Well, my good friends at webtrends we’ll forgive me my implying they’re a bunch of old farts, with their 15 years in the market now. No, they’re of course not; I’m writing this post to react to their Digital Marketing Maturity Model, which they made public today (in beta, which is surprising for a model, since it’s not an application after all). BTW, you can get the PDF here.

Brace yourselves, dear readers (are there still any left? I sometimes get the feeling everybody is on Twitter now); you are going to hear a lot about this topic in the coming months. you see, we professional web analysts are just plain fed up with repeating the same Web Analytics Greatness Mantra over and over, and still see so many organizations not really doing anything with it. We figured that the problem had to be on their side, right? We finally came to the conclusion that analytics is like a glove; it fits or it don’t. And it comes in various sizes. Hence the recent works on evaluating how ready, analytical, mature, an organization is.

We are seeing various types of classification in Darwinism-like models (hey, nothing wrong with that, and we all know that only the fittest survive in business (or the too big to fail)): people like Thomas Davenport, Josh Manion, Bill Gassman, Stéphane Hamel, Moeller & Landry, Wayne Eckerson, to name a few recent works, have all proposed some ways to determine where in the evolutionary scale of analytical competitiveness a company can be placed.

Webtrends’s model is a similar attempt to identify the main components of maturity and what elements compose each one of those components. Here, of course, as with all other models mentioned above, we could spend hours debating specific items, how they can be concretely evaluated, why here and not there, etc. And I’m not so sure I like radar graphs that much; I certainly wouldn’t want to see them becoming a standard! Which is what webtrends hopes to accomplish with their DM3 model. They are also quite open to feedbacks from the community to improve it.

I just hope we will not see commercial interests around this maturity model thing, and witness debates motivated more by those interests than intellectual pursuit (some commercial stuff is OK; I’m not that naive!).

In the meantime, O Reader, start asking yourself questions if this is the first time you’ve heard about “analytical maturity”, or worse, “analytical competitiveness”. Organizational resistance still is in my mind the biggest obstacle to Web Analytics adoption, to its true operationalization.

You should have a hard look at your own organization; and the wiser you think you are, the harder you should examine your situation.

Give a try to webtrends model, or any other one for that matter.

Just do it this week.

This is the second instalment of this series about the analysis of www.cuisineduquebec.com since the day of its launched. If you missed the first one, well, look down on this page! or find it here.

For this post, I will examine in more details what’s going on with the organic search engine traffic, keyphrases, etc, since search engines, namely Google, are responsible for 85% + of total traffic.

In the period we examine, those visits came via 10,121 different keyphrases. Only 391 expressions have generated 5 visits and more, which is 3,9% of total. And what percentage of keyphrases brought 10 visits of more? A mere 1,4%. So, you got it, we are talking about a long tail. In this case 81,3% of all expressions generated 1 visit, a single one, in 3 months of activity!!

So, if we have a look, nothing surprises us with the line shape:

The 5 red crosses represent the top expressions where we find “cuisine du quebec” (see box). Note that we cannot speak of “brand” searches here since they are very generic expressions, except for the cuisineduquebec.com, which obviously is the site. This means that the very top keyphrases are generic topic terms too, by that I mean that those organic search visits were not from a navigational use of Google, which I have commented a lot here and here. This is not surprising, because the site is not known in its market, and that’s why organic search here is true search, and brings so many new visitors. It is important to differentiate that phenomenon from what Avinash Kaushik presents in his analysis of keyphrase Head-Tail attributes; I think Avinash negelcts the navigational use of Google, which I believe makes his interpretation of the Head questionable.

The dotted vertical line is where the 10–visit plus keyphrases end, which I define as being the Head of the graph. So, as you’re telling yourself I am sure, the graph presents no surprises, since it is quite the typical head-long tail representation.

However, it does not tell the whole story of where www.cuisineduquebec.com is at with organic search. The main expressions, the Head, again represent only 1,4% of total traffic from searche engines (the tail should normally be 20 times longer than what it appears here on this graph). If you look at the number scale, that Head still does not produce that many visits. Why? Because the site does not compete well yet for the high volume keyphrases. If it did, I believe that not only the number scale would be higher, but the relative contribution to the site traffic from the Head would be much more important than 1,4%. If we expected to be Pareto here, we would say that it basically should bring 80% of all visits!

