branding me 2.0

The Multi-Lingual World

In globalizing societies, it appears increasingly advantageous to fluently speak/read/write Mandarin, Spanish, Arabic, and so on.

The other day I realized I have a modification to this: in globalizing societies, especially those diving head first into information technology, it will be almost as important to be able to read/write in programming languages as well. In other words, the “other” languages of globalized capital will soon include PHP, JavaScript, Objective-C, etc.

How To Make Google Out Of Date

I saw the TED demonstration of Microsoft Labs’ Pivot the other day. Gary Flake describes it as the point between searching and browsing information. In another video, he says to think of it as business intelligence meets a search engine.

This demonstration came at a great time for me. A side project I’ve been working on (my non-analytics/web/business reading) led me to realize the limitations of Google, or, more precisely, the idea of a search engine that’s based on keywords.

I don’t think we’ll ever outlive searching for keywords. With so much based on language, there are not many ways around that. However, it is limiting when you are looking for more than just some keywords or a quote or some lyrics to that song you just heard.

For instance, for my project I would like to have digital/searchable copies of all the works of a philosopher that I am currently reading; then, I want to compare all the text in those books to a database that contains the names of other philosophers. I would like to do this so I can trace this philosopher’s thought, and the people that he wrote about.

There are numerous problems in trying to accomplish this. Here’s just a few:

1. We are extremely behind in digitizing content and making that content available online…this is a whole other post in and of itself.
2. HTML typically does not provide semantic markup; that is something like XML’s domain. For example, on Wikipedia I find that Jacques Derrida’s name is surrounded by this: <h1 id=”firstHeading” class=”firstHeading”>Jacques Derrida</h1>. The H1 tag is not helpful, since it is only “structural” in nature; I need something like <philosopher>Jacques Derrida</philosopher>. Sure, I could write a program that looks for the name in <h1 id=”firstHeading” class=”firstHeading”></h1>, but I don’t know Python (as well as most Americans).
3. Finally, it would be a lot of information to search through. The ability to instantly visualize it (e.g., the philosopher cites person X only once, whereas they cite person Y 115 times) is important.

As far as I know, Google et. al. do not allow this. I have not had time to do my research on what other people have written about this topic, but after thinking about it I realized that the search box we all know and love on the homepage of Google may be our generation’s 8 track, especially when it comes to accessing the so-called deep web.

The future of search is so much more exciting than a text box and ten blue links.

Thoughts On Google Analytics Intelligence

Intelligence is a new feature for Google Analytics. It sounds like it sprung from a challenge by Avinash Kaushik to the GA team. Eric Peterson has called it brilliant. Stephane Hamel says it challenges traditional web analytics vendors. Jacque Warrens says that at this point, he’d pay for Google Analytics.

That’s a lot of positive buzz. So, here are some comments about this new, exciting feature for little ‘ol me :-)

Guess I’m Floating Comfortably In Data, Not Drowning In It

Let’s start with one general thing: it sucks to have small websites. I won’t go into details, but the average amount of visits we get is on the smaller end because what we advertise is definitely in a niche industry. So, with the sensitivity bar set to low, there is hardly anything of note:

Such is life; makes you want to work harder for every single visit.

A Math Brain For An English Major

What is all this intelligence about, though? After using it for a while I’ve come to the conclusion that Analytics Intelligence is about making my life easier and reports more actionable.

If your position at XYZ company is at all even remotely similar to mine, you get to work in the morning, fire up the computer, open your browser to Analytics and BAM, you have graphs! Then, reality sinks in…what I am to do with all these graphs? What do these lines means? In a word, what really matters? And at this moment your degree in English isn’t helping (that only comes in handy when it’s PowerPoint time and you present insights for all the HiPPO’s).

Easier insights come about because Intelligence automatically surfaces starting points. Some examples:

1. A PPC campaign experienced a huge spike in bounce rate; ruh-roh. Wait, wasn’t there an update to the text ads for that campaign? Yikes, let’s make sure the destination URL’s are kosher. Or something. Maybe the ads just stink.
2. Or, you see that spike in the chart and you get really excited and you’re wondering what in the world could have happened, what wonderful thing took place yesterday and, and, and…actually, nope, nothing to see here. Organic traffic was slightly up, but not really to the point where you should spend an hour digging for insights. Unless you have some time to spare.
3. Or, there are days where Intelligence just goes wild. Direct traffic up! Organic through the roof! One of the highest conversion rates you’ve seen all year! Tons of new visitors from New York! Pageviews…time on site! At this point you can’t even talk; you’re so excited you’re just spewing metrics. Later you find we ran a newspaper ad that day. Sweet, I think there really was a great response. Let’s start digging and see just how great the ROI really was.

These are just a few examples of how Intelligence can do some of the legwork for you so you can get better starting points and act quickly.

Things That Make You Go HUH?

But, there are some things that did remind me this is a new system.

When you start getting to higher sensitivity, some of the alerts kind of leave me wondering. For instance, this alert tells me the conversion rate on the landing page we have for a specific campaign grew (look on the left):

But then, it tells me the exit page was the landing page itself (look at the middle). Is there really a correlation between the conversion rate going up on the very same page where people are exiting? This alert really had me scratching my head.

One thing that might be a great feature is to be able to search for alerts; a small box that you can start typing a metric or dimension that has an alert associated with it. I found myself finding an alert, getting sidetracked, and not remembering what day the alert was on, but I knew it was associated with direct traffic.

And, funny, this draft has been on going for a while now, and Google read my mind about releasing ways to annotate analytics, so that solves my other complaint.

The Verdict

Use it. This is great. As soon as you get used to reading the reports and figuring out how everything is connected, it gets much easier to find springboards for analysis.

An Aside: Some Philosophizing

This is a huge step in the direction of putting business data and analysis in the hands of even smaller companies. I was surprised by the comments on Eric Peterson’s review of GA Intelligence. Here’s a section of my response (for the full version, click here and scroll down):

Will people misuse the information? Sure. Do people understand how it got there? Probably not entirely. But now the stakeholders for small to mid size businesses who don’t know how to calculate an upper and lower control limit with the SUM, AVERAGE and STDEV function in Microsoft Excel have at least another valuable piece of information about their business than they did yesterday. That’s great.

Interestingly, Avinash Kaushik already had this in mind. In a brief email correspondence, he mentioned to me that:

[…] the world of online analytics will be ruled by the messy masses. It has to be. Decision making is so much faster and democratized now. What you see from Google now is a manifestation of that belief. Tools should go as far as they can to make data more accessible, give better starting points.