First of all, thank you to Avinash and the rest of the Google Analytics team for getting me into the new version of G.A. while I was at Emetrics. Avinash, in particular seemed interested in my comments on the new visualizations, so without further ado, here is some feedback for the team.

I’d really love to start off by talking about the addition of sparklines, but I just can’t because I am too distracted by the big trend chart.

Here’s what that looks like:

Let’s take a quick moment to compare that to a Google Finance chart:

Notice that they are both ‘web 2.0-ey’ and ‘ajaxey’. However, the GA chart seems to have gone a little over the top with the very thick trend line and large diameter data point markers.

It’s worth noting that the Google Finance chart offers multiple ways to traverse (interact) with this data. I can both traverse and zoom (or unzoom) the chart at the bottom as well as zoom the chart from the top-left text links. In the GA chart my only option for interacting with the data is through the pull down menus at the top right (one for date(s) and one for dimension which are kind of kludgey if you ask me - they don’t respond -hhmmm- smoothly and they rearrange the page whereas the controls in the GF charts don’t require the page to be reordered and react both quickly and smoothly.

None of these are my biggest criticism, the simple fact is that the GA chart with the heavy line and large radius markers makes me feel like it’s a kid’s chart, that is a chart designed for a kid.

I find it very odd that the GA Dashboard chart is liquid - that is it stretches to fit the browser width. This changes the horizontal scale of the chart and can either obscure important trend details or create artifacts in the visualization - a fixed width chart that is appropriately scaled would be better.

Maybe it’s just me but I actually feel patronized by the Google Analytics chart like I’m so clueless that I wouldn’t notice the trends or the data points if they weren’t pessimistically obvious. I *think* that I actually like the roll-over data labels because it puts the detailed data into the chart as opposed to how the GF chart updates the information way up in the right hand corner is you hover over the map. But, again the heavy black border just screams ‘MAKE SURE YOU LOOK AT ME AND IF YOU DON’T THEN YOU DON’T KNOW WHAT’S GOING ON HERE!’.

It’s not that I’m totally in love with the GF charts, as noted above I’d really like to see the detail data down next to the chart in a data label and while I theoretically like the idea of marking key news events on the chart, the implementation can get quite messy and distracting. Unfortunately I don’t have a good solution in mind yet but I’d really like to see it cleaned up.

For the GA chart, I’d recommend:

1. Don’t talk down to the user, move towards a more elegant presentation like that of Google Finance
2. Build in interactive controls like the ones used in GF
3. Make the chart fixed-width, in this case liquidity is bad
4. Potentially allow for multiple dimensions to be charted (within reason)

Ok, that’s enough about the dashboard chart, sparklines will have to wait for another day.

By the way, overall I like the redesign of GA - for me it is generally more intuitive to navigate and use.

I’d love to hear what you think…

Avinash posted today on “Leveraging Statistical Control Limits” in web analytics as a way to cut out the noise in your KPIs and therefore being able to focus on the true outliers - data points that are truly worthy and in need of your valuable brain power.

In his post, Avinash challenged his readers to provide an Excel solution since there are no OOTB (Out Of The Box) solutions for introducing constants, statistical or otherwise, into charts - Excel tries to box us into using just vertical and horizontal axes scale controls.

Well, I can’t pass up a challenge like that right?

What do you think of this?

So what’s going on here?

1. Named Ranges - Named ranges are your friend. In this case, I’ve got named ranges for each important series of data, especially my key metric - page views per visit. That way, I can insert a new row of data and my constants (Standard Deviation, Mean, Upper Control Limit, and Lower Control Limit) will update automatically. BTW, the chart uses the named ranges as well so that it will update automatically every time a new row of data is added.
2. The chart is a combination of lines and stacked bars. The stacked bars create the gray biorhythm (normal variation) area. Line charts are used for everything else.
3. For the top 5 outliers (either upper or lower) I’m dynamically adding a data label so that I know, at a glance, the specific values for them.
4. I’ve eliminated all horizontal lines other than my own constants because they are just distracting

You can download the spreadsheet and play with it if you want.

A couple of caveats:

• Named ranges won’t do the dishes for you. In other words, when you insert a new row, you’ll need to copy down the formulae from the previous row
• Excel doesn’t always do a good job of placing data labels so the outlier labels may need to be massaged after an update
• My spreadsheet only uses 1 standard deviation. While 3 might be a six-sigma best practice, three standard deviations would have put the lower control at 0.1 and common sense tells me that it shouldn’t be possible to have an average page views per visit value lower than 1.

