Building the single customer view

In 5 steps with Linked Data

Single Customer View (SCV) — or “single view of the customer” — means an aggregated, consistent, up-to-date and holistic representation of all the data a company holds about its customers.

Delivering the SCV brings evident advantages: a better customer experience, better results from marketing initiatives, better conversion of the sales pipeline. In many cases it is indispensable for running an omnichannel and profiled management of the customer, and for orienting the whole organisation to the market. That is why many companies start SCV projects that all too often never reach completion, wasting resources and energy.

The following paragraphs explain why this happens, and how a different approach can deliver the SCV using linked data in 5 steps:

• STEP 1. Understanding the Single Customer View
• STEP 2. Quantifying the Single Customer View
• STEP 3. The barriers to the Single Customer View
• STEP 4. How to build the Single Customer View using Linked Data
• STEP 5. The stages of a Single Customer View project

STEP 1. Understanding the Single Customer View

First of all, the SCV is not a product or a solution but a process concerning all the customer’s data.

Representing the customer means being able to “draw” and “describe” them using at every moment every possible piece of data available, organised by meaning and type. This representation makes it possible to understand and analyse the customer by considering them in their entirety. It is like seeing a puzzle when it is complete: the picture reveals its meaning only by putting the pieces together and looking at it as a whole.

Aggregating the customer’s view means bringing together and “connecting” the very many pieces of data a company may hold about its customers. Internal data such as registry, socio-demographic, geolocation, transactional, administrative and behavioural data, data on their actions and interactions, their purchases, their privacy consents, their use of services and channels, after-sales, support, their preferences and their contact history. External data such as data harvested from the web, or coming from other sources such as suppliers or partners. To have an SCV these data must be brought together in a single point.

Having a consistent view means guaranteeing that there are no duplications and, above all, no inconsistencies. Without an SCV, companies often duplicate data across systems and applications — the customer’s registry in the billing systems and in the contact-centre management system, for example. Keeping this information up to date and consistent requires complex and multi-layered governance.

Updating the representation is fundamental, because customer data are dynamic: they change over time. Having an up-to-date and clean representation of the customer at every moment is essential. If a customer calls the contact centre to change their privacy consents, that variation must be immediately available in the customer’s representation to any other organisational unit that requests it.

A holistic vision means that the customer is not just a code, a list of purchased products, a table of contact data, a commercial value. The customer is an entity that evolves and takes shape progressively and constantly. This is the paradigm of digital profiling, and it can and must be applied in traditional contexts too.

The SCV is often confused with building a data warehouse, a raw-data lake or a Data Management Platform, which turns it into a project generally driven by IT. As said at the start, the SCV is a process that reorganises data around the customer to answer the business objectives of the company’s different areas. It is therefore fundamental that this process be designed for and with the business “users” — the people who will then use it, such as marketing, customer operations and sales.

STEP 2. Quantifying the Single Customer View

The SCV brings multiple benefits, some of them specific to the company’s sector or activity. In most cases the most important ones are enough to justify an SCV project, and they are relatively easy to quantify.

Improving the customer experience: having a single, up-to-date view of the customer means being able to manage them consistently across any touchpoint, channel and process. The customer will have the perception of being known, managed, remembered and considered unique. Picture for example a customer who calls the contact centre to get help with a faulty product and goes to the shop to return it. Obviously the customer expects the shop assistant to know everything already and to be ready to serve them at best. The principal value driver is increased satisfaction, with positive effects on loyalty and repeat purchase. Some KPIs that can be used are the Net Promoter Score and churn.

Improving customer profiling: aggregating a single view of the customer means having far more data to develop analytical models, more or less complex, to segment and to profile the customer’s journey descriptively and predictively, anticipating their wishes and offering a service consistent with their profile and preferences. The principal value driver is better performance of marketing campaigns and commercial activities. Some usable KPIs are campaign redemption and the conversion rate of the sales pipeline.