This site is currently building a deep, long, and very diversified Long Tail. The variety of keyphrases from which it gets most of its traffic grows everyday. But it still gets a lot of competitive pressure indirectly, since it does not position itself yet in the high-volume keyphrases. This means that when it does, traffic will explode.

Question is if and when this will happen. In three months? Two years? This remains to be seen. I certainly wish I had a predictive model here, but external conditions are always hard, if not impossible, to factor in.

In the next post, I will examine entry pages, which are strongly related to what’s going on with keyphrases, as you might expect.

As a Web Analyst, one does not always have the chance to analyze and follow a Web site right from its first day online. I have had that chance in the last 3 months, following the evolution of my wife’s site www.cuisineduquebec.com (in French). You’re lucky too; my wife has given me permission to expose everything about her new baby, just so we can all have fun! Thanks Danielle!

NOTE: I use Google Analytics and Excel for all the analyses I’ll be talking about in the coming posts.

Cuisineduquebec.com is a content site about gastronomy in Quebec, but more precisely (and passionately!) about the people who make the great food products we enjoy here, the chefs who use them, the history of food, and of course recipes from the producers, chefs, and general public. It was launched on October 29^th 2008. As of Monday February 9^th, it had not been the object of any promotional activities with the media. That media blitz has started this week, and we will hopefully see the impact of that on the data in the coming weeks.

So, basically, Danielle has been relying on organic search to generate traffic, with the addition now and then of sites referring to hers. If we say organic search, guess what, we’re basically talking Google. In November, Google organic accounted for 65% of all visits, 89,5% in December, 87% in January, and 84% so far in February. That’s only one search engine (and its various properties)! In a word, we LOVE Google.

Here’s what the traffic (in visits) looks like since October 29^th up to yesterday February 10th:

We saw the traffic steadily grow up to Christmas (the first drop is me messing up a filter!!), then goes up again up to New Year’s Day; after that it kind of stabilizes, and restart to grow from the end of January. Of course, we don’t know yet for sure if the pre-Holiday traffic was due to seasonality, but even then, visit numbers are still low compared to the expected growth potential.

For the record, and as of February 11th, Google records 4,430 different URLs, although it identified up to 8,000 before Christmas, and that the latest sitemap.xml file we submited a month ago had 6,779 URLs in it. Keep these numbers in mind, because we will see interesting stuff, notably the number of different entry pages we got, a clear indication of efficient indexation the site benefits from.

In the next post, I will examine how the site is currently building its keyword long tail, but still fighting for the head keywords where competition dwells.

Am I too stubborn?

Recently I had lunch with a client and another consultant who’s there to help him choose a new Internet agency. Most probably, that agency will be the digital arm (or sub-contractor) of a traditional advertising one. The conversation focused for a while on my role, and what type of guidelines would I propose, so that early on in the RFP agencies would know where Web Analytics stood in the general picture.

The client in question is an agro-business lobby, the kind advertising agencies love, you know, those clients who spend millions on “branding”. I give them credit, however, for having developed in the past year a strong measurement culture when it comes to their Web investments. Not that it has percolated throughout the whole organization yet, but the Internet people have been asking very hard questions to their agencies about the benefits of all those online stunts. In one occasion, I was asked to analyze several months worth of data from a Web site that had been through two phases. You know, your typical full Flash concepts. The news was generally bad. Was not really worth all that money. The meeting with the ad agency was quite tensed; nobody shook my hand when I left.

Back to that conversation: we were wondering to what extent an agency should be subject to precise success measures. It was important to give them all the leeway they needed, (“Isn’t brand recall increase a good thing?”), etc. Wouldn’t being too strict with analytics neuter agencies creativity?

Would it? What is wrong with asking measurable results? Are we analysts killers of great concepts? Can’t a good idea be accountable to the business?

We did bring up the risk of the agency becoming too passive, too “go-fer”, forcing my client to basically play project manager all the time. I said at one point that I obviously respected the creative geniuses at ad agencies, that they were the best at what they did. However, I could not equate creativity with freedom to do anything.

Can’t creativity bloom with constraints? Isn’t architecture art + budget + gravity?

I believe Web analytics can, and should, establish the framework within which agencies can demonstrate their great creativity. It would actually be an even bigger demonstration of their abilities. A box, full of their magic.

I think many agencies still need to learn the analytics language, to evolve amidst new gravitational fields.

And there’s a right to be wrong we should not forget. An analytical culture is not a punishing culture, but a learning one. This means rewarding mistakes when they spawn great insights.

Ad agencies will need to understand the new grammar, and still write great poetry.