I’ve deliberately not gone into great detail on the inner-mechanics of the spreadsheet (not that their profound or anything) because that would make for an impossibly long post - if you have a question please post it in the comments or email me directly (civy at instantcognition.com).

It’s been 43 days since my last post.

I’ve been busy getting/starting a new job.

New jobs are great, but unlike some, a new job and its associated learning curve drive every little thought from my head.

I’m still in information overload but I’m trying to dry-shift the brain in blogging-gear.

I have been paying attention though. Jeremiah, who has been traversing the far east, has left Hitachi for PodTech.net. Jeremiah, through no fault of his own, was one of the bloggers I studied before I launched mine back in March ’06.

There’s been a bit of a kerfuffle stirred up by privacy advocates over behavioral targeting. There is a discussion about it going on over at the Web Analytics Yahoo! Group.

Scoble asked for some new KPIs around engagement, which just drives me nuts because engagement is not a new topic and not limited to “Web 2.0” – we’ve all (probably) been working on engagement a long time and there are KPIs available – I guess either they’re not special enough or “Web 2.0ers” don’t pay attention to web analytics *sigh*.

On a related note, Eric Peterson has kicked off his Web 2.0 Measurement Working Group to start ferreting out if there are any special measures for new media and what technical requirements need to be met in order to support them.

Scoble and Jeremiah talk about “Content Malls”, maybe I’m old school but isn’t that a suburban way of saying ‘Web Ring’?

Finally, the guys over at Juice Analytics report that marketing guru Seth Godin challenges visualization master Edward Tufte on Minard’s chart. That’s just silly.

I just watched Seth’s Preso on Google Video and he definitely doesn’t think that the Minard chart is a good chart – for the ‘lay’ person. I still disagree, the fundamental points of the chart are easily grasped, but since he pays Tufte a back-handed compliment I guess I’ll let it slide. (By the way, it’s an entertaining presentation, no surprise there, but the Tufte bit is in the last section at around minute 17 or so.

Hopefully I’ll have something useful to write about soon – any thoughts or questions that I might address? Let me know.

Ok, so it’s more like what I’ve been keeping my eye on and this is a desperate attempt to make sure that I actually write something this week, but still here are the things that have held my interest.

• Manoj Jasra over at Web Analytics World had four guest posters this week, Avinash Kaushik, Jennifer Laycock, Gord Hotchkiss and Rand Fishkin. They all cover different topics and each one is worth the read.
• Robbin Steif has an excellent post on what web analysts do all day at her Lunametrics blog, so if your bosses, peers, family, or friends have a difficult time understanding what you do - point them to Robbin.
• Both Stephen Few and the guys over at Juice Analytics pointed out this new tool for creating sparklines and bullet charts (among other things) in Excel, so I’ve been playing with it. By the way, it’s called Microcharts by BonaVista Systems
• Of course, the Hewlett Packard BoD fiasco has been absolutely fascinating (and horrifying) - Scoble is keeping tabs on it: Day 1, Day 2, Day 3, Day 4
• Mom’s back to blogging, this time from Egypt where they’re on a 2 year teaching stint - check out the new masthead for the blog which is the view from her classroom!

Or, boy I am I oblivious.

Eric Peterson lobbed this my way a week ago and I just noticed it a few days ago (I’ve been hard at work on a response).

His basic question, if I may restate it, is “What do I think about using a ‘tachometer‘ to display KPI data?”

The short answer is, I love to do this - when it’s practical.

If you work in Excel, it is often not practical because the level of initial work and maintenence needed to operate a custom chart element like this violates the efficiency rule (remember, that in order for reports to be truly effective, they need to be easily repeatable so that they can be delivered efficiently and on time). In other words, I’m lazy and it is, generally speaking, too much work for not enough return.

That being said, I’ve spent the past couple of days approximating the dashboard that Eric refers to (via Dashboard Spy and courtesy of Subaru) in Excel, to see if it could be done - with the proper wherewithal.

Here’s a screen grab of the finished product:

Ok, so it can be done.

Here’s a list of resources you’ll need to pull it off:

• Andy Pope’s Gauge Work Around
• John Peltier’s Speedometer Work Around
• Apps Pro X/Y Chart Labeler
• Excel Camera Tool

Scared yet?

Ok, here’s my file for the above picture, you’re welcome to try it out (no commercial use, please), but I make no warranty as to the accuracy or stability of it.

Long story short, custom dashboard elements like the above are cool, sexy and can effectively communicate important information such as KPIs, but hopefully you have a better platform than Excel to create them in (this is where something like Xcelsius becomes helpful).