One of the critical parts of an Analytics Ninja work is related with the way humans interact. Analytics Ninjas must understand what people need to make decisions and what they don’t. However, this task that seems to be pretty simple, in fact it is not.

First of all every human has a different mental model which process the information in a very particular way. The communication also differs based on gender, man and woman communicate in a very unique way.

Culture is something that pushes the differences among people. People from Argentina process the same information different than someone from the States (look other posts like “What’s a Latino?“, “Analytics, Focused on People?” and “Is time Just money?“.

All the above mentioned differences make our job very complex. Why? Because the first part of an Analytics Professional work is understanding the “System” and you wont be able to do that if you are not able to understand the way people communicate it to you.

So my suggestion is using the kid’s very same technique…the Why? Why? Why? Technique. Normally the less people talk the more they agree. This is usually desvastating…

I Leave you a very interesting video from the Monolog of Mark Gungor called the Tale of Two Brans, enjoy it!!!

I do agree with you if this post title sounds a little funny to you but I’ll show you that it makes sense.
I’ve post several times about the importance of integrating information, it provides us the total picture of our situation avoiding inferring part of it, which means building better scenarios for most efficient decision making.
Lets use an example the ecommerce site from a computer company. What information this company normally use to understand what is happening with their businesses? Traffic and sales. Both are behavioral metrics which means that tell us what are people doing but not why.
What happen when there is a slow down in sales? Well, even when the available information is not enough for making an efficient decision, people tend to infer the part of the situation that is not answered with the available information.
So when there is an slowdown in sales people will infer that the problem is the promoted product, a non so efficient marketing campaign (paid search, banners, offline, etc) and several other reasons that sounds logic to the person that is analyzing the information (and off course, enough logic to convince the rest of the team).
However making decision in the above mentioned environment is not so healthy, not just because the decision may not solve the slow down in sales, but also because the company may be investing money in the incorrect place increasing the acquisition cost situating the company in worst situation. Lower sales at a higher cost.
Fortunately today we have lot of information sources not only with behavioral information but also with, for example, attitudinal information, which gives lot of relevance to the behavioral information.

So lets see an interesting, and very simple, four lines chart. The lines represent:
1-    Visits (Behavior).
2-    Sales (Behavior).
3-    Buzz (Attitudinal).
4-    Server Performance Monitoring (Environment).
In the X axe we have time (in days) and in the Y1 axe we have Q (quantity) and in the Y2 axe we have percentages (for Buzz and Server Performance).

I also recommend adding information about Events that are relevant to the company and its projects. That is qualitative information that gives relevance to the qualitative information. I called that information “Internal notes”.
In the following example have two situations that require making decisions that are based in the following triggers.
1-    Alert – Drop in sales (9th of January): The marketing manager receive and alert from the Web Analytics tool because sales are down our daily average. The Marketing manager look at the four lines chart and finds out the following.
a.    Visits: No important variation in visits, no further analysis.
b.    Sales: As the alert informed, sales droped in more than 43% so further analysis is required.
c.    Brand Buzz: There is no important change in the brand or product buzz which means that we cannot asign the drop in sales to an attitudinal problem or perception regarding our brand or product.
d.    Server Performance Monitoring: There is a important drop in the server performance right on one of the conversion pages. Some minutes later in the shared notes appear a note from the Technology department notifiying the server issue.
In just 10 minutes of analysis we’ve got the answer and can correct the deviation. Sounds like actionable information, isn’t it?

2-    Alert – Drop in sales (18th of January): The marketing manager receive another alert to his email from the Web Analytics tool, apparently there is a almost 25% drop in sales. Whitout the four lines chart the same manager could try to infer that his company is experiencing the previous experienced server issue. With this model the Marketing manager look at the four lines chart and finds out the following.
a.    Visits: Visits are slighly growing, however nothing very important at this moment (howevery look at the data table and you can see that if the manager waits untill the jump in traffic is higher than normal, its too late for solving the problem).
b.    Sales: Almost 25% drop in sales.
c.    Server performance: Even when the manager look at this metric first based on the previous issue, he easily understands that there is no problem in the performance of the server.
d.    Brand balance: There is a negative variation (higher than the normal fluctuation) in people’s perception about our brand or product. Great! That tell us a lot about the problem. Visits increase as a result of the increase in buzz (when people talks about us, no matter they do it good or bad, our brand is more present in the web), this is because people reads about an issue about our product and inmediatly look at the company’s site in order to colect more information and, off course, to confirm the rumor.

In this example the problem was that our Laptops batteries had a contruction problem that prevent the bateries to charge after four of five charges. An important blogger commented about it and in just three days the information was distribuited all through the internet.

The negative information about our product makes that the higher traffic could not be converted in sales. People visited us to know about the problem, not to buy a deficient laptop.

The company found the problem in 10 minutes of analysis, and converted the information into action focused on return the company to the previous situation.

PR: wait… I: wait… L: wait… LD: wait… I: wait… wait… C: wait… SD: wait…