Mymotech

Google Analytics recently updated their product to allow users to trend data across a week or a month instead of just daily. This is a very nice improvement in their trend viewing capabilities.

Which one of these tells the story best at a glance?

A.

B.

Gotta give credit, this is a nice small enhancement that adds a lot to the instant analysis capabilities of Google Analytics.

I just got back from Portland, Oregon, where I had the good fortune to have lunch with Eric Peterson, dine with WebTrends (twice), and spend a whole day on the cutting edge of web analytics with the WebTrends Customer Advisory Board.

First, lunch with Eric. It is pretty cool that a guy who barely knows who I am would be willing to take time at the drop of the hat to have lunch. But that is just what kind of guy Eric is. He picked me up at my hotel in a black porsche (I didn’t get the model). It was indicative of the fast paced time we had. I had a blast; talking about web analytics as an industry, talking through engagement as a metric and the challenges around standardization of it, and general web analytics industry observations. Like me, Eric has a passion for web analytics and the web analytics industry, and that made lunch go by way too fast. Since my wife is having our next baby a month before the Xchange conference, Eric let me off the hook for that, So there’s one spot that somebody could grab if they hurry.

The WebTrends Customer Advisory Board is something I am very honored to have been a part of for the last three years. Barry Parshall, the Director of Product Management, and his product management team work so hard to provide a platform for us to evaluate and think with them about WebTrends Products and Strategy from every angle. The setting could not have been better. We had consecutive days of sunshine, and a great venue at WebTrends headquarters in downtown Portland. Here are some of my takeaways.

• WebTrends Dynamic Search. This product offering has officially clicked for me, and it completely validates the acquisition that WebTrends made a couple years ago. One anecdote that helps to explain the potential value of a product like this is that in the course of using WebTrends Dynamic Search over a year’s time the algorithm that tunes and optimizes bids made six million bid adjustments for one client, to achieve the best possible search marketing ROAS (Return on Advertising Spend).
• Bruce Coleman, CEO of WebTrends is a genuine and warm individual who completely gets it. Not that my saying something nice about him adds anything to his huge list of professional accomplishments.
• Saw for myself the excitement and energy that my friends that work at WebTrends have been telling me about since the executive change last fall. This was evidenced by a stronger sense of ownership, transparency, and urgency across the board. I was completely caught up in their enthusiasm. The resurgence within WebTrends is perhaps best evidenced by the recent re-joining of several key executives, that had left WebTrends and are now back and re-energized. That doesn’t happen at a company on the ropes.
• I got a tour of Tech Support. It was cool putting some faces to the names I have talked to so many times throughout my years working with WebTrends.
• I finally met Aaron Gray, blogger and thought leader on web analytics and Web 2.0 measurement. That was really cool.
• Got to meet Mike Keyes. Another blogger and web analytics mastermind that I have looked up to for a long time.
• Got to hang out with the inimitable Chris Grant. Nuff said.
• WebTrends definitely acknowledges that they are the number two in the market right now, but is not content to stay there.
• Despite the turmoil of the executive team turnover WebTrends continues to be profitable and growing.
• As an organization they are turning up the focus on the WebTrends Analytics product. Expect lots of improvements.
• The executive team has thought clearly about the needs they want to address in the marketplace, is able to enunciate it clearly, and has aligned their product strategy accordingly.
• They are acting on customer demands to communicate more openly about product news, development, and best practices.

My overall impression after meeting with the WebTrends Customer Advisory Board and the WebTrends leadership team is one of potential about to be realized.

All in all, it was a great trip to Portland. I would recommend it to anyone.

One of the nice things about having my own web site is that I can express my own opinions. Here is one of them.

I have been annoyed for a long time with Omniture’s sales methods. Omniture is having their extremely well attended Customer conference this week so, I figure this is as good a time as any to get a little critical of the world’s largest web analytics vendor.

First things first, kudos to them for doing an exceptional job leveraging their position in the market to become the largest web analytics vendor. They have some eye-popping stats.

• 66% of the most innovative companies use Omniture
• 40% of top 100 retailers use Omniture
• Omniture process 8.2 billion transactions daily (events and page views?)

