Google predicted that Auburn and Kansas would combine to take 55 3-pointers. They took 49.getty images

Google Cloud is not the typical NCAA corporate partner. It can’t hand out samples of its product at the fan fest. It doesn’t have a celebrity spokesman delivering catchy lines.

But in its second year of a multiyear arrangement to be an NCAA sponsor, Google Cloud already has embedded itself into college basketball and, specifically, the NCAA Tournament in ways that typical sponsors can’t.

The intent when Google Cloud signed on prior to last year’s March Madness was to create a unique two-pronged partnership that makes Google the official cloud provider to the NCAA. One component of the sponsorship is traditional media and branding. The other is technology-focused, with an emphasis on showing what Google Cloud can do with the NCAA’s data. Google’s analysts have studied 80 years of the NCAA’s statistical data to analyze March Madness trends and make predictions.

This year, Google Cloud is diving much deeper into the NCAA’s data to create a method for judging a team’s relative strength called NET — NCAA Evaluation Tool.

Google Cloud’s Eric Schmidt (not to be confused with the company’s former CEO of the same name) oversees the analytical side of the partnership. He said the analytics created for the NCAA provide examples of what Google Cloud can do for potential customers outside of the NCAA’s sphere to drive business.

“We’re not focused on driving revenue. Our main focus is to drive awareness,” said Schmidt, Google’s Cloud developer advocate for big data and data sciences. “By using the NCAA as a real-world case study, we can show businesses how this platform can work and grow over time.”

NET, the NCAA’s new method for determining college basketball’s best teams, is the most glaring example of how Google Cloud is contributing.

When the NCAA decided last year that the old analytical method for gauging a team’s strength, the Ratings Percentage Index (RPI), had outlived its usefulness, it turned to Google Cloud to help develop a new power index called NET that does a much deeper dive into efficiency stats and strength of opponents when judging a team’s competency.

The NCAA introduced it last fall and as the season progressed, NET became a talking point and source of information for college basketball analysts. It also became a data point used by the NCAA’s selection committee when selecting and seeding the field of 68 teams. Of course, like the RPI before it, NET became a source of debate over how effective it is at ranking teams.

“The things that you can come up with from the NCAA’s data is almost limitless,” Schmidt said.

Dan Gavitt, the NCAA’s senior vice president of basketball who oversees the tournament, described NET as a “modern analytical tool. It can be used for specific analytics or for fun data for fans,” he said.

For example, if the NCAA wanted to know what type of play would most likely trigger a shift in momentum during a game, Google Cloud could go through years of data to determine if a dunk or a steal or some other type of play is more likely to start a scoring run.

“We’ve provided Google with access to several years’ worth of NCAA basketball data, which gives them numerous data points that allow them to study plays, the sequence of plays and the timing of plays in different game situations,” Gavitt said. “That enables the NCAA to draw some conclusions based on past game performances.”

Once Google took possession of the basketball data, drawn from 80 years of box scores and play-by-play sheets, it became Schmidt’s job to figure out what to do with it and how to give it value for the NCAA and media partners Turner and CBS, who occasionally reference trends during broadcasts.

For the tournament, Google Cloud brought in 30 college students who are aspiring analysts and charged them with finding interesting trends and other predictive analysis, which is posted at g.co/marchmadness. The students from all over the country were selected from a hackathon event Google hosted at MIT in January. Among them are majors in chemistry, engineering, data science and machine learning. What they have in common, Schmidt said, is a passion for basketball.

Not only are they providing Google Cloud and the NCAA with high-level data analysis, Google is identifying future employees.

“Is a team performing above or below their season averages, based on certain rotations and lineups?” Schmidt said, citing another example. “We looked at competitiveness. If a team is favored by 12, how well are they playing relative to the expectations and the competition?”

In other words, what data will help determine whether a score is more reflective of the winning team playing well compared to the losing team playing poorly?

The NCAA works with Genius Sports, a data and technology firm, to deliver the stats real time to Google Cloud. The software, called LiveStats, became the go-to method for keeping stats and moving them into the cloud this season. For the first time, media and analytics entities have access to data that tracks the location of every shot, foul and turnover.

Among the predictions on the Google site last week that came from some of the students:

n Kansas and Auburn will combine to attempt 55 3-pointers in their second-round game. The final tally for the two teams was 49.

n Gonzaga and Baylor will combine for 22 offensive rebounds. They actually had 25.

As the NCAA compiles the feedback from Google Cloud, it can “answer questions related to how the game has been played and progressed in terms of style of play, decision-making and how teams handle certain situations with regards to strategy. It also shows us how successful teams react in various situations,” Gavitt said. “We’re just scratching the surface of what this could be in the future.”