We will be solving practical challenges through MBA concepts. No theory only applications !

While planning to launch a new product or a new idea, smart way is to conduct market research and take data backed decisions than gut based. It helps in staying connected to market demands.

The central question is to decide GO/NO GO ! Today, we will get little bit deeper into how statistics again come to the rescue of answering the most important question before a product launch/market entry.

**Step 1: Start with a hypothesis **

This is similar to what consultants do day in and day out. They start with a hypothesis for eg, we believe that average satisfaction level of the market for the product ABC will be 3.5/5. Later, we approve or reject the hypothesis with data. We reject in case of strong adverse evidence.

**Step 2: Collect information**

This goes back to our previous lessons of creating a simple random sample which has unbiased and independent information. Get statistics parameters on it such as mean, standard deviation etc on jmp software.

**Step 3: Apply Statistics **

There are two parameters in discussion here.

* Alpha *: It is the value of significance. For eg: your tolerance of error i.e. alpha can be 0.05. You can take alpha as 0.02 in case of solving a murder mystery where you definitely dont want to go wrong.

__p-value__* *: Now let's plot market research data on jmp and draw a beautiful normal curve like below. It will give you a p-value for eg: 0.01 **If the p value is less than alpha (0.01<0.05), then you reject the hypothesis because the evidence is strong enough**. To quote a real life example, in court trials the suspect is innocent till proven guilty so that's your hypothesis i.e. status quo. To reject the hypothesis you need a very rare evidence eg: the suspect was found holding gun and leaving the hotel room at the time of the crime. Now that's very rare data point and the probability of that happening with other suspects is low.

Refer figure below- only look out for giving a picture to p value. You can ignore rest to avoid getting confused.

The key question remains the same. Why am I doing this ? Why is it required ?

Because you always collect a sample which gives you point estimate and never details on full population because it's time consuming and expensive. That's why you rely on alpha and p value to safely conclude answers received from your sample.

In the end I would again emphasize the point of article is to share a new world of decision making MBA learnt tool. It will never be exhaustive. It's up-to your interest if you wish to delve deeper.

If you liked this post, please click that heart icon to show it. I appreciate it.

If you are new here and wish to know the story how MBA in 2 minutes was started, you can click __here____.__

## Comentários