March 2007

27 Mar 2007 01:07 am

violetA “aspiring SEO & online Marketer” in Germany asked a interesting question via email and I as I replied to the email it seemed like one of those things that others might be interested in. Then I got another one from someone here in the US and that was a good one too. Net net here's a blog post!

The first part of this post are the two email exchanges and the third part is a realization I have come to recently about a key ingredient that every great Analyst I have met seems to have.

Does a analyst need to know statistics?

Here is a excerpt from the reader’s email and my response…..

If I want to become solid at web analytics, will it make sense for me to learn and become very proficient at the bases of traditional statistics (while Im in college)?

I would absolutely recommend getting a solid grounding on basics (and maybe just a smidgen beyond basics) of statistics. This is helpful for two reasons:

    # 1 Great analysts will know how to leverage the awesome power of statistics in their analysis, such as I have outlined in these two posts:

    Both posts outline how you can apply even the simplest of the statistical principles to great benefit in doing analysis of your web data.

    # 2 Increasingly traditional clickstream analysis is becoming less and less insightful (because of the inherent complexity of the web and customer behavior is increasing with every passing day).

    Experiment smallFor this very reason I have often recommended a focus on customer experience and on experimentation / testing, because both are the keys to unlocking the insights that are truly actionable. To truly exploit these options for your business you’ll need to be good at statistics.

    As an example in my past life I have used the ACSI (American Customer Satisfaction Index) model for doing surveys and it is essentially a complex multivariate regression model on top of your survey data. It helps bring hard core quantitative power to something so qualitative as survey responses.

    Likewise I if you do multivariate testing (or even A/B) you will need to know the basics of statistics as well as you analyze the data and make decisions based on when the result set has reached the desired confidence levels.

You don't have to be a PhD in statistics, but being a smart person with a familiarity of statistics (basics and just a bit more) will be greatly helpful in your career in web analytics.

Perhaps the most critical skill, IMHO, is your ability to do analysis, your ability to think analytically and look at every problem differently. Anything you can get exposed to in order to become a great analyst will be awesome for you. This post might be helpful : Top Ten Signs You Are A Great Analyst.

Are you on the “business” or “IT” side?

Here’s the second interesting question….

My silly question is that do you see yourself as on the "business" side of Web Analytics or the "IT" side? Should there even be separate "sides"? The reason I ask is that my interest in Web Analytics is mainly on the what to measure, how to read results, how to present and communicate the result, and how to influence changes based on results. But then I want to have good understanding of the capabilities of the different tools and data quality, etc. It's a dilemma to have to choose to be on one side or another.

choice From my experience if you want to be a Analyst then you are on the "business" side automatically. I have spent a few years as a part of IT but IT is not business and there is something core about the integration and mindset that is quite different. That's based on my experience. So if you want to read data and analyze data and present it then it is better to be on/in the business side.

That said the best analysts have a deep technical side, in the sense that they need to be comfortable understanding javascript tags and urls and parameters and redirects and web pages etc so that they are good at one singular thing: How does the data get collected and then how it is interpreted by the analytics tool. That will ensure your long term success.

So in summary, get your salary and comp aligned to the business side and keep your friends, and smarts, on the IT side so you’ll can understand what's happening behind the scenes. Then you are set for success. :)

The common trait that great analysts share.

Writing that last paragraph in the first email made me think of all the people I had thought were great analysts, and what did they have in common. One obvious thing was that they were good most or all of the things mentioned in the Top Ten Signs post. But after a bit more thought it struck me.

All great analysts I had ever known had amazing (maybe even crazy) formative life experiences.

They had done ok in school, they were not even remotely straight A’s (in terms of grades). But they had rebelled and had a crazy youth. Or trekked through the woods of Laos on a whim. Or had studied to be geologists or forest rangers. Or never finished college. Or had lived in four countries (mostly on very little money). Or had gotten into trouble with the law (in minor ways). Or… and I had a bunch of these examples.


Each person was, perhaps even unknown to them, a masterful student of life after having lived it very richly, and that had imbibed a ton of common sense and reality in them. They had observed the true complexity of life.

All of them ended up in Web Analytics more by chance than design. But their experience made them particularly adept at solving mysteries (and web analytics is more a mystery and less a puzzle). 

I think their experiences encourages them not accept obvious answers, but rather look below the surface. They are very good at unstructured environments and don’t get paralyzed when all the questions don’t have answers.

I find that they are great learners and rather than being satisfied with being report writers they dig, they probe, they make noise, and hence are amazing at analysis.

life experiencesSo if you want to be a great analyst then take a detour or two in life, you don't have to backpack through Laos but you can volunteer at a local shelter or start your own business or participate in a study abroad program or… well you get the drift.

If you are looking to hire a Analyst look a bit beyond the stand resume entries and look at the "life entries", look beyond the nineteen years of experience with WebTrends or Omniture or HBX. If you have a choice then choose someone with a life experience because you can always teach them how to press buttons in a web analytics tool or write a formula in Excel.

[In case Damini or Chirag are reading this: Daddy is not saying that education is not important or that you should not shoot through school, college and university with straight A’s. It is important, and I am just sharing my a observation that might encourage you all to stop and try bungee jumping. Ok maybe daddy would strongly discourage that particular activity! :) ]

What do you all think? Do you agree, both on the value of statistics and aligning oneself to the business side and mushrooms? What's been your path to Web Analytics? Any interesting stories you would like to share? How about some insightful career advice? Please share your feedback via comments.

25 Mar 2007 05:12 pm

union 20jackI am speaking at the Emetrics summit in London this week and looking forward to the event (Mr. Finlay and Mr. Sterne have ensured each event they present is better than the last one!).

I was wondering if anyone would like to get together for dinner on Friday night, March 30th.

When we did this in November, over some excellent tapas, we talked about web analytics, our challenges and solutions, the weather, the Queen and pretty much everything in between.

We can meet at Hotel Russell (Russell Square, London WC1B 5BE) in the lobby at 1800 hrs and go some place close by (I am not familiar with the location so please suggest a appropriate place for dinner).

If you are interested please email me at blog at kaushik dot net (or add a comment below).

I would love to meet some of our blog readers, or web analytics practitioners, since we have so many in the UK.


Next Page »