Managing customer data, that vital asset without which CRM can never be a success, has always been challenging.  In the past that challenge was largely limited to tackling silos within the organisation itself and differences in how people defined ‘a customer’.  What data should we collect? What format should it be stored in? How can we eliminate duplicate customer records? Are all longstanding questions CRM professionals have grappled with since the outset of the discipline.  You might well still be asking such questions, in my experience many people are, or you might be asking some slightly different questions.

That’s because the world of data has moved on again, now encompassing ‘Big’ data – some structured, some semi-structured and some completely unstructured.

If we take the Internet of Things (IoT) as an example we can see how the data underpinning CRM has evolved. A chain of gyms can now gather data from a customer’s wearable device and exercise equipment to gain insight into how customers use the facility. They can then turn this insight into value by personalising the customer experience or by offering an additional product such as specific personal training sessions or health food regimes. This information could include workout intensity, duration, technique, heart rate, location, recent injuries and more. Immediately you can see how we’re moving very rapidly past static demographic and past purchase data.

Social media offers a whole host of valuable CRM use cases. But let’s say you work for a travel brand – wouldn’t it be useful to know when your customers are on holiday, where they’ve been recently and which destinations they are discussing with their networks? Again this is becoming a reality for brands across a number of sectors today from financial services to retail.

These two examples signify both opportunity and challenge. Opportunity, as companies have the ability to gain insight like never before that can underpin new strategies and deliver competitive advantage. Challenge, because both examples necessitate a drastically new approach to managing customer data. Both include examples of semi-structured data being harnessed effectively.

Let’s go fishing

While a significant number of companies have established data governance for structured data scenarios, there are very few that have governance for what is known as ‘joined-up data’. Currently, CRM data governance initiatives are typically fragmented as different teams work on specific data silos. Because of this, CRM initiatives have often only had appropriate and assured information for a specific silo. This doesn’t work full stop. But it really is a problem for ‘big’ or ‘joined-up’ data scenarios as the information naturally covers many aspects of the business and what’s more, it comes from a plethora of new sources often outside the company. What’s called for, is a more holistic approach.

The tricky bit is being able to associate the customer information you’re collecting with other relevant data such as social media profiles. How do I know John Smith in my CRM system is the same @JSmith that’s active on Twitter?

The data that’s being drawn from social media and other sources needs to be prepared in order to fully analyse it. For example, machine-generated and social media data often contains duplications and contextual data, which needs to be converted into a consistent format and in some cases additional data preparation may be required before the data can be analysed. Only once all the data has been prepared, cleaned and analysed, can the business start to gain a better understanding of that customer. This has led to the proliferation of data lakes where companies amass great swathes of data in a wide variety of formats – but which on their own are of little use.

There’s no point in preparing data if you don’t have the ability to spot correlations between different data sets. This process of ‘fishing for relevant information’ takes place within and across different data lakes. With data lakes there’s loads of data in many different formats, information that’s incomplete, inaccurate and sometimes can include personally identifiable information (PII) to which access rights must be restricted.

If you’re going to address these challenges, you’ll need a more holistic data management approach that will enable correlations and links to be made between the information and customer profiles. Big data in all its guises offers perhaps the greatest opportunity for customer-focused marketers to demonstrate the value their department can bring to a company but it necessitates effective data profiling, quality and governance tools that can handle any type of data from any source. This, coupled with the ability to integrate all relevant data into existing CRM systems, will help overcome complexity.

At Trillium Software, this is exactly what we do; it’s our job to improve the quality of data by enabling data sets to be brought together, validated and enriched to empower organisations with fit-for-purpose data. This allows CRM leaders to spend time focusing on how they want to put the new data insights to best use for the firm, rather than on data preparation. You could say we provide the fishing rod.

Ed Wrazen

Ed Wrazen


Ed Wrazen, VP Product Management, Trillium.