Delivering an omnichannel experience: having a single view of the customer is the only way to offer an omnichannel experience. It is anachronistic today to design one or more customer journeys and assume that the customer will follow what we have arranged. That kind of approach assumes that customers are segmented into a limited, predefined number of personas (customer stereotypes), each associated with a journey that assumes a given customer behaviour. What is delivered in that case is only a multichannel experience, because it is limited to the channels predefined in the journeys we thought of. The reality is that today the customer wants to use the channels they prefer, as they prefer, moving along a journey that is their journey. We speak of “segment of one”: every customer is to be understood as a single segment, a single persona, and must therefore be considered in their specificity, enabling all channels through the SCV. The principal value driver is the increase in interactions with the customer, which translates into cross/up-sell opportunities. Some usable KPIs are customer value (Average Revenue/Margin Per User — ARPU/MARPU), the average number of products or services purchased, and RFM indicators (Recency, Frequency, Monetary).

Improving attribution models: thanks to the punctual collection and tracking of all data relating to the customer, their contact history and their position in the “journey” can be reconstructed in relation to the marketing and sales activities carried out on them. This allows attribution models to be created or improved, to measure the business impact of those specific actions. The principal value driver is the return on marketing and sales investment. Some usable KPIs are the conversion rate between marketing budget and sales pipeline, or the ratio between revenue growth and growth in marketing investment.

Building a customer-centric culture: having a single, consistent view of the customer means that every organisation and person inside the company will have the same understanding of the customer and will act consistently. Moreover the vocabularies people use will begin to be mutually understandable, because they refer to the customer and to their univocal state. This not only reduces information asymmetries but makes it possible to orient the whole value chain to the customer — and therefore, in short, to the reference market. The value driver is less dispersion of corporate energy and greater focus, translating into higher productivity. Some usable KPIs are revenue or margin per employee, and the dead times in information-transfer workflows.

One last benefit is the ability to hold a single, certain custody of the customer’s profile and state, which is fundamental for complying with regulatory obligations such as the GDPR — General Data Protection Regulation — for checking the customer’s risk level or for intercepting fraudulent behaviour. An SCV holding clean, regulation-compliant, accurate and certain customer data is a key asset and the basis of any corporate governance. The principal value driver is the reduction of sanction risk and reputational damage.

STEP 3. The barriers to the Single Customer View

So if the SCV offers important benefits, why do so few companies actually deliver it? Better still, one could say that many companies invest in and start SCV projects, but few reach the end. Why? Let us try to understand.

Barrier #1 The first barrier is data fragmentation. It is normal for a company to have created, over time, different databases with different data models and taxonomies that in many cases contain the same entities under different names, with a concomitant proliferation of keys, duplications and inconsistencies. Think of warehouse, accounting, CRM, production and after-sales data.

Barrier #2 The second barrier is the layering of systems and applications. The fragmented databases feed different systems and applications, often introduced at different historical moments, by different business areas and suppliers. These systems and applications — even more than the data — are often totally disconnected. To deliver the SCV, systems and applications must be able to access customer data as they require it.

Barrier #3 The third barrier, which explains the failure of so many SCV projects, is the complexity of tackling the project by trying to solve barriers 1 and 2 in one go while also guaranteeing operational continuity. In other words, it would mean rebuilding the company’s entire data and application infrastructure from scratch, in parallel with the existing one, which must keep working until the new one is ready. This also entails an initial phase in which the company must invest almost twice the resources (capital and people).

Barrier #4 The fourth barrier consists of people. Since the SCV is used by, and matters to, every organisational area of a company, it is hard to establish who should be responsible for it. Almost always the mistake is made of thinking that the SCV belongs to marketing, or to CRM, or to IT. The result is a partial view, often useful only to those organisational units and not recognised by the rest of the company. The SCV must be relevant at every moment to the organisational unit that has to use it.

STEP 4. How to build the Single Customer View using Linked Data

First, it helps to understand what linked data are, as briefly explained in this video.

Video: What are linked data?

In short, Linked Data — invented by Tim Berners-Lee, the creator of the World Wide Web — are a very powerful way to aggregate data coming from any source, by creating links between them and making them consistent, that is, by eliminating duplicates and inconsistencies.

In other words exactly what the SCV needs!

Let us look more closely at how linked data can bring the barriers above down.

Solution #1 To deliver the SCV, data must not be fragmented. Using linked data it becomes possible to describe every single data set, attributing the relevant meaning to every piece of data, and to join them through links that simply connect the data on the basis of the meaning attributed. Consider for example the figure below, where the CRM dataset has been joined to the warehouse dataset with a simple link stating that product 1 bought by customer 1 is the same product 1x of the warehouse. Powerful!

Figure: the CRM dataset joined to the warehouse dataset, from the source article.

