Analytics Notes

A couple of weeks ago I had an engagement in New York, and was staying at a big hotel near the client’s. Since I like to keep it simple when on the road for work, I had all my meals (except for lunches) at the hotel restaurant. While eating, I would get asked by the staff if everything was to my liking. I mean everybody. Anyone with a hotel badge would stop by. That was about 5 times each meal. You got it: it ended up annoying me a lot. Especially since I would not tell them if I really did not like the food, which happened to be the case a few times. Maybe it’s me, but it’s got to be very bad for me to tell a waiter to bring back the plate. It’s actually rarely happened in my life.

This made me think about the Web, and how much easier it is to give our honest opinion. This is why Web site managers should get into attitudinal analysis if they haven’t already done so. On the Web, you’ll get a good idea of what people think of their experience with you online. Do it, but…

That experience in New York also got me to think about maybe how it is actually too easy to ask for people’s opinion online. With attitudinal analysis getting now into the web analytics main stream, I wonder if we do not run the risk of sollicitating our visitor’s feedback too much. Just imagine if 3 out of all 5 sites you visit started with asking you if you would please answer their survey since they care so much about your opinion. Would be pretty unnerving, wouldn’t it?

I guess it’s a matter of balance: when should we ask about visitors’ opinion, and when we should wait until they want to give it to us? I don’t have an answer to that, but I suspect that it will be soon a very relevant question.

So, how’s everything? Everything is “fine”, thanks.

Last week I had the pleasure to have a demo of XiTi, the leader in behavioral analysis applications in France and in good parts of Europe. I had heard of it for several years, but had never had the real opportunity to see it first hand (and frankly I was a little more interested in what we find here in North America). Fehmi Fennia, the head of Xiti’s still small office in Montreal, was kind enough to proceed to show me what that application has in its guts.

And it’s got a lot!

First, XiTi has a free version, and I understand that this is how it got so well known in Europe first. What I saw, though, was the full-blown XiTi Analyzer. I was first impressed by how much slice’n dice that can be done on the fly with the data vizualisation (tables, charts, etc.). It’s got a powerful page grouping structure that allows you to configure 8 levels deep! It’s possible to take out specific days in a custom date range, say the last 10 days except last Monday and Thursday. One can even drill down to an individual visit! They are the only ones, I believe, who will create separate reports for brand-related keyphrases in the search reports. And they have a heatmap module together with a data overlay one.

In short, I was impressed with its flexibility.

I think that the product is particularly interesting with its segmentation capacities. One can create segments on the fly, with the reports that will then all reflect that specific segmentations. I have never had the pleasure to see Visual Sciences work during its prime, but this is definitely the most flexible application I have seen (except maybe for WebTrends Visitor Intelligence, but that product is not really in the same category, since it sits on a relational database. And yes, there’s Omniture Discover).

A XiTi implementation, in terms of tagging, seems to be pretty much like an Omniture one; you need to do a lot of planning beforehand. For now, XiTi will offer free implementation services to its new North American clients.

I know, nobody knows where the web analytics application market will be in a year or two, with all those acquisitions, and free (ahem, subsidized) stuff around. But if you believe that your organization should take the investment way, which I think it should if you do anything substantial on the Web, you should probably put XiTi on the list of products to consider.

P.S. Fehmi, don’t hesitate to post a comment if I made a mistake in my summary.

I was recently on an interesting case, analyzing the behavioral data from one of my client’s Web site.

Here goes:

This company operates in a geographically delimited market, with a very unique product. We know from research that the client enjoys a 96% brand awareness. Needles to say that their clients, and their prospects too, know who they are.

Analyzing the search traffic, I saw that organic search actually drives a good third of the online sales! Paid accounts for close to nothing, and has been falling (both in terms of percentage of sales and conversion rate) for some time now. As for Search Engine shares, Google owns that space with MSN and Yahoo barely moving dust.

When I look at the keyphrases, the top 20 grab 83% of all used expressions, with mention of the brand in all but one! I mean, it makes no sense that any Internet user would in 2008 not try the company name with .com in the address bar to get to the site (the .com is actually one of the top keyphrases). Still, many people used the company name in their search.

