On The Trail

A recent ClickZ article by Shane Atchinson has caused a ton of good discussion on the Web Analytics news group In a nutshell, Shane Argued that traditional linear conversion funnels are fairly useless because few people browse a site using a common conversion path. He argues that a hub and spoke model is much more accurate, wherein the hub is a particular page/section of interest and the spokes are the paths away from it. Some spokes move towards a conversion (this is good), while some move away or are dead ends (bad).

He is right, of course. Linear conversion funnels are great for shopping carts where you have to complete a series of steps in a specific order, but that is really all they are good for. Looking for common multi page paths in pages that are not part of a defined process is a pretty big waste of time. The hub and spoke model, whether your hub is a single page or a group of pages, is usually a much more productive use of time because it reflects the reality of what is happening in most cases.

Actually, that brings up a good refinement of Shane’s argument. The “scope” of the hub and spoke can and should change. At the most macro level your hub and spoke model shows how people are getting to your site. The hub is your entire site, the spokes link out to the internet. Zooming in a bit you can see how people move from your entry pages towards conversion, or to dead ends (hubs are the entry pages, spokes are where people go from them), The final zoom in would be figuring out how people end up on your conversion page. In this case the flow of the model is reversed, the hub is the conversion page and the spokes are incoming links from other pages on your site. The metaphor we use internally is that of ripples on a pond. The conversion is where the rock hit the water, the concentric circles are the various levels of “zoom” as you move farther and farther out.

This is just a model and it doesn’t work for every site; there will always be exceptions. On some sites the traditional linear conversion funnel fits perfectly. But I would argue that on most sites a hub and spoke model of traffic analysis makes much more sense.

Well, I’m just back from Vegas. I’ve never been there before and I had a ton of fun. [[image:paris.jpg:Paris Casino Sign with Calculator:right:0]]

While I was there, just to keep my head in the game, I tried to figure out how they were measuring the success of their multiple marketing techniques. The advertising and marketing that goes on there is amazing and all pervasive. I suspect that everything from the shape of the casinos (very few right angles) to the scent of the air freshener (Mandalay Bay smelled like coconut butter) can be classified as marketing in Las Vegas. Most of it left me wondering if or how they measured the success.

The picture is a snapshot of the video billboard on the Paris casino sign. You can’t read it on this crappy camera phone picture, but the box in the upper left is a bit of the Windows Calculator. Oops! Maybe they were using it to tally up their profits!

I use WebTrends as my primary tool, so I’m very interested in their new version. I watched the WebTrends Marketing Lab Webcast today. I tried to live blog it. It’s really hard to type and listen at the same time! Plus, some of it would make more sense with pictures. All in all it was an OK presentation. As these things always are, it was too general and lacking in specifics for my taste, but they laid out a good vision of where Web Analytics is headed and what WebTrends is doing to go down that path. The Warehouse portion of the product looks great. I wonder how much it costs.

In any case, What I wrote is below, fisked in italics. All I can promise is that it will take you less time to read the description below than the 25 minutes it will take you to watch the webcast

.

Continue reading Webtrends Marketing Lab Webcast

There’s a new KPI making the rounds called the “Net Promoter Score”. The score is pretty simple. It’s based on a single survey question: “How likely are you to recommend us to a friend”, 10 = very likely, 1 = not likely. Subtract the percentage who answer 1-6 from the percentage who answer 9 or 10 and you have your score.

Fred Reichold came up with the metric and wrote a book on it (Ultimate Question: For Driving Good Profits and True Growth). The book just came out, so the buzz machine is starting to pick up on it. I haven’t read the book, but there are two good articles on it here and here.

According to Reichold the big deal about this metric is that a high Net Promoter score is a very good predictor of high revenue growth. It’s kind of a “Duh” conclusion if you think about it. Obviously companies that have satisfied customers will tend to have higher growth rates, and those who have dissatisfied customers who complain to their friends will tend to be less successful. The cool bit is that it provides a concrete measure to something squishy like “customer satisfaction” — and Reichold has data to show that differences in this measure actually matter.

The other cool bit is that a single question is pretty easy to ask on your website or in a follow up email. Before today, if someone asked me to measure “Customer Satisfaction” I’d break out in a cold sweat, but maybe it really isn’t so hard!

WebTrends just started the promotion for their new version, apparently it is slated for release in late April. They’re calling it “WebTrends Marketing Lab”. They’ve put a PDF up on their site about it here:
WebTrends Marketing Lab (PDF)
From the pamphlet there look to be a few new and noteworthy features, so it looks like a good upgrade:

1. Performance Dashboards You’ll be able to put your goals right into your dashboards. The graphics show some fancy meter style graphs that I assume allow you to see if you’re hitting your numbers or not. The graphics are neither here nor there for me, but I really like the idea of encouraging users to define their goals and to try to hit them. There is also a “what if” section on the dashboards that looks pretty cool. The graphic has sliders, so I assume you can see (for example) what would happen to your revenue if your visitors doubled. We use something like this now for our PPC campaigns in Atlas One Point and it is pretty cool. Of course the difference there is that you have complete control over PPC, while control over what people do on your website is, well, a bit less complete.
2. Business Intelligence Integration This is a huge one. They claim that you’ll be able to access WebTrends data via a standard ODBC connection. That will make WebTrends really easy to integrate with non-web data. In my book this is the biggest feature enhancement.
3. Interactive Reporting Console They claim this will allow segmentation of reports on the fly. In theory this is an excellent feature and one that their competitors have had for a long time. Whether it is worthwhile or not depends in exactly how it is implemented. I hope it isn’t lame.

