January 27, 2020 | 9 min read

Stitch: the Brains Behind a Truly Comprehensive View of the Customer

To serve your customers, you need to make sense of all this data, across systems, day in and day out.

Graphic depicting identity resolution: connecting lines forming between white orbs

A room full of customer data

Hello folks who love their customers. Imagine you’re in a room with a very long table. The table is covered with index cards, and when you pick one up, you see what looks like customer information written casually across the front.

The card reads:

First name: Amanad Last name: Davidson Email: amwarren@gmail.com Phone: (262) 602-5877 Address: 565-66 Ho Estates Ave. City: North Samantha State: New York Postal: 11461 Birthdate: 1900-01-01

Intrigued, you pick up another card and see:

Name: AMANDA Last name: WARREN Email address: amwarren@batespaceandbrown.com Phone: +1-262-602-5877×172 Address: 5656 Ho Estates – Unit 66 City: North Samantha State: New Hampshire Zip code: 11461 Birthdate: 1982-04-09

Using your super-powered human brain, you immediately think that these two index cards probably represent the same real-world person. Sure, some of the fields are named differently and there are is a typo in the first name, but the email aliases match, as do the core phone number, city, and zip code. The differences in the information could easily be chalked up to lazy fingers on a keyboard, a change of last name due to a marriage, and someone too rushed to enter their real birth date. After a moment’s thought, you correctly determine that these two cards both represent the same person.

Now imagine you read and match every card. For some people, you would find 10, 15 or 30 cards. These represent your brand’s best customers. For others there are only 2 or 3 cards. These are new customers. At the end of an hour, you’ve used your human intuition to connect maybe 200 cards with 99.9% accuracy. You’re pretty proud of yourself.

As you study those piles of cards, you learn everything there is to know about those customers, based on every piece of information they’ve shared with you. You know who they are, what they buy, and how best to reach them on any channel. You know which customers are your best and highest value, and which are likely to churn. And you know how to inform your marketing and advertising decisions to deliver the highest returns.

Now imagine that someone comes in and dumps a hundred thousand new index cards on the table. Now you have to assess and match every one of those as well, not forgetting to also connect them to your existing piles of cards. How many hours would that take? And what if the next day you got another hundred thousand? And the next day another?

What sounds like a bad dream is what real companies face every day: the ever growing onslaught of customer data from new and existing sources and channels, fraught with typos, inconsistencies, data fragments, and no linking keys across systems. To serve their customers, they need to make sense of all this data, across systems, day in and day out.

We've taken steps at trying to unify our customer data and every single time the scope of that project was too big and then we then ended up having to scale it down. We had to scale down the scope so much just to get it done within two years that it basically became a communication preference center as opposed to an actual single view of the customer.

Now imagine you had a machine to take care of all of that for you. Introducing Stitch.

Stitch: human intuition at scale

Amperity’s Stitch uses AI to deliver human-like intuition at scale, finding the hidden connections across all of your customer records – eCommerce, point-of-sale, loyalty, email, clickstream, mobile and more – enabling you to build the richest, most accurate customer profiles possible, finally delivering on the coveted Customer 360 view.

Using a purpose-built, scalable infrastructure designed for massive computing power, what would take an army of humans months or years to complete is done in minutes. Better still, the whole process is refreshed daily, as new data enters the system or as you add new data sources (which only takes 1-2 weeks), so the results are always current.

Stitch leverages patent-pending algorithms to scour and understand your records at the feature-level, ultimately connecting all of your disparate data, across all of your data silos. With no requirements for joining keys, deterministic semantics, or heavy IT involvement on your end, we process your data with speed and intelligence, from applying relatively simple searches on common name patterns per geographic region to advanced sequence measurements like Levenshtein distance – and everything in between.

Along the way, we have reinvented what it means to intelligently match customer data, throwing out the antiquated concept of a binary matching score and designing a threshold scoring system instead. This allows us to deliver a much more complete and accurate picture of your customers and their data. This, combined with cluster-based matching and the retention of all source data, supports a durable solution that prevents your customer records from decaying over time.

There's no other solution on the market with customer data unification capabilities that come close to Amperity's. We're finding connections in our data where there are no unique identifiers. Without the power of Amperity's algorithms, those connections would remain hidden and the data would simply go to waste.