It is important to consider that it is not necessary to find all the links at once: they can be added little by little. Moreover, with linked data it is possible to identify and clean duplicates, errors and inconsistencies automatically. Joining the databases creates a so-called “Knowledge Base”, which is in fact nothing other than the Single Customer View!

Solution #2 The knowledge base will be created in parallel with any system or application present in the company, without requiring any change to them. In other words, while the knowledge base is being created, the company’s systems and applications can remain as they are and keep working, guaranteeing full operability. The resources needed to create the knowledge base are contained, and it can be built by aggregating one dataset at a time. Conversely, it will be the knowledge base that can be used by the applications or systems, whenever an SCV is needed. It is the knowledge base that takes care of:

• interpreting the data of the systems and applications in order to join them and make them consistent;
• translating the requests of the systems and applications and serving the answers in the correct format required by each of them, as represented in the figure below.

Figure: the knowledge base serving each system and application in its own format, from the source article.

Solution #3 For the reasons above, the advantage of linked data is the possibility of:

• aggregating data step by step, starting from the most urgent, important and immediately useful ones;
• operating in parallel with the company’s systems and applications without compromising their operability in any way.

This will allow the company to tackle the SCV project progressively and in an orderly way, involving the business areas concerned one at a time. The result is a project able to manage and resolve complexity by breaking it down into a path that uses resources progressively and makes it possible to collect immediate benefits. For example, one could start by bringing together marketing CRM data with contact-centre data, immediately putting the two organisational units in a position to use both sets and improving the view of the customer. Moreover, tackling an SCV project in this progressive way makes it possible to verify at every moment what is being done and to understand what is needed, changing the requirements and the corresponding activities in an Agile approach. Standard projects, after all, often require the a priori definition of a blueprint containing all the requirements, which almost never manage to be complete and final, creating in the course of things the need to introduce compromises and sub-optimal solutions.

Solution #4 For the SCV to be useful, it must be meaningful and natural for any organisation of the company. These organisations often use different languages, with vocabularies and taxonomies well known to them but not necessarily to the other units. Think of how many times, changing jobs, one has practically had to re-learn a language of terms and acronyms. To tackle an SCV project successfully it is therefore necessary to:

• decide the priority of the company’s organisational units that need the SCV, so as to start the project from the most important ones;
• identify their vocabularies and taxonomies (often referred to as an ontology);
• identify the data and the corresponding systems and applications;
• map the data according to the vocabularies and taxonomies identified.

This work not only resolves the barrier but produces a result of great importance for the company: the mapping of the taxonomies and vocabularies used by the different organisational units, and their “translators”.

This means a substantial improvement in the processes and in the working of the whole organisation, in a logic of customer orientation. Misunderstandings and information misalignments will for example be substantially reduced, if not eliminated, with the corresponding benefits.

Finally this approach makes it possible to use, with each counterpart, their own language, and to translate it into technological solutions that are not invasive for the company’s IT, which plays an essential role in the project. The objective is indeed that the governance and maintenance of the knowledge base be carried out by IT in companies that have one. Otherwise it is possible to use “turnkey” solutions that are totally managed, such as the one offered by LinkedData.Center.

STEP 5. The stages of a Single Customer View project

In conclusion, thanks to linked data a new and powerful approach to delivering the SCV is available today, one that overcomes the barriers which are often the cause of failure. The principal stages of an SCV project using linked data are:

1. Definition of the expected benefits and quantification of the SCV’s value;
2. Definition of the KPIs and monitoring metrics;
3. Identification of the organisational units involved and their priority;
4. Definition of the project governance and identification of a third party responsible for project management;
5. Mapping of the taxonomies, vocabularies, data, systems and applications of the organisational units involved, according to the priority defined;
6. Creation of the knowledge base, and joining, cleansing and mapping of the first two databases;
7. Immediate verification of the result and feeding of the systems and applications;
8. Possible revision of the requirements;
9. Addition of further databases to the knowledge base by repeating points 5–7;
10. Maintenance of the knowledge base.

The company will thus be able to proceed recursively and progressively, even at a distance in time, adding data to the knowledge base and improving the SCV step by step, always consistently with the business objective and without the constraints of the systems and applications in place.

LinkedData.Center has developed a set of solutions based on linked data that have made it possible to deliver all of the above for several companies, successfully.