Customer 360 ! What , Why and How?

One of the most important targets for any B2C model businesses is “customer satisfaction”. Be it through providing a better value for money or through customized consumer support or through any other means of business.

But how would a business know who the customer is , what the customer is trying to do, what have they done before and how best to serve the customer?

What:

Customer 360 is an engineering process of creating a SINGLE VIEW of a customer by capturing various different attributes across several different channels within the business, which helps in predicting the customer behavior

Why:

Lets say a retail company runs business both online and in store and has unique customers in both the segments and omni customers as well. The value of sale in these segments can be very different and are driven by several factors like marketing campaigns, product needs based on demographics, offers & discounts.

When companies start to lose market share after a period of time, it becomes essential to run sales through customer service and customer retention. This is where knowing the customer and predicting the behavior helps in accurately determining the appropriate investments required in marketing, loyalty programs and also in customer support.

A Single View customer 360 simplifies the objective of data democratization by capturing multi dimensional attributes that are accessible from a single source for various internal teams like marketing, IT, customer support, campaigning etc..

Here is a “Customer View” generated from Customer 360 information for a business user: Such a derived metric helps the business users to exactly get a view of type of customers and the various segments they belong to, there by enabling to make better decisions.

And finally here are few use cases that can be developed simply by using the data from Customer 360.

1. Customer life time value
2. Customer propensity
3. Churn prediction
4. Segmentation
5. Conversion/Cross Sell
6. Loyalty optimization

While the above uses cases can also be built without implementing C360, the engineering cost in moving data for specific use cases is redundant. And this is where C360 helps in optimizing the cost of data engineering.

How:

Now that we looked at what and why a Customer 360 view is required for businesses and IT, lets look at how to implement this.

There are several off the shelf products in the market today which can enable implementing the 360 view quicker. However the decision of building an in house platform vs purchasing an off the shelf product is determined by the customization vs cost.

Building a Customer 360 view in house is relatively a less effort for a company whose analytical maturity is already established.

Here are the steps involved in building a Customer 360 view.

1. Identify features

The first step involved in building this view is to identify the various different customer touch points. Here is a list of attributes that are required in a retail industry

1. Demographics
2. Transactions
3. Products
4. Browse
5. Loyalty
6. Campaigns
7. Customer support and other features more specific to the brands.

The features & attributes are very tightly coupled to the business model and vary depending on the type of business and the domain.

2. Collect and consolidate

Given the nature of the business and its underlying architecture of implementation there could be several software and tools a company would have used to solve the specific IT needs.

For example: CRM, SQL & No SQL databases, blob stores, web traffic and analytical data systems, etc… Such varied tools poses a challenge in collecting and consolidating the data. While data collection is one aspect of the problem, what data to be collected is all together a different one and is the most important factor that determines the quality of the feature market. The data should contain hard and soft information.

Hard data is the information that is readily available within the datasets.

Example: Customer’s age, gender, location, support requests, emails sent, discounts applied, etc..

Soft data is the information which needs to be extracted from the given data set.

Example: Customer’s purchase value in the last 3 months, last interaction to an email campaign, preferred shopping segment based on previous transactions, number of purchases through social media ads, number of purchases vs browsed & saved products.

3. Create a Feature Market

A data lake is where all the raw information collected is stored and processed to create reusable features stored within a data ware house. This activity is engineering heavy and is solved through big data solution implementations.

The reusable features to be created are driven by the business problems and the ML models that need to be built to address these.

Customer 360 view is often built using several variables created out of these features and that is what makes it rich and highly reusable.

The process of engineering to create a feature market is for another post, hope this articles address the what, why and how to build a Customer 360 view.