Now let’s go back in time a couple of years. I was sitting in a room with a group of Omniture people, and they were doing a presentation for a client of mine. As the presentation and demo kept going I was shocked that they kept promoting common features in their tool as the only vendor to provide that when it is common knowledge that the exact same feature is available in almost every other web analytics tool on the market. I felt like I was taking crazy pills. We ended up not going with Omniture in that particular instance, but obviously many, many other companies bought similar logic hook, line, and sinker.

It looks like this trend has only continued based on the recent press release for Omniture Fusion. Their “unique” approach is just a retread of what every web analytics consultancy worth its salt has been doing for many years. Sure, it continues to be a challenge. And sure, I am glad Omniture has decided to try this, but why act like you are the first people to slice bread. In the meantime, it looks as though doing a worthwhile implementation of their tool has been left to surrogates.

Why does Omniture do this? Well, obviously they are a vendor with a need to sell as much of their products and services as they can. And if they can get you to believe that they have introduced a true differentiation between their product and their competitors they will have leverage in the sales process. It is a strategy that has worked wonderfully well up to this point. The truth is that Omniture is selling its sameness as differentiation all over the place to resounding success. But how long can fooling some of the people some of the time continue to be a successful strategy in this industry that is getting smarter by the day?

Some people are going to read this and think that I dislike Omniture, and that is not true. Omniture deserves its place as the leading web analytics vendor, but their sales approach over the years annoys the crap out of me. Truthfully, if I had to do a vendor selection right now, they would be on the short list, but not because they have provided a truly differentiated offering.

These are some of my thoughts related to Avinash’s recent post about how to make HiPPOs hungry for more web site analysis and optimization.

I think this goes deeper than employing this or that tactic or strategy to advance web analytics. It is fundamentally how to get business decisions to be made in your favor and for the good of the company. The keys to getting buy-in from management in any area are proving value and overcoming fear. In the mean time, managers are saying, “I bought you Omniture, and now you want me to buy Offermatica, Comscore, and/or ForeSee?”

The Truth That We Need to Face
For web analytics specifically, I think we need to be honest about where we really are. For most companies there is a pretty big gap between intent and reality. The intent is to spend more on analytics while the reality is that 76% of retailers don’t do any kind of testing. That tells you:
A. Now is a great time to be a testing vendor.
B. The majority of companies are possibly still only doing squat-grunt-report for web analytics. And if they are doing something meaningful, it is probably compressed into one acquisition channel like search marketing (not that there is anything wrong with that).
Anyway, this post isn’t meant to depress you so let’s move on.

Proving Value is Comparatively Easy
At what point do you stop spending when you have a 3:1 ROI? That is a trick question. In most cases, you don’t. The internet is full of great ROI whitepapers for whatever web analytics tactic you want to employ. In fact there are even a few handy dandy ROI calculators out there. This is all Business Case 101. Here is my favorite template.

Overcoming Fear is Hard
Getting past people’s fears is a lot more difficult than proving value. Fear in a lot of instances is irrational. You can’t overcome irrational fear with a good plan. What you will get instead is a very strange and irrational negative response to your very well thought out idea. The only way to overcome fear is with trust.

Fear Factors
These are areas where I have noticed a lot of fear in the area of web analytics. You could also substitute FUD (Fear, Uncertainty, and Doubt):

• Fear of the data. You cannot do more and better web analysis when there is a fundamental distrust of the underlying data in your tool. As web analysts we do a lot of damage to our own cause by introducing a lot of uncertainty/fear in the data.
• Fear of failure. Many organizations punish failure, but as iterative testers/optimizers of the corporate web site we are supposed to fail in small ways to gain great truths about our site. The fear of failure, especially if this is something the organization has not attempted, or worse yet never done successfully will make any good plan go nowhere.
• Fear of the unknown. There is a fear of the new and the unknown. Eventually leading edge ideas become mainstream, and there is a pretty well established technology adoption curve. Some day your company is going to get their act together and do some of this cool stuff. Remember when Google Analytics came out? That was the crest of the broad adoption wave in web analytics, just so you know.

Don’t Give Up
Think hard about which fear is in action when you get a negative reaction after proving the business value. Alter your approach, and start attacking the fear by building trust. Sometimes you have to approach obliquely. Sometimes the thing they don’t trust is you. So find out who they do trust and use them as a vehicle to get buy-in. This has nothing to do with web analytics, and everything to do with building trust. Most importantly don’t give up. Being part of the solution in improving web site analysis and optimization is preferable to giving up and embracing your own fears.