The more I look at it, the less I can call that search, though. It looks much more like browsing to me, as if Google was replacing the URL field of the browser.

Is Google the new browser?

Actually, I am noticing that phenomenon more and more with companies that have strong brands, and whose domain name is really obvious, i.e. people could know right away from trying their name with .com. Still, the brand (and variations/associations with other keywords) often occupies almost all the territory of organic search terms. Quite interesting.

I began asking around, and quickly people started telling me about that new habit they’ve had for some time. It is so much simpler to key the name of the company in Google (and for a few others MSN or Yahoo); you know you’ll get the site as the first, or very first, choice at the top of the SERP. No more typo in the URL, etc.

And you? Is this how you sometimes use Google?

If so, this brings the important question of how we should analyze organic search, especially in relation with PPC. It also pose the question of SEO relevancy, and strategy. Scoring well on one’s brand(s) and name(s) should not then be seen as a success. Sure, there are the odd cases of companies that don’t even own the best positions for their brands/names. You can imagine how much trouble they’re in then when people are just looking for them, and not finding them.

This also questions the relevancy of buying one’s names and brands in PPC. I am no SEM expert, but I am wondering if Paid Search in that context is not simply bringing traffic that would otherwise have come to the site via the free organic results.

Anyway, I am not drawing any conclusions today. I will need to investigate this much further.

And you? Have you noticed this phenomenon in the site(s) you analyze?

Yet another post about my fantastic adventures with Feedburner. After writing that things were back to normal, my friend started to play tricks again, almost immediately in fact, leading me to think that I was getting punished for claiming victory to fast (I was raised a Catholic after all).

Since my previous, naive post, here’s how the situation has evolved:

You can tell what ride I’ve had; swinging from sentiments of being loved one minute, and rejected the other. My doctor actually told me that it might have caused permanent damage to what was somehow a healthy 45–year old heart.

Can I sue Feedburner/Google for “inconsistent reporting” reasons? Sleepless nights? Existential crisis? I guess not, since no one can touch “free”.

So, what are my real numbers? Well, I guess it depends on my mood, if not accuracy. Shall I calculate an average, or just plain decide to care only when the numbers are in a high phase?

I know more than one manager whose attitude towards numbers is just that!

Did we, the people who left comments on my post about Feedburner and I, have anything to do with this? Was someone at Feedburner listening? Google? Maybe even…? The thing is, it’s back!

As June Dershewitz said: “As bloggers I think we all ought to track new subscribers and unsubscribers as success measures, but we can’t because the numbers are so sketchy.” (my emphasis). Well, in the absence of a better solution, I guess we’ll have to live with suspicious numbers. I certainly look forward to seeing the Google ownership effect in Feedburner.

An integration in Google Analytics v3?

OK, I’m throwing the towel.

Dear readers (how many of you are really out there??), is anyone among you able to explain to me how the Feedburner stats work? Following daily the numbers of my blogs, I have come to understand that the displayed number of subscribers was based on how many readers connected each day. Well, is it? I have noticed that, on weekends (and holidays), the number could go down a lot, making me think that many subscribers were connecting from work. Made sense. So, I quickly realized that it was not because people were actually unsubscribing (although I am not naive enough to believe nobody has!). The number of subscribers would still slowly go up month after month.

But now, I don’t get it. The number went down by half some time ago, to sharply go up again, and back down by half (114 on time of writing). It’s been there for days now, making me depressed. Or should I be?

Here’s the graph for the last 30 days:

I know something is fishy with Feedburner. One of my blogs, the too neglected The Big Integration, is supposed to have between 0 and 2 subscribers. I know it’s way more, because I personally know about a dozen (unless they all lied when they confirmed receiving my feeds), and each time I talk about WebTrends, about 25 people visit from the company on the same day. How would they know if they hadn’t subscribed?

So, please, help me make sense of this nonsense!!

What am I missing (besides brains)?

I do not really like to repeat rumors. In fact, I am categorizing what I am going to say as a rumor, because I can not reveal my source. All I can say is that source is very reliable.

