Customer Data Unification
Last updated May 22, 2024In the last phase, you brought together data sources from across your organization to create a single, live, trusted source of truth. Now you can use this data to create 360-degree, omnichannel profiles of your customers.
When you have complete customer profiles, you have a deeper understanding of the customer’s journey. For example, before customer data unification, you might identify five contact points:
- A person read your blog
- A person browsed your product pages
- A person visited your brick-and-mortar store
- A person made a purchase online
With omnichannel customer data, you can confirm that these five interactions were from the same person. You can then correctly attribute what led to the conversion, and deploy more relevant follow-up to continue the relationship.
In the next phase, we’ll use these unified customer views to derive insight and create a closed loop of feedback and nurturing.
Here’s how to get started with customer data unification.
Why Does Customer Data Unification Matter?
The customer journey is far more complex than it used to be, and it takes place across far more channels, too. Customers might hop on with a display ad, then continue through social media, your website, review sites, a coupon app, a local search, a brick-and-mortar visit and an online purchase.
Without the ability to combine physical and digital data in a single customer database, you wouldn’t see the totality of the journey. You might send the brick-and-mortar visitor a follow-up email, badgering them to buy something they just purchased from your online store. You might not attribute the display ad to the eventual purchase. You could end up tracking all of these touchpoints as separate individual encounters, rather than the journey of a single customer.
In short, customer data unification enables marketing that is more efficient, more relevant to the customer, and ultimately more effective in building relationships.
Thanks to customer data management, customer data unification and personalization at scale, Shiseido saw a 20% in-store revenue increase and a 38% growth in net income year-over-year.
Blasting emails to everyone who tried samples or bought a particular product won’t lead to customer delight. Detecting a mood swing in each customer and changing the tone of push notifications, does.
Kenji Yoshimoto
Chief Analyst for Direct Marketing
Shiseido
Identity Resolution: Creating a Single Customer ID
One major component of customer data unification is identity resolution. Identity resolution, or identity stitching, is the process of correctly attributing multiple data points to the same person, and assigning each profile a permanent customer ID.
Such a massive task would be all but impossible to do manually. Imagine having to evaluate thousands of touchpoints, with live data constantly adding more, and assigning them to the correct customer ID! Fortunately, this process can be virtually entirely automated with AI and machine learning.
CDPs use two types of matching to create permanent, persistent customer IDs:
- Deterministic matching unites customer records by searching for identifiers such as email, phone number, and usernames. This approach provides high confidence that the data points are directly related to the same customer.
- Probabilistic matching uses artificial intelligence to match through an estimate of the statistical likelihood that two identities are the same customer. The identifiers used could be things like IP address, device type, operating system or browser type. This type of match is less certain than deterministic matching, but is useful to expand reach or make up for limited first-party data.
With persistent customer IDs, you can finally recognize your customers as individuals, not just disconnected data points. You can use this information to:
- Improve the accuracy of your attribution models
- Optimize campaigns based on more complete data
- Increase marketing efficiency: Avoid sending ads for products customers have already bought, fine-tune your targeting, and deploy cross-channel frequency capping to avoid oversaturating your audience.
Employing a CDP is not the only way to unify customer data. You might try it with a CRM, or by manually collecting data in a data warehouse and manipulating it with business intelligence software. But a CDP is by far the most powerful, efficient and intelligent way to go about it.
CRMs and DMPs are limited in the types of data they can intake and process. CRMs are optimized for sales and marketing data, but are less-equipped to handle customer service data and data from offline interactions. DMPs aren’t configured to handle personally identifiable information. And manual processes can work for small businesses, but it’s tough to achieve any kind of scale processing data by hand.
In addition to the above, CDPs are built to handle the regulatory and privacy considerations that come from analyzing customer data. The right CDP gives confidence that you will stay on the right side of the law — and the right side of your customers’ trust in your brand.
Finally, it’s important to understand that customer data unification isn’t a one-time activity. It’s an ongoing process of updating profiles as new data comes in. Your customer profiles will continue to be enriched over time, bringing even more data to bear on your marketing efforts.
CDPs track anonymous users via Javascript and mobile tags, along with cookie or mobile device IDs. This data is stored alongside the persistent customer IDs and analyzed for matches. CDPs can even combine work and home email addresses with their master identity. When a customer opens emails to both addresses on the same device, the business and personal IDs can be stitched together into a known customer profile.
Once you set the rules for your CDP to follow in stitching together customer identities, it can continue to process billions of datapoints at a time to continually develop these profiles. These profiles are then available to your other marketing automation systems to drive personalized campaigns.