Thanks for reading this long post. I am really curious about your thoughts on this.

Surprise! Surprise! I recently had a chance to do some analysis of reporting tasks when done with Google Analytics vs. WebTrends. The idea here was to see if I could save some money using Google Analytics for web analytics reporting. Make sure to read at the bottom for caveats galore.

The Challenge
Compare reporting effort between Google Analytics and WebTrends for 15 web site profiles that utilize custom segmentation, event tracking, and goal analysis reporting. Being very experienced with WebTrends, I have a pretty good idea of the amount of time it will take to get the data I need from WebTrends. I then did similar tasks in Google Analytics to create a comparison. Keep in mind that this comparison only relates to the time to get the needed data out of each tool. Implementation and Analysis times are not considered. They are roughly the same anyway in this scenario.

The Results
Google Analytics can do most of the required reporting (The only data point I am unable to collect from my requirements is unique visitors by country). However, due to limitations in the way that data is accessed and exported using Google Analytics it would take more than three times the hours to collect the reports needed to perform analysis.

So using my ultra-complex mathematical algorithm: 5 hours per week with WebTrends now equals 15 hours per week with Google Analytics. Hmmm, over the course of a year that is 500 hours of my time. If I use Eric Peterson’s average web analyst salary data [pdf] that is $22,000 that Google Analytics is going to cost in salary. Ouch.

What causes the Gap

1. ODBC. ODBC. ODBC. WebTrends having an ODBC connector that I can use to directly query the report database is a monster of a time saver.
2. The persistent export of interval data before the main report in Google Analytics. This is sometimes useful, but not always what I want.
3. Inability to create a custom report in Google Analytics. Especially around custom segmentation. I need to see multiple values in the reports 1st dimension and GA will only show me one at a time. It is enough to make a grown man cry. And before some kind soul recommends creating separate profiles for all of these just know that before doing that I am at 15 profiles already in this particular situation.

My conclusions

• Google Analytics is a robust web analytics solution. But we already knew that.
• While my reporting requirements were not all that intense, there were some aspects of it that really made life hard for me when using Google Analytics. I know there are probably ways around these issues that ingenious GAAC’s have thought up, but you begin to see diminishing returns at some point. I want to stay away from bailing wire and magic solutions that keep me from being replaceable.
• When you evaluate web analytics tools, it is important to carefully consider your needs. Just because Google Analytics is free doesn’t mean it is the lowest cost.

Caveatville:

1. I use Excel for web analytics analysis. I think that I am in pretty good company until I can afford some sweet software.
2. None of this is meant to imply that people should use WebTrends over Google Analytics. Needs differ.
3. This analysis was done by someone who is very experienced with both tools, so results may vary with familiarity, however it is probably fair to say that I am more skilled with WebTrends.
4. This is not meant as a comparison of the features and functionality of each tool. The requirements did not utilize all features available in either tool.
5. Why are you wasting away in caveatville? You should be making an awesome comment or subscribing to my feed.

Update: I made a mistake initially and wrote 2nd dimension only shows one value where I should have said 1st dimenstion only shows one value. Sorry for the confusion.

I love web analytics. It is one of my life’s highlights to have gotten involved with it when it was still fairly young, all the way back in 2004. For all of you that are longer in the tooth than I am, before scoffing just think about what was available online back then. Basically a couple of vendor sponsored white papers [PDF], ClickZ articles, A few blogs, and the Yahoo Group. Today, I have 45 web analytics blogs [OPML] in my Google Reader (and I am fairly selective). I recently wrote about some bloggers that need to post more. As evidence of my “pull” in the industry none of them have actually taken me up on it, but a couple have indicated a return soon. Yesterday I stumbled on another jewel of a web analytics blogger: Paul Strupp.

My endorsement is not exactly the Colbert Bump, but here’s why I like his blog:

1. Very intelligent and very experienced.
2. I can relate to his style. He doesn’t take a know it all attitude in his writing.
3. He is a real practitioner solving for real problems a web analyst faces.

Here are some links to good content to get you started. After that you should subscribe to his feed.

