Creating a single customer view – harder than you think

It’s fascinating to see the contradictory responses that one gets when asking large organisations whether they have a single view of the customer – everything from “Of course,” to “Partly” to “Not at all!”

And these responses often come from the different stakeholders within the same company.

No consistent understanding of what a single view means

From an IT perspective, the single customer view is regarded as a data integration problem.

But to business, having customer data consolidated into a master data management application often leaves a bad taste.


Well, one reality is that the understanding of what customer data is is also changing.

Where a pure master data approach might consolidate customer identifiers and contact information, business users may require other, related data to be present to give them business context. Service agents need to have a record of service requests and resolutions offered; marketing teams may want to add demographic data, or buying patterns; and for customer experience we may want details on preferred channels or times of engagement; and so on.

For these users, the single customer view may be better represented by a Customer Relationship Management (CRM) application.

And for analytics purposes we may deliver a data warehouse, or a data lake, to allow us to track customer trends and to feed our artificial intelligence models

Quality data the common ground

What almost every one can agree on is that quality customer data is the foundation of a single customer view.

Whether in your CRM, in a master data system, or in your data lake duplicate customer records erode the value of the solution by creating confusion and inaccuracy.

In our experience, a data quality strategy for a single customer view must achieve two separate goals.

  1. We must identify the “core” data fields that can confidently group individuals, households or businesses. We need to recognize that data may be missing, or inconsistent, across records or systems, and our match strategies must accommodate these inconsistencies – and present results that our stakeholders in business and IT can trust.
  2. We must link and, where necessary, enrich the related data fields that serve specific uses.

For example, to assess your customer’s buying journey your marketing team may wish to bring together four sets of data:

  1. Contact data: This includes names, titles, roles, email, postal, and phone information; gender, date of birth and so on
  2. Demographic data,:such as industry, region, company size, and could also include things such as technology platforms, etc.
  3. Engagement behavior data: This includes channel preferences (online, phone and other offline channels, such as events), touch point engagement (website, help desk, inside sales, outside sales) and content preferences (white papers, videos, demos, etc.)
  4. Sales data: This includes transaction level data such as product history, buying history, payment history, etc.

Without this combination of data sets it can be very difficult for marketeers to define buying personas, build predictive models for buying journeys, or even create simple marketing segments.

To achieve the single view we must first group related records (individuals, household or businesses) and create a common record. This typically will use contact data – which may in its own right be drawn from the CRM, the billing engine, product silos or a variety of other systems.

Once we have done that we can link the related data – demographics, product and engagements to that one single record. This is the only way to ensure an accurate assessment of the customer’s history.

Ultimately, the single customer view is not a technology.

It is an approach to customer data that delivers quality, complete, consistent customer data to the user that needs it, irrespective of the purpose.