The email and notifications marketing automation firm Boomtrain has taken another funding round, this time a $12 million chunk from Sierra Ventures, Cota Capital, and Lerer Ventures.

Boomtrain uses machine learning to help businesses send customized emails and app notifications to customers. The company said it now holds 350 million user profiles and 7.3 trillion data points in its database, and it has targeted 1.35 billion messages.

The company said it will use the new funds to hire the engineers needed to expand its delivery channels beyond email, mobile, and web, saying it intends to allow for “full-circle recommendations across all channels and optimize the entire audience journey.”

Boomtrain currently has more than 100 customers using its platform, including CBS Interactive, The Guardian, and Thrillist Media Group. For instance, CBS Interactive uses the platform to tailor-fit the content of its newsletters to individual readers in different regions.

Boomtrain claimed that publisher and ecommerce clients using the platform regularly see “50 percent to 100 percent-plus increases” in engagement, including higher clickthrough-rates, page views, time on site, and conversions.

Founded in 2012, Boomtrain is backed by top-tier VC firms including Sierra Ventures, Cota Capital, Crosslink Capital, and Lerer Ventures.

The Boomtrain team at play.

Above: The Boomtrain team at play.

“We need an understanding of two buckets of data to do this,” Boomtrain CEO and cofounder Nick Edwards told VentureBeat in a phone interview Tuesday. “First we need to understand what the customer is reading, what they’re sharing on social, what they’ve purchased, and what they’ve commented on in social media.”

Edwards said the second bucket is filled with data about the client’s product. For instance, Boomtrain might conduct an audit of a client’s marketing content library, then “semantically read the content and categorize the topic models,” as Edwards put it.

The platform then uses what it knows about an individual customer to deliver them the most relevant of the content it’s crawled and analyzed. Much of the customer and content data comes to Boomtrain from the client company itself.

Boomtrain pulls in no third-party data sets to match consumers with content, Edwards said. It’s a privacy concern, part of the “don’t be creepy” maxim used at Boomtrain. For instance, Boomtrain wouldn’t use Facebook data about some photo a user shared last year, but it might use data about a user making a comment about the client’s product.

“It’s like when you go to your local coffee shop and the barista says ‘Is it a cappuccino day or a latte day?'” Edwards said. “That’s awesome because the barista is making a suggestion based on past experience with you.”

“But if the barista said, ‘Hey I was looking through your bedroom window this morning and saw you put on red polka dot underwear, so I’m guessing it’s probably a cappuccino day today,’ that’s just creepy,” Edwards said, likening this type of knowledge to third-party data.

Boomtrain was just a four-person shop until it raised its $2 million funding round in March of 2014. Then, the company started signing on customers and making hires. Today the startup has 100 paying “enterprise level” customers and has grown to more than 40 employees.

VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.