This is a quick crash course video where we’ll talk about customers analytics, data science,

and how the two work together!

The topic we will be discussing here is price elasticity.

Leading companies are always on the lookout for savvy data scientists to join their fast-growing

Customers Analytics teams.

In that sense, considering a career as a data scientist in customer analytics is a super

smart choice.

But here’s why exactly: First, companies need people who know how

to use data to understand their customers' needs.

Once they understand their needs, they can provide the products customers want to buy.

Second – and that’s a bit more technical – companies need people who have the skills

to build the analytics capabilities that will help them provide these innovative customer

experiences.

In these videos, we’ll be focusing on the customer part of customers analytics.

Why?

Because even if you know how to do the technical analyses well, unless you understand the customer,

you won’t be able to meaningfully help your company.

So let’s build those foundations, shall we?

Just one more thing before we get started!

We’d like to mention something else we’ve put together – a very comprehensive data

science training.

The 365 Data Science program contains the full set of data science courses you need

to develop the entire skillset for the job.

It’s completely beginner-friendly.

For example, if you don’t have any maths or statistics knowledge, we’ll teach you

that first.

And if you’d like to build a more specialized skillset, you can do that with courses on

Time Series Analysis, Credit Risk Modeling and more.

If you’d like to explore this further or enroll using a 20% discount, there’s a link

in the description you can check out.

Perfect!

Now, let’s get into customers analytics and more specifically price elasticity.

In the broadest terms possible, price elasticity measures how purchasing behavior changes when

the price changes.

For instance, let’s assume a bottle of Coca-Cola costs $1.

If the price increases to $2, many people would stop buying Cola, as it would be too

expensive.

On the other hand, if the price decreases to 10 cents, many more people are going to

start purchasing Coca-Cola.

The measure we use to quantify this phenomenon is called: ‘price elasticity of Coca-Cola

demand’ or ‘price elasticity’ in short.

Now, as you can imagine, if we assume that a Pepsi also costs $1 and Coca-Cola cuts its

prices to 10 cents, then most of the people that used to buy Pepsi, would immediately

transfer to coke.

If prices of coke increase to $2, instead, people will likely stop buying coke and turn

to Pepsi.

That is to say that there seems to be another important phenomenon - ‘price elasticity

of Coca-Cola demand with respect to the price of Pepsi’, or the so-called ‘cross-price

elasticity’.

Now, in the first case, where we measure the elasticity of Cola with respect to itself

only, we call that own-price elasticity, price elasticity of Coca-Cola demand, or simply

price elasticity of Coca-Cola.

However, in the second case, where we’ve got 2 products, we must say the whole name:

‘price elasticity of Coca-Cola demand with respect to the price of Pepsi’, in order

to be sure that there will be no confusion.

Alright.

Now that we have an idea about what price elasticity is, let’s discuss it in economics

terms.

Price elasticity stems from the basic economic law of supply and demand.

The cheaper the product – the higher the demand.

The more expensive the product, the lower the demand.

Simple as that.

It is extremely important for businesses though, because there is this sweet spot which maximizes

revenue.

Since revenue is equal to the ‘price’ times ‘units sold’, we can use this price

elasticity concept to find the point at which ‘price’ times ‘units sold’ is optimal.

Okay, but how does this look in mathematical terms?

Well, price elasticity is the percentage change of an economic outcome of interest in response

to a 1% change in the respective price.

Usually that economic outcome of interest is the number of units sold.

Let us denote the economic outcome of interest with Y, and price with P. Then, the price

elasticity of the Y must reflect the percentage change in Y in response to a 1 percent change

in the P. We can obtain that by taking the percentage

change of Y and dividing it by the percentage change in P. Well, the percentage change of

Y is the difference between its present and past value, divided by the past value.

Similarly, the percentage change of ‘Price’ is the difference between the present and

the past price, divided by the past price.

Okay.

Now let’s tie price elasticity to the subject matter of our course.

Remember, we’re going to address three questions related to economic outcomes of interest:

(1) Will a customer buy a product from a particular product category when they enter the shop?

(2) Which brand is the customer going to choose?

(3) How many units is the customer going to purchase?

These three questions boil down to the estimation of the following economic outcomes of interest:

(1) Purchase probability, (2) Probability for brand choice, and

(3) Purchase quantity.

Naturally, we would be interested in price elasticities of each of these economic outcomes,

so we’ll look at: (1) Price elasticity of purchase probability,

(2) Price elasticity of probability for brand choice, and

(3) Price elasticity of purchase quantity.

Okay.

It’s time to see why we need each of them.

What can we expect when we calculate the price elasticity of purchase probability?

Well, there may be a lot of different brands with differing prices from the same product

category on the market.

For instance, the price of different brands of beer.

Suppose we can calculate an aggregate price for the whole category.

The law of demand says that the greater the price, the lower the quantity that customers

want to buy.

So, if the aggregate price increase, the probability of purchasing a beer would decrease.

And calculating price elasticity will show us how much exactly.

What about price elasticity of probability for brand choice?

Well, that’s the most interesting one.

And it’s the most important for marketers, as well.

If you’re working for Oreo, you’d be interested in the Oreo brand, not in the aggregate biscuits’

sales across the board.

That’s why all marketer’s efforts are devoted to influencing customers to choose

namely their brand over competing brands.

Similar to purchase probability, we can assume that if the price of a product from a given

brand increases, the brand choice probability for that brand decreases.

Again, calculating price elasticity of brand choice for a brand with respect to the price

of that brand would show us exactly how much.

Accordingly, if another brand increases its unit price, the brand choice probability of

the brand of interest would increase.

This is precisely the Coca-Cola – Pepsi example we started with.

Such elasticity calculations will show us how much the brand choice probability of our

brand would increase with a one percent increase in the price of a competing brand.

Nice skill to have, isn’t it?

We’ll learn how to do that and even more.

Finally, we’ll discuss at length price elasticity of purchase quantity.

As you might expect, following the law of demand, the greater the unit price of a product,

the lower the quantity that is going to be purchased.

For a car, the difference may be from 1 to 0.

If the price of a Tesla is acceptable to us, we will buy 1 unit.

If it isn’t, we won’t buy any.

Alternatively, if we’re considering, let’s say, avocados, depending on the avocado price,

we may decide to buy 0, 1, or even 10 avocados at once.

Calculating the price elasticities will show us exactly how the purchase quantities move

with the change in price.

I bet you’re impatient to see how all this is done and learn to do it yourself.

Well, let’s get to it.

To do all this, we could use many statistical software packages, such as SPSS, SAS, and

R. Here, we chose to use a statistical computing environment, which is widely used, has growing

popularity, and indicates it has the potential to become the most popular amongst data scientists:

Python.

We hope you found this video helpful.

And if you enjoyed it, please take a second to subscribe to our channel, hit the like

button, and share the video with your friends!

Thanks for watching!