The document discusses building a customer 360 view using big data and modern data management. It describes key challenges in creating a customer 360 like data silos, large and growing data volumes, and new data sources. It then presents an architecture using an Enterprise Data Hub to ingest diverse data sources and enable analytics to build a holistic view of individual customers. The approach advocates starting with core customer data and iteratively expanding the view by adding new data sources and delivering specific use cases.
Before I go into detail on the Customer 360 - At Cloudera, what we see is three areas of opportunities within businesses generally are:
Data-Driven Products – How can I build better data-driven products and services, at lower cost?
Security, Risk, and Compliance – How do meet compliance regulations and preserve data security to minimize our corporate risk profile?
And for today, focusing on Customer and Channel – How do I build a 360 picture of my customer?
Customer 360 is not necessarily something new. Many industries such as the Telecommunications & Insurance have been doing this for many years.
The thing that has changed however
diverse set of data sets that available
Richness that this data brings
And as we use data to build a profile about the customer, it enables us to
Understand better how are customers
So that we can better serve and meet their needs
And these factors are driving how a modern Customer 360 view is built:
Look at the new sets of data sources and emerging data sets to build out a real-time holistic view
Could be Location Based Data
Clickstream Data
External data such as Mobile & Social Media feeds
And for organizations that deliver via omni-channel
customers are now demanding a seamless experience across channels
If your customer profile or view is based only on data from a single channel, what do you think experience is like in other channels?
This needs to drive the buildout of the profile, to extend across the channels that are being delivered to.
And all of this is done to ensure that:
You can deliver a consistent experience across channels
You can make personalized recommendations to the right customers
You can deliver targeted experience by knowing the behavior and patterns of your customer
However there are still many challenges for customers in building their 360 degree view out.
And consistently across the board we are seeing the same issues:
Data Silos
Data is spread across multiple sources, often no consolidated view. And a company with multiple business units often has duplicates sets of common data; but this data is often conflicting, from system to system and BU to BU making it hard to identify which is correct.
2. Volume of Data
As systems are generating more and more data (often doubling year on year), keeping this data is posing a challenge.
3. New Data Source
Gone are the days of storing just batch, structured relation data. Many systems now are generating semi- or unstructured data and often in real time. And its often these that are critical for complete view.
4. Cost of Data Storage and Processing
Traditional systems and architectures are expensive to storage large amounts of data (between 30,000 – 100,000 per TB depending on the system). To be able to keep the volume of data required and process it is not economical at those rates.
And if those challenges are not addressed- what you are left with is a customer view that is not a full and complete representation.
And generally what that means is that its:
Only focused on Structured Data
Only contains internal data
Only has what someone has deemed “important” (but is that really the case)
And Only Limited history is kept.
And this approach is not sustainable and won’t organizations address the complex challenges they are facing.
First, Lets take a look at a Traditional Data flow common today…
On the top left you see some of the traditional systems with structured data – Billing & Ordering Systems, Product catalogues, CRM Profile Data & a lot of info from Marketing campaigns & Surveys – and these are very well structured
On the bottom left you see some of the newer sets of data sources – including sources such as Location data, social media streams (for example Facebook, Twitter and Linked In) or could be website clickstream data.
This new data is either not collected today or a only very small subsets are; or data is highly aggregated and then all of this is commonly fed into the Data Warehouse
But the problem with this approach is that it has many issues with it:
First of all these systems are really not able to handle the data that is being generate – both from the volume of data that is coming in and also the streaming aspect that they commonly have to deal with.
Plus - some of these new sources may not be easily be able to be ingested into traditional database schemas
From a processing point of view, systems that are doing the ETL and stored procedures have limited processing power; and the data marts end up need to do the aggregations
What you end up with is a loss of granularity with your data that is in your data warehouse. You no longer have the full set of data, you only have aggregations of data that have passed through the ELT processing.
So, there is a new way forward and a new way to address these challenges:
But lets contrast this with a Cloudera Enterprise Data Hub. It provides a new approach and a number of key differentiators to the traditional architectures:
It can ingest data from multiple sources; streaming as well as Batch processing while enabling you to can keep unlimited data online without needing to archive
Economically feasible to store more data (cost can be 10x cheaper than traditional systems)
Powered to predictably process large data sets
Ability to build your data asset at linear scale
Collect data in native format – enables agility
2. Diverse users can get direct access to all business relevant data, through the best tool for the them. That could be
SQL,
search,
or your existing BI or analytics tools such as MicroStrategy, Tableau etc.
Users who previously had no way to benefit from data can now find and generate insights.
3. Finally All of this can be done with confidence, thanks to Cloudera’s enterprise-grade security, governance, and management tools Either on-premise or the cloud.
While this afternoon will talk about Cloudera tools in the cloud, certainly many customers are choosing Cloud to deploy solutions such as Customer 360.
Cloud enables a new infrastructure paradigm with flexibility & easy scalability
An when combined with Cloudera’s automated tools makes deploying managing and growing clusters in the cloud quick & easy.
And the great thing is because Cloudera’s solution can be deployed in a similar manner across all Cloud Providers you can avoid any vendor lock in or even have Cloudera in Multiple Cloud’s simmultaniously
And so then in contrast, the Data flow architecture with Cloudera is simplified; addressing the key challenges as:
All the sources are ingested into one location
We are able to handle all the different types of data
Standard way to process and analyze both batch and streaming data
While Enabling multiple use cases off a single platform
Cloudera we like to say “Think Big, Start small and Iterate Often”
Start with ingesting the “best” version of your customer profiles from a transactional system or an existing data warehouse
Identify your common identifiers across datasets: customer name, number, IMEI, IMSI
Enrich with additional demographic information from other systems
Deliver your first use case with this information, e.g.: Lifetime value modeling, Device and plan modeling, Next device offer
Continue to add datasets – such as purchase behavior - and explore common identifiers across your datasets
As you explore those new datasets, enrich your customer profile with the additional information
Continue to deliver additional use cases,
Although there are many applications that can be applied to Customer 360 – I wanted to go highlight some of the common ones that we see time and time again:
Targeted Marketing & Personalization
So - making sure you are focusing the right things to the right customers. As you can imagine most people don’t have unlimited marketing and advertising dollars – so ensuring this is spend to the maximum efficiency is critical….
And this can be done through activities such as
Offering personalized product offerings or derive specific upsell/ cross-sell opportunity to and existing customer
Or proactively present the right offer, to the right person at the right time based on some event that has happened.
Proactive Care
Which is all about improving the customer service experience.
Organizations are building intelligence and analytics tools so as to proactively identify issues and fix it or offer a solution before it impacts the customer.
And Not only does this provides a compelling customer experience; but it also deflects and prevents calls to the customer care centers thereby lowering costs.
Finally -> Churn Prevention & Customer Retention.
Given the impact of customer churn affecting the Insurance and Telco industry today, we are seeing Big Data & analytics to bring together various data points including - quality of service, network performance, billing information, details on calls to the care centers, and social media sentiment analysis to design and build an effective model to predict and prevent customer churn