Nmedia - Fotolia

MIT panel shares tips on how to get into the data business

Data wrapping -- in this case, bundling data and analytics services with products -- may entice more companies to become data businesses. A panel at an MIT symposium considered some best practices for doing so.

Growing interest in the value of data is leading more organizations to consider going into the data business. But...

they will have to make special preparations for the new undertaking, according to panelists who discussed the matter at the 2016 MIT Sloan CIO Symposium in Cambridge, Mass.

Those about to make the jump into so-called data wrapping -- which, in this sense, is a phenomenon that sees companies reselling data and analytics that enrich their core products and services -- should first look at the practices of veteran data services companies, according to Barbara Wixom, an MIT researcher who led the panel discussion.

"Things are changing in this digital world. We're starting to look at data as a part of our products," said Wixom, a principal research scientist at the MIT Sloan School of Management's Center for Information Systems Research. As that happens more broadly and businesses increasingly adopt data monetization strategies, it's worthwhile, she added, to "learn from the companies that are really good at doing what we need to do."

What new players may find is wrapping data around existing products for delivery to customers is no easy task. Panelists from the world of data services said going the product route requires top in-house data management skills. New types of sales skills often are needed, too. Also necessary is an understanding of the various levels at which customers will want to work with data.

Start with a data model

"The first thing is to model your data to make it consistent. At the same time, you have to think about the abstraction level at which you will deliver the data," said James Powell, a panel participant and CTO at ratings agency and marketing research firm The Nielsen Company in New York.

For some, he suggested, modeling data for use by people outside one's organization may require a different mindset. What might have been clear enough to use internally will need to be made more clear when it starts to be used by others, he said. For Powell, creating an adequate structure around unstructured data is part of that process.

"We do a lot of careful modeling, because we have a lot of specific needs. If you don't structure the data, it becomes too hard to analyze effectively," he said.

Powell suggested that would-be data wrappers will find their new data offerings will become embedded in the structure of businesses to which they sell. That can be a problem if the data service isn't flexibly designed with APIs that can reflect the different levels at which customers will want to work with the data, and the changes that have to be made as tools are updated.

It's possible to design data products "for the lowest common denominator" in order to cover a broad set of uses, according to Powell, but that can diminish the value of the product for niche data businesses. "How you design that API can radically change the ease of use of the data," Powell said after the conclusion of the panel discussion.

Skills are us -- or need to be

Overall company skills may need to be upgraded when making the leap to being a data business -- and the press to gain such skills is on in full in many organizations, as seen by MIT researcher Wixom. "Talent is the elephant in the room," she said. "We all know there is a data-skills shortage out there."

Panelist Mona Vernon concurred. Vernon, a vice president at the Thomson Reuters Labs unit of New York-based media and information services company Thomson Reuters, said the situation calls for ingenuity in hiring and organizing teams.

A lot of the time, data management is the hardest part of analytics.
Mona Vernonvice president at Thomson Reuters Labs

"A lot of the time, data management is the hardest part of analytics. You don't want to hire an experienced machine learning hand and have them spend six months trying to chase someone that can get a database. So, you have to think about skills creatively," Vernon said.

When going beyond basic data wrapping to selling algorithmic analytics tools, additional attention is required, she said. That's because the analytics algorithm being sold becomes a part of customer processes, and the buyer and the seller need to be clear on where the business risk associated with using it lies.

"The competency you need is to be data-literate in the product marketing and management," Vernon told session attendees. "You have to know the implications, even legal implications, of the algorithm for users."

Data at your service

Newbies to the data business will find preparation is generally required to succeed at selling new data and analytics packages, according to Ivan Matviak, an executive vice president at State Street Corp. The Boston-based banking company has expanded its original securities finance and investment management services to include a data as a service platform, risk and trading analytics services, and other new offerings.

Matviak said some of the new data services had previously been offered in different forms as part of financial services bundles, so enhancing and unbundling them was challenging, especially for the sales team.

"We underestimated how difficult it would be to sell these products," Matviak said. "Costs were different. The salespeople we had didn't have experience with that, so we had to make an honest effort to build that up."

Clearly, many steps are required to turn the dream of new data revenue into reality. Companies that join the data-business fray may expect more competition, too. In late 2014, forecaster IDC estimated 70% of large organizations had already purchased external data, and 100% will do so by 2019. In this light, IDC expects more organizations will begin to sell data or offer value-added content also aimed at monetizing data.

Next Steps

Learn about ways to connect data and business value

Find out about transformative tech as uncovered at a 2014 MIT CIO symposium

Read more MIT Sloan symposium coverage on SearchCIO

Dig Deeper on Big data management