’s should be developed using both Uniques and session based metrics to determine success. Simply picking one over the other is not enough.
So there you go landlubbers I’ve drawn my cross in the sand. Let me know your thoughts on all this. Post a comment.