On June 25, 2012 I was honored to be invited to present Meta Analytics, Configuring the web analyst’s mind at the eMetrics Chicago.
The first thing I have to mention is the exquisite organization of the event, everything works fine as a Swiss clock.
Tiffany Harris was running behind the speakers to be sure that all was going perfect. Terry Cohen was the host and coordinator of my presentation, thanks for that Terry!
You can see the tireless Jim Sterne everywhere, making sure that everything goes perfect in the event. You can feel how much Jim enjoy eMetrics, not just like the founder, but just like any other attendee.
But let’s talk about my own experience there. I have to be honest, I was not expecting many people there since I was not talking about the most popular topics like conversion Attribution, Social Media Analytics, Conversion Optimization and Multivariate Test. But I was wrong, lot of people into a huge conference room.
My theme was aimed at understanding what happens before all of the above. Before all that you have people that even when they can generate logical reasoning processes at the end of the day they behave chaotically. Some tips of the presentation:
1. The more professional, less professional: The human mind creates something I call layered knowledge. What is this? When you face a challenge or problem your mind will try to find an answer based on a past event. If it can’t find it, because it’s a new situation, then it stats a research process until get the answer for the problem. When it does it, it saves the analyzed variables like in an encrypted way so in the future when you face a similar situation your mind will just the answer how to solve the problem but you will not analyze all the variables again. For instance, if the first time there was a drop in conversions and there was no record about a similar situation in the past. After a researching time you find that the problem was caused by a poor server performance. When you face a similar situation, a drop in conversions (no matter if generated by the same cause), your mind will try to answer immediately with the tag “poor server performance”.
So even when we learn from specific situations we will atribute the same cause to every situation that looks similar. But in an industry with that is constantly changing like Internet the probability that a present event that looks similar to a past event is generated by the same cause is very low…I would say close to null.
We called this learned knowledge ”transparency”.
2. Breaks: The method to keep learning and seeing each new challenge without the influence of the subjectivity of our mental model and the past, is to create “breaks”. How is it done? In the same way that the kids learn… constantly asking “why”. Do not take anything for granted.
3. The inference is a direct slide to hell: Digital Analyst love the inferences. That is, using the available information guessing the rest by using learned lessons along with “the logic”. As we know the past or “learned lessons” are not very useful (at least if we use that information superficially) and THE logic is not THE but YOUR logic, since it’s based on a particular mental model of each person so imagine how useful it is. For instance: registrations fell, then surely there is a problem with the server (based on the previous example). The “fall in registrations” is based on behavioral information that can just only answer the “what” but not the “why”. In that case you should go for “attitudinal”.
In general we tend to believe that a good digital analytics professional is the one who knows everything about metrics and tools … would that be the answer of why is so hard to find good Digital Analytics Analysts?