**Averaging averages**

Let’s for example consider that we have data broken down by gender, age, and location. We now want average income by gender. Therefore we will average all the female values and average all the male values. However, this will be a wrong approach as in case the age and geographic distribution is not perfect according to gender we will get an error.

e.g.* 1) Females , 26, New York – Average: $30, 000 out of 10,000 people 2) Females, 26 New York – Average $31, 000 out of 11,000 people*

Then normally we would calculate the average income as $30, 500 but this is wrong number as the second has a bigger sample size and the correct answer is $30, 524 due to the hidden variable. The same mistake can happen when we calculate ROI of successive months and try to obtain average ROI for two months.

**Correlation Coefficients**

Correlation coefficient is a relatively simple concept that can be used to measure how closely two variables are related. Excel has the ability to find this for you in second. It has inbuilt functions that help you find it. There are several YouTube videos that will guide you how to scatter the data and then find the correlation coefficient.

**Probability **

Probability helps us predict the probability of event B given that event A has happened. Mathematicians write that as P(B│A) the mathematician Bayes found out that this = P(B and A)/ P(A)

In short it means that probability of both event B and A happening divided by the probability of A happening

**Writing programs**

You may use programs in languages like C or pascal to write and analyze data. However, when large datasets are concerned opt for a simple language rather than a difficult language like C.

**Nash Equilibria **

This word sounds too geeky but if you looked up the film A beautiful Mind you will understand it lucidly. This theory helps us understand how two partners cannot do any better by changing strategies given the opponent’s likely to response. This theory is great in determining pricing strategies.

**Time series**

This refers to data on a time line, such data comes from **analytics**. Again Excel is equipped to do a time series analysis. This is peculiarly important when doing decomposing analytics. For example when you want to know the effect seasons have on your sales. This type of analysis gives you insight on underlying trends that otherwise would not be so evident.

In the same way you can use concepts like Prime numbers/RSA, Eigenvalues, and Markov chains to gain insight to your marketing efforts. The best part is that this type of analysis gives you a very objective picture that certainly helps in understanding and making corrections.