They’re also releasing something called the WebTrends Marketing Warehouse as a separate product. There is scant info in the brochure, but it looks like it is meant to be used in real time, probably so you can dynamically modify content for individual visitors. It looks like this product has a desktop application interface instead of a web interface.

Excellent! I’m looking forward to seeing more of this new version.

There is an interesting article on Slate today regarding analytics and the difficulties of knowing exactly how many people visit a website. He compares numbers from comScore, Nielsen/NetRatings and Slate’s own IBM SurfAid analytics and finds that they are quite different. These differences ALWAYS crop up whenever you measure a site using different systems, and the panel based comScore and Nielsen numbers are rarely comparable with the onsite analytics numbers. The author does a good job of explaining why the differences occur (and why the onsite analytics is much more reliable), but it highlights one of the big frustrations with Web Analytics in general: it seems like there is very often an asterisk next to our numbers.

I was watching Aladdin with the kids this weekend and sometimes I feel like the genie in this scene:

ALADDIN: You’re gonna grant me any three [reports] I want?
GENIE: (As William F. Buckley) Ah, almost. There are a
few provisos, a couple of quid pro quos…

So what can be done? I think the best solution is to pick a good measurement package and stick with it. Question and try to verify the numbers occasionally, particularly if the metric in question is very unexpected. But don’t waste too much time on it — it will drive you insane and it will get you nowhere. The biggest value can usually be had by trusting your numbers spending your time acting on them.

I just finished reading Freakonomics : A Rogue Economist Explores the Hidden Side of Everything. I’m still not sure exactly why they called it that, but it was interesting from an analytics perspective, and it was quite fun to read. Stephen Levitt, the economist who co-wrote it, seems to like to find odd data sets that can tell him something unexpected about the world. He answers questions like “Why do most drug dealers live with their mothers?” (Because most earn less than minimum wage) and “How are school teachers like sumo wrestlers?” (They both cheat in similar ways when given similar incentives). He also has a long chapter on the economics of baby names that was very entertaining.
It was interesting from a web analytics perspective because he takes seemingly useless data and manages to draw very useful conclusions from it — either by combining several different useless datasets, or by mining existing data in a new way. Hey, I do that every day! Or at least I try to.
[On a different note, mouse over these Amazon links. I inserted some beta code from Amazon. They're running a test of a new widget that pops up a description and review of the linked book. Apparently half of you won't see it though because they're A/B testing the widget.]

Nature.com reviews a study that shows that people can make accurate assessments of a website in 50 milliseconds. This actually isn’t that big a surprise. Malcolm Gladwell has written a whole book on the subject called Blink. This just underscores the importance of your key entry pages. Have a high bounce rate? Have someone take a quick look at your page and maybe they can tell you why.

via Inside Analytics

I’ve been poking around in Google Analytics a bit more, and I must say that I’m still impressed. It does an awful lot for a free product. Google isn’t a charity, so why are they giving it away? Many theories have been floated around: to help people optimize Adwords, to help sell Adwords to your competitors, to help make the web a better place. All or none of those may be true, but I think the real reason is something different.

Google is, primarily, a search engine. They’ve branched out into blogs and groups and a billion other things, but the business is built on having the best search engine out there. Adwords would be worthless without the natural search results right next to them. If those natural results aren’t as good or better than results from MSN or Yahoo or New Competitor X then Google will quickly disappear.

So how does Google determine if their results are as good or better than their compitition? Before Google Analytics they could see what people were doing on Google.com but they could only infer that a successful search occured. How many people clicked on a natural search link in Google and then bailed and went to MSN? Did people find what they were looking for when they clicked on a link? There was no way for them to know.

Enter Google Analytics. Now if someone clicks through to a site with GA, Google can see if they spend some time there. They can see if they complete one of the goals that the site owner has defined as a success. They can also see this information for competing search engines. This information will be incredibly useful for improving the quality of their natural search results — which is really the most critical part of their business.

Now, this strategy will only work if a significant portion of the web uses GA. So, that’s why it’s free and fully featured.

Of course, if this is right it has a ton of implications. Will sites that use GA have an advantage in Google’s search rankings? Will optimized sites have an advantage in the rankings? Time will tell.

A nice succinct recipie for success from someone who has proven his ideas with several paydays:

“People come back to places that send them away. Memorize that one.”