First, you can not put 20+ brilliant people in a room, with no restrictions normally attached to being a P&L, ask them to code, and expect them to stop once they got to a nice little application you can give for free. At one point, ambition starts to kick in, with its sister, pride.

It was just a matter of time before Google Analytics got bored with being the little widget of Web Analytics. The secret? Well, not a big one to anyone who has been following the field for some time: Yep, GA is going “Enterprise Level”, for whatever it means. I guess they resented the name calling last year. Expect a full fledge product, with all the functionalities, with support, and yes, sales efforts (even though it will be free). People are actually calling and visiting big companies to convince them to use GA!

I don’t have anything against Google Analytics. Well, I do a little. Like basing my trust in the numbers on the sole reputation of the brand. No access to the data. To a professional Web analyst, that’s a major drawback. No, I totally agree it’s a neat product; I use it myself all the time. I use it, because I can’t afford the Big Ones, which I would by far prefer. But I certainly am grateful to have access to a product the quality of GA.

I am simply wondering about how the application scene will evolve if GA goes after the market, the real one, the paid one with its big bucks. If GA finally behaves like a “company” that is in the Web Analytics market. Why would anyone not choose a subsidized product over an expensive one? And what’s in it for Google? And what if, with the power of such brand, money, and “free” (the nicest word in the English language), GA kills off all the competition, and become the only available product?

Of course, this is science fiction. Won’t happen. I just think it’s funny we can now start talking about that great company the way we have loved to talk about Microsoft for so many years.

Needless to say, if what I am saying has any merit, the coming 18 months are going to be really interesting.

I’m not int the habit of using this blog for personal stuff. I know you read it to get my take on Web Analytics, and I want to respect that as much as possible. However, following my announce that I was leaving Facebook, my friend David Sadigh published this video on his blog, which I thought was just plain hilarious. I don’t want to start a movement, no, Facebook is fine. I just want those who can’t stand it not to be ashamed, and do their coming out!

Enjoy!

Aren’t you bored with all those blogs (now over 70), including this one, and those PDFs, and those books telling you how to do Web Analytics, and how full of insight it is?

No? Yes?

Well, I am.

When was the last time you got some true amazing insight from analyzing your Web data? Better yet, when was the last time you did act upon that new information? This brings the hard problem of adoption. And that is why Web Analytics is hard. Yes, I know, nobody’s job is easy; heck, that is why it is worth giving you your salary! I mean hard in the sense that once all that nice information and insights are around, it is still hard to move our big marketer behind and do something about it.

But whose fault is it?

As an industry, the Web Analytics one that is, I think we are at a crossroad. I tend to agree that we have somehow under-delivered on several promises. Or rather, we seem to have not kept the promises we made. True, many WA projects out there finally laid very small eggs. But I would suggest that the main culprit is still a company’s culture.

It has amazed me, throughout the many analysis projects I have been involved with, how hard it is to make people act on one’s discoveries. It is as if Web Analytics was there only to talk about what happened, and not about what you should make happen. I have put my finger on problems that were costing tons of money, and still, companies would take months before even trying to find a solution.

It truly escapes me…

In 2008, we have to stop wondering how well Google Analytics compares with the paid solutions. Please, I beg you, stop asking me that question! If you don’t care to pay, then don’t, and implement whatever tool you want. Of course choosing a good tool is important, but if to you price is the only criterium, go ahead.

But start squeezing insights of what you implemented right now!! You should get something to act upon at least once every quarter. If you’re not, you either ask poor questions (most probably), or you have reached the limit of your tools. However, I bet my money on the former; it takes time before becoming analytically smart to the point of rendering a tool useless.

Don’t get me wrong: I know there are many success stories out there. But what is YOURS?

When was the last time your Web Analytics investments helped you discover a $500,000 opportunity?

I just canceled my membership on Facebook. After several months of getting all kinds of invitation, and trying to care about what all those people were doing, I got bored. I should say, I have better things to do, and I need a life!

So, to my ex-Friends on Facebook, don’t look for me; I’m gone. Hey, you can stay there, it’s perfectly fine. It’s just not for me; maybe something to do with age…