The single customer view has been dubbed the holy grail of marketing ever since the phrase was first coined in early 2000s, yet widescale implementation seems to be as far off as ever for many firms.
At a recent DataIQ round-table event only a handful of delegates claimed to have built a true SCV.
For the majority, it remains on their to-do list. One reason for this ongoing delay is the lack of clarity about what SCV really means. Delegates were keen to have a textbook definition of what it involves and what data it should contain.
A spectrum of possibilities emerged from this consideration, ranging from a website visitor to a service user to a repeat customer. Each business tends to have its own definition of when an individual can be described as a customer and this definition can itself vary between departments in the same organisation.
An energy provider offered the challenge of deciding which is the most important data point for the company – the meter, the household or the bill-paying customer?
If this fundamental definition can be agreed upon, it should help to set the parameters of how much data will need to be included in the SCV – another area of debate on the day. A narrow view, such as only customers who have carried out paid transactions, might be limited to ten columns, whereas a broader view across the entire customer journey might contain hundreds of columns. It is this business view of the customer which will ultimately establish the technical specification for the SCV.
Building the business case for SCV
If SCV is understood as a concept, it is less well understood as a business requirement. Many delegates said they were struggling to build the business case for investing in SCV and this extended to whether the project would cost several thousand pounds or several million. One reason for the upper limit was the view that, to justify its name, the SCV would need to be perfect. By contrast, the view was also expressed that perfection was not necessary, it was just important to make a start.
Three main foundations of the business case emerged across the day:
• Better business – understanding the full profile of customers, from their light-touch interactions on a web site or app through to transactions, is recognised as best practice. It also allows for better marketing planning and customer experience management through being able to track the customer journey and identify opportunities for up-selling and cross-selling.
• Better compliance – with the enforcement of GDPR from May 2018, organisations will need to be able to respond to customer requests for access and deletion of their personal information, as well as being able to document exactly what information is held and how sensitive it is for regulatory reporting. SCV supports both of these demands.
• Better customer satisfaction – delegates agreed that consumers have become savvier about how organisations are capturing and using their data. To ensure this is sustainable, those brands need to be able to demonstrate the value customers get from sharing data through the personalisation and responsiveness they experience.
Not all data is equal
A commonly-shared view among delegates was that creating a single view is harder for well-established organisations than it is for more recently-created entities. This is because, as a business grows and expands, it tends to create data silos. Customer data may be shared and spread across these, especially if each product or service maintains its own customer database. Getting a view across all of these sources is at the very heart of any SCV project.
In identifying where data is being held on customers and what type of data exists, an important issue to keep in mind is that not all data is equal – the reliability of different data points, as well as their value to the business, can vary considerably. Matching points for data integration are a key factor. Delegates identified the following challenges around data types:
• Multiple email addresses – it is a common experience for customers to appear across multiple files with different email addresses. A consensus view was that the most recently provided (or responded to) has a higher value.
• Date of birth – although generally thought of as a gold standard data point, a DOB provided when an individual has applied for a loyalty card may be less reliable than one given to a healthcare provider.
• Telephone-captured data – information which is captured during the course of a customer service call was viewed as less likely to be reliable due to the time pressures, risks of mishearing, mistyping or incorrect transcription. These kinds of data problems routinely enter SCVs upstream and can be – but are not always – resolved during downstream data cleansing.
• Online customer-volunteered data – despite the similar risks that a customer may mistype their personal information, the combination of online validation and visual self-checking – we are more likely to spot an error in our own name and address than in somebody else’s – give this source a higher rating than telephone-captured data.
Although this hierarchy was generally agreed on, there was less agreement about its consequences. To some delegates, it supported a strategy of data-minimisation which focused only on high-value data points. To others, it was an argument for retaining everything for as long as possible.
The problem of finding the right match
The problem of having multiple records for the same customer seemed very prevalent. As one delegate pointed out, two people with the same name or same initials may – or may not – be the same person. Another reported that their organisation has 28 customers called S. Smith. Elsewhere, one customer with three email addresses was noted.
To resolve such problems, “fuzzy matching” is commonly deployed, especially for marketing data. But this may not produce a match to the confidence level required for a legal join within a regulated business.
One solution is to encourage customers to confirm their data whenever they are contacted, with incentives offered to do this, such as better offers or discounts. This had been used by both an energy company and a grocery chain.
For those regulated businesses, making the effort to ensure a legal join between duplicate records has specific business benefits – one example cited was avoiding the risk of offering financial products to a customer who has been declared bankrupt.
The underlying fear is of creating “false positive” matches where two records are joined that are actually different people. To avoid this, many delegates kept matching rules tight, resulting in an ongoing volume of duplicates and potential overlap between customer records. Emailing customers ahead of merging records to reconfirm key details was one method being used to minimise the risk in this situation. The general consensus was that matching should be done on the basis of, “over-dupe for marketing, but under-dupe for compliance.”
Technology to the rescue?
Much discussion took place of technology and the platforms that could help companies create SCV. Delegates from companies that were already using data management platforms offered advice to those that were perhaps considering it.
A wide range of approaches emerged, with three key solutions being used:
On-premise – a significant change has been taking place within the on-premise solutions stack being deployed against SCV. Many organisations are switching out of long-established database architectures into commodity or open source solutions, such as Hadoop. These can be lower cost, but may also present integration issues, especially where the data being managed needs to flow into marketing campaign management and CRM systems.
Cloud-based – elements of SCV, the analytics deployed around it and the marketing systems drawing on it are being pushed into cloud-based services. In addition to the perceived cost benefits, a major factor in this migration is the improved compliance offered by cloud service providers as well as their strong focus on data encryption, which can often exceed the levels experienced on-premise.
Hosted and hybrid – for marketing-focused SCV, a wide range of marketing services providers are being used, including the roundtable discussion host Blueberry Wave, as well as providers focused on specific aspects of customer management, such as data on-boarding and identity resolution. All of these can be integrated into the wider enterprise environment through APIs, removing many of the complex data architecture issues.
Looking to the future, the requirement for SCV may even be removed as GDPR impacts fundamentally on the way individual consumers manage their own data. Through the new right to data portability, an emerging sector of data middlemen to arrive who manage this interaction, handling consumer data and offering it to organisations based on the rules set out by each person.
GDPR changes everything
With GDPR enforcement coming into effect next year, many delegates felt the need to begin the process of SCV implementation. One assessment of the current situation was blunt: “We’re not ready.” Notably, this same organisation is taking a longer-term perspective of GDPR as a macro change in the data culture, rather than a short-term shift.
There was a real concern about the lack of guidance coming from the Information Commissioner’s Office of what needs to be done in preparation for GDPR. As one attendee said: “It’s coming, but who has said ‘here are the five things you need to do’?” Another pointed to the Regulation’s focus on the importance of the individual’s rights: “This is why they created GDPR in the first place.”
With this lack of guidance from the ICO, some companies are hazarding a guess at what they need to do. Several delegates said they will update consent from all their customers – a practice others are also considering.
Steve Mattey is head of operations at Blueberry Wave