A single customer view (SCV) draws together all the data a company has about any individual customer into one single page. It's thus a great asset for any customer-centred business as a company to have what our data science consulting team describe as joined-up data.
Most datasets are stored and then used for reporting and modeling for specific purposes, such as CRM, subscription services or datasets from apps. Joining up these datasets by creating a single customer view makes this data more potent and useful for any company, especially regarding driving KPIs and executing core business goals and strategies. Given the ferocious competition in the TV industry since Netflix and the OTT model broke the mold, we at Dativa state unequivocally that those companies who use TV data to know the most about their customers, powered partly by a single customer view, will emerge as the winners.
Benefits of a single customer view
Our data science consulting team has identified several specific business and marketing benefits to having a single customer view. These benefits begin with a more thorough and complete understanding of customers, which will then facilitate improved business planning and product development. The company can then share these better insights it has into customer behavior with advertisers. Both the business itself and the data which flows through it will become more efficient, especially concerning avoiding waste and misguided efforts.
The marketing benefits begin with insights, both detailed and comprehensive, into each customer. Assuming there is a large enough pool of customers, these insights will help customer acquisition too. However, a single customer view also has great value when it comes to organizing your marketing campaign across multiple platforms in a logical sequence. Perhaps a customer began using your website but now uses your Facebook page or a third-party app or simply a different device to access your website. What is important is the customer and not the platform they are using, and a customer who regularly switches between the platforms you offer should get the same smooth and uninterrupted marketing experience as someone who stays on the website, and this is much easier to achieve with a single customer view.
A single customer view also allows the company to offer a personalized marketing experience to each customer. This experience includes making the customer's journey as easy and interesting as possible by using the single customer view to tailor the marketing experience to individuals. Customers have many ways of conveying their needs and desires to a company if the company are willing to listen. If a male customer clearly dislikes sport but is still being treated as a sports fan - because, using demographic logic, all men like sport, right? - the company may be about to lose a customer. Perhaps the customer is actually fascinated by historical documentaries, as revealed by his viewing behavior, and will respond well to a marketing campaign full of details of all the historical documentaries the company are going to broadcast this winter. For us, the critical factor is that a customer should only have to reveal their preferences and how they want to engage with the companies and services once.
A single customer view is also known as a customer real-time propensity model. In propensity modeling, past behavior is used to predict likely future behavior by using machine learning. This can help create a great user experience, including recommendations, i.e., if you liked that, you'll probably like this too. Propensity modeling can also help to retain customers by spotting and then minimizing churn as well as trying to persuade the customer to make decisions which will be profitable for the business.
How to build a single customer view
Each customer needs a unique identifier which is used to identify them across all the databases and data collection points which the company has. If you don't have a separate, primary email address for each customer, deciding which identifier (or match fabric) to use can be tricky. An IP address is both unstable (some IP addresses change frequently) and would tie the customer into a particular location while cookies would tie a customer into a specific device. These identifiers may have worked a few years ago but not in 2018. At Dativa, our data strategy consultants encourage companies to make customers register to access their services; this is now so normal that they will lose few if any customers by making this demand.
There are many issues to be considered when building a single customer view. When and where will you be using the single customer view? How do you prioritize among the various data feeds you have about your customer, i.e., all the info will be on one page but which data comes first and which goes to the bottom? Are there any third-party sources you could be including but aren't? Are you updating your data sources? How long before a piece of data becomes out-dated? Is the quality of the existing data sufficient? Can it be improved? Asking these questions is essential for the data science team who are going to build a single customer view.
Data matching and single customer views
At the heart of creating a single customer view for each and every customer is data matching. At it's simplest, if there are two datasets and one has my email address and the other my television preferences, based on my previous television watching habits, then matching is linking these bits of data as about the same customer. While deterministic matching is by definition accurate, there is also probabilistic matching where two records are likely to match. In this post-GDPR world, arguably the most important thing to get right in data matching across multiple datasets is the privacy of any personally identifiable information, which your company can do through anonymization or pseudonymization.
Data matching is useful both to help companies learn more about their customers and to make ad inventory more valuable, e.g., if one dataset contains a list of email addresses on a mailing list and a second dataset which includes those customers who use one of the company's subscription-based apps. Data matching these two datasets might help sell ads to particular users if the app contains an advertising element. Data matching can also be combined with segmentation, i.e., selling to people based on segments rather than demographics. Acxiom and Experian are examples of information service provider companies which have built up large databases of users where they characterize the users with literally hundreds of behavioral segments based on purchasing habits and more general attitudes.