All of this data intelligence, computing scalability, system flexibility, and raw speed add up to a solution to a problem that until now has simply been out of reach for even the largest, most modern companies. And the best way to prove how Amperity solves the customer data unification problem is to allow our customers to visualize the process hands-on. For that, we provide an entirely new experience called Stitch.

Seeing is understanding: launch of Amperity’s Stitch experience

Amperity’s mission is to help people use data to serve the customer. Today, we’re launching a new Stitch experience that gives users an easy way to see and understand how we’re stitching their data together, so they can use it intelligently and effectively.

Stitch has always been the brains behind our AI-powered customer identity resolution and record matching. But until we launched the Stitch experience, it was hidden from view. Users didn’t see how their data was being processed until it was populating their rich Customer 360 Database.

We launched the new Stitch experience because we believe that for something as business-critical and nuanced as customer data – the foundation upon which you build and shape your organization – you need full visibility, flexibility, and understanding. Not just for yourself, but to help you educate and enroll others in how to think about your customers and their data.

Exploring Stitch

When you first open Stitch, we immediately give you a big picture snapshot to help you understand the size of your incoming data and the speed and freshness of Stitch.

The first thing you see in the Stitch experience

As you scroll down the page, you’ll get a deeper look at how your data is being matched. In the Overview section, shown below, you can see how many total customers Amperity’s algorithms found in your data (total Amperity IDs) and the overall Deduplication Rate for your data. In other words: how many records you started with and to what degree Amperity reconciled them, enabling you to better understand the overall connectivity of your source data.

Stitch overview shows the power of machine learning-driven customer record matching

When you click the “Explore Amperity IDs” button shown above, you can drill into sample customer clusters. A cluster is a group of records that all belong to the same individual, according to our algorithms.

This section of the product allows you to view the actual records that make up each cluster, along with the score we assigned to every pair. Scores are important because Amperity gives you the flexibility to match records more or less conservatively depending on your unique use cases.

Finding Amanda

A sample cluster for our fictitious customer named Amanda

Remember Amanda Warren, the fictitious customer with disconnected data on index cards at the top of this post? We ran her records through Amperity, along with many others, and our algorithms agreed with your human intuition and made some matches. Above is a cluster of her records, color-coded based on the strength of each match.

Here, platform user with PII-access can click on any link in a cluster and view the actual records that were matched, along with the Amperity score. Looking at a “high” match for Amanda, we see the actual records side by side. Notice the differences in the email address, last name, and other key data points. Amperity was able to discern that these are in fact matching records, albeit with a lower Amperity score than if the data matched more cleanly.

Two of Amanda’s records from eCommerce and Loyalty

If you want to view all the records in Amanda’s cluster simultaneously, there’s another view that lets you do just that, for the most complete picture of your data possible. This shows the magic of Stitch in a single screenshot. Records from across systems, with varying degrees of completeness and error-rates, are connected into a single cluster that tells a rich story of the unique customer. All keys from your source systems (Source IDs) are retained, existing alongside the unique Amperity ID that we assign, so you can truly understand who your customers are across your entire technology ecosystem. Note that this view is also only available to users with PII-access, which lets you keep your customer data secure.

All of Amanda’s records pairs with scores

Find out if Amperity is for you

Stitch is a critical portion of the Amperity platform, but it’s only a fraction of what we do. Amperity is an end-to-end customer data management solution that ingests raw data from any system, matches records using the intelligence of AI, builds comprehensive yet flexible Customer 360 views, derives customer insights, and lets you easily segment and send data to any downstream system.

All of this is done for one primary purpose: to help you use data to serve your customers. If you don’t know who someone is across all your systems, you can’t serve them. If you don’t have a comprehensive view of your customers, including all their transactions, preferences, and identifiers, you can’t personalize their experiences and target them with meaningful offers. If you can’t see trends in customer behavior based on an accurate and complete picture of your full customer base, you can’t optimize your initiatives across the business to drive long-term loyalty and growth.

Amperity is built to give you comprehensive, current, and connected views of your customers, for the first time. Many of the world’s most loved brands trust Amperity and use it to supply the customer data management foundation of their businesses. To see if our platform might be right for you, request a demo today.


Let us know what you think about Stitch: the Brains Behind a Truly Comprehensive View of the Customer on Twitter.