• White Lies
• Cold Fusion Moments
• Two Types of People

Starcom, Tacoda, and Comscore jointly published a study yesterday revealing that 6% of the online population accounts for 50% of all display advertising clicks. This audience fits the demographic of 25-44 and household income less than $40,000 annually. Additional details regarding this segment are that their online buying is not proportionate to time spent online, and they frequent auction, gambling, and career services sites. This is interesting data. I like to think. And here are some of my thoughts.

• Click-Through rate shouldn’t be used by anybody as a success metric, even for brand-building campaigns
• Is this the signal that the “unwashed masses” have come online? The heritage of the internet is smart, but as the internet becomes more available to everyone that means the internet will reflect society more and more. Go get your driver’s license renewed to see what I mean. A lot of times we are too NY/LA/ivorytower to really get a grip on this.
• I love Tacoda. I mean really love them. Behavioral targeting is hella fun.
• We have to keep pressing to understand people online. Web measurement is only part of the puzzle. So often, I see people attacking an online objective with only quantitative data. This is perturbing especially in the area of brand measurement where visitor behavior tells you very little about how your customer actually feels about you.

Every one’s favorite analyst, Dylan Lewis, started a Web Analytics Wiki back in June 2007. When it launched I thought it was a great idea. A community around web analytics that was a living breathing Wiki. Well then I got busy again and forgot about it. A couple of weeks ago I remembered the web analytics wiki, so I thought I would check it out and see how the community was coming. I jumped to the “recent changes” page and to my delight there were updates from that day.

I jumped to one of the pages to see what the community had developed, and lo and behold this is what met my eye.

Needless to say I was horrified. I made a couple of edits to try to get rid of some spam pages, but if this wiki is ever going to get going again it will take the community to get in there and fix it. So what do you say, let’s all get our iShovels out and get rid of a few pages of spam and help get this wiki back on track again.

Two of my favorite web analytics stars are at it again. It wasn’t difficult to see before that the divergence in their thinking centered around Eric Peterson stating that Web Analytics is hard, and Avinash Kaushik stating that web analytics is complex, and now it is out in the open.

But what is truly at issue here? Is this really the throw down between complexity and hardness?

When I got involved in web analytics back in 2004, I was surprised at how easy it was get good reports out of the tools. But I was also amazed at all the little details I needed to know to make sure my reports were as accurate as possible. Things like whether or not my PDF downloads were throwing off multiple 206 return codes and giving me way too many download counts, or trying to figure out why 80% of my visitors came from Reston, VA. and other stuff like that. Because of this I started saying, “Web analytics isn’t rocket science it is just a lot of information.” In other words, Web analytics is complex.

As I progressed in this field I started to realize that producing the world’s most accurate reports was one thing, getting someone to do something about it was quite another. The actionability crisis was on. And this is where things got difficult. To make an organization that was unaccustomed to paying attention to web metrics to make decisions on this new data set was a monumental task that required not only a strong understanding of their business but also enlisting the participation of executives who could effect change within their organization and bring about a metrics revolution. This is where web analytics is hard. Really really hard. That is why many organizations today still struggle with what most web analytics experts think is commonplace; picking a KPI and then doing something about it. It is hard to get organizations in the habit of doing that. Especially organizations that don’t reward risk taking. In this way, web analytics is hard.

When the Giants beat the Patriots in the SuperBowl, they used the same combination. It was hard to beat the patriots, and to do it they utilized complex schemes of attacking the QB with different guys coming free at Brady. And on the offensive side of the ball they used complex patterns and routes for their receivers and tight ends. Beating the Patriots was hard and to do it they had to implement a complex strategy.

So, at the end of the day web analytics is hard and complex. As excellent web analysts we must be comfortable with both the complexity and the hardship of web analytics.

Eric Peterson and Avinash Kaushik are both right.

You know you have arrived when you get a reciprocal linking request. This arrived in my inbox today:

Hello,

We visited your site [name] and are interested to swap links
with your site.

We would add your link at the home page of
[URL removed] which will actually
help to increase the search engine rank of your site and give you some
targeted traffic too.

We’d appreciate a link back from the home/internal page of your site for
the mutual benefit.

If you are interested please reply to this email with your link details.

We look forward for your positive response.

Best regards,

Well, it must still work, I guess. There is only one problem, the URL they sent me where my link would go was to a domain that expired five days ago.

Woops. I mean seriously, get it together. You are going to ask for a reciprocal link, but then tack on a domain that is not there. Maybe the barrier to entry is too low online.