On The Trail

Ciceron hosted it’s own version of Web Analytics Wednesday yesterday. It differs from the one Eric Peterson is organizing in that, well, it was on Thursday. More importantly, we also had it during the day in our offices instead of at night at a bar. This totally changed the tone from a networking party to a working meeting. (Of course, this is Ciceron, so we still had wine and beer by the pool table…). Andrew (Ciceron CEO) stressed that he didn’t want this to be a Ciceron thing, which is great. If it works out maybe it will rotate around to other offices after a while.
Several of our clients showed up and one person who isn’t a client. The discussion ranged from integrating tracking of offline promotions (big pain point) to ways to engage upper management in web metrics. One way to engage upper management is to bring them to meetings like this, so maybe next month we’ll see some new faces. Anyone is welcome at these things, if you’re interested drop me a line and I’ll get you the details for the next one.

Peter Merholtz of Adaptive Path comments on the (very important) link between analytics and design. Linking updated design to (hopefully) improved results is a pretty obvious win for everyone: the designers prove their worth and the business people understand what they got for their money.

One thing he didn’t mention is the idea of testing designs to determine the most effective. Testing is a key part of any redesign since the redesign may very well NOT improve your KPI’s despite the best intentions of the designer. The designer’s best guess is just that, a guess, until it is validated by some data.

After a rough start on Monday Google Analytics is finally starting to come on line with actual reports. We set up ciceron.com on Monday morning and now reports are available for Monday and Tuesday. At first glance the available reports are very impressive and the user interface is intuitive and clean.
Here are the things I find most impressive for a free tool:

1. The segmentation buttons. Many of the rows of the standard reports have the ability to segment on the fly. For example, you can see what keywords brought new visitors to your site as opposed to returning visitors
2. Integrated ROI reports. Only Google could do this, but it turns out you can only do it with Google campaigns/keywords.
3. Integrated A/B testing. It is relatively easy to do this in most other tools but they’ve provided some reports that specifically call out this optimization tactic.
4. Geo Map overlay. The program produces a nice looking map with a clear indication of where your traffic is originating. Most other tools just produce a list of cities/states/countries. A map is much more powerful.

All in all I’m very impressed. Over the next day or so I’ll be exploring the browser overlay, the funnel reports and the conversion tracking. I’ll put a full report here when I’m done.

As for the more strategic question of what does this mean to the web analytics marketplace I think this is a good perspective from Jim Sterne (quoted in a Clickz article):

“The Google Analytics announcement is a tsunami that will wipe out the lower-end tools and lift all higher-end tools after a short flood of delayed sales cycles”

Google has released a free tool that matches or beats the functionality of any mid-level analytics tool on the market. With Google’s money and marketing muscle it looks to me like this part of the market is done for anyone but Google.

The good news is that this will only serve to increase the need for people who can actually use this data to optimize their online marketing. Google has made the tool a commodity, but the analysis will be in more demand than ever.

A few months ago Google purchased Urchin — a very solid mid-level web analytics tool. This weekend Google released a re-packaged version of Urchin as Google Analytics . It’s a tag based system that is hosted at Google (so it requires no software installation, only the addition of a bit of HTML code to each web page) and, most importantly, it is free. Unfortunately the tool has been down all morning so I can’t tell you specifics about the available features, but by the descriptions I’ve seen it looks like a very robust offering. Eric Peterson (Jupiter Research’s web analytics analyst) has a good post on the features and some of the implications of this release. This is a big deal in web analytics and I’ll be posting more on it soon.

The cookie debate has died down somewhat, but the problem hasn’t gone away. I have 3 interesting datapoints from three different WebTrends installations:

1. Client #1 uses third party cookies set from “statse.webtrends.com.” They have 19.5% of “Visitors Not Accepting Cookies” (a WebTrends term that is pretty clear…). Now in this case these cookies would be third party cookies from a well known tracking domain, so this would be a worst case scenario. Client #1 is a large manufacturer of expensive products. Their audience is not particularly tech savvy or particularly corporate.
2. Client #2 uses third party cookies from “sdc.clientname.com” (in other words, a not well known tracking domain). They have 6% of “Visitors Not Accepting Cookies”. Their audience skews towards tech savvy and corporate.
3. Clients #3-#8 use first party cookies set via WebTrends’ tracking Javascript. The audiences are all across the map in terms of tech savviness and work/home mix. All have 1.1 +/- 0.2% “Visitors Not Accepting Cookies.”

All the percentages have been constant over the last 6 months.

Moral of the story: Use first party cookies.

Continue reading One more cookie data point

Here’s a interesting report from
DARTmail based on their aggregate statistics.

Summary:

Open rates down,
current average 30.2%
Clickrates down
slightly, current average 7.9%
Click to Purchase rates
from email range between 3.3% and 4.9% depending on the
season.

Full
report: 

DoubleClick Q1 2005 Email Trend Report