Amperity, a machine learning-focused CDP, lately updated its platform with new AI capabilities that can combine disparate consumer data into unified profiles, said Chris Jones, SVP of product management, who joined Amperity from Microsoft last year.
Additionally, a new visualization tool and UI helps marketers analyze their data and extract insights, like customer affinity across channels or which channels produce the most loyal customers. In the past, you had to be quite technical and know how to run SQL queries if you wanted to mess around with the data through the Amperity platform, Jones said.
Running shoe brandBrooks, an early client of Amperity’s, has been using the AI capability – named Stitch – to get “a powerful snapshot” of its customer ecosystem, said Nicole Des Rosier, a senior ecommerce web and digital analyst at Brooks in charge of the Amperity deployment.
“At first glance, our CRM manager immediately focused in on the lack of overlap in certain areas of our business and identified multiple opportunities to further connect with runners,” Des Rosier said.
Brooks customers are multichannel, she said. They shop online, they take in-store running assessments, sign up for emails, attend events and follow Brooks on social media.
“Across each of these touch points, we had great information, but no way to bring it together so we could use it – no linking key or consistent email addresses,” Des Rosier said. “We needed a platform that could ingest all of this siloed data, make sense of it and let us use it across the business.”
Most vendors create preset deterministic rules that they use to compare discrete pieces of information and determine whether they’re related. A simple rule might be something like: If two customer records seem similar, compare the associated email addresses, and if they’re the same, join the records.
The idea behind Stitch is to create more complicated or “transitive connections” – for example, Record A connects to Record B and Record B connects to Record C, so Records A and C are probably correlated – the sort of linkages that would be difficult for a human to make at scale.
And for that, you need artificial intelligence, he said.
Amperity’s largest brand client, for example, has somewhere around 972 million different source IDs scattered across Wi-Fi connections, ecommerce, clickstream data, point of sale and the like, which Amperity resolves down to around 227 million customers on a daily basis.
Many of these connections would be fairly obvious to a human, “but a human isn’t able to look at and then reassess 972 million records every day,” Jones said.
From there, marketers can use the visualization tools to slice and dice their customer databases into segments, and then send the segments off to their downstream partners for activation. Amperity has integrations with Facebook, Google Ads, data warehouses Snowflake and Azure, Microsoft Dynamics, Listrak and others, and also supports custom integrations with a client’s existing systems.
Which is all well and good – but what about the marketing clouds nipping at the CDP category’s collective heels?
“What they’re doing validates what we’re doing,” Jones said. “But the problem we’re trying to solve is really hard and [CDP technology] isn’t just an add-on or a bolt-on, and that’s what I think other entrants in the market are doing – making bolt-ons.”