Data scientist Hilary Mason wants to show you the (near) future

Marc Ferranti

How to formulate a data strategy

There are three basic questions about data every company should be asking themselves. The first is simply, "what data do we have?" This is not necessarily a simple task, especially for large business with multiple departments and silos of information. Fast Forward Labs has clients look at their products and write down the sort of data they collect. The process includes checking servers to confirm what data is stored.

The second question is, what data should the company have that it currently does not? Businesses should reflect on how the business operates, and the sort of questions that might provide answers that could generate growth. Companies should also look ahead and figure out how they would store new data they could collect, for example in Hadoop clusters, on site in internal servers, or on a platform like Amazon S3.

Finally, the Fast Forward Labs team asks companies about the assumptions they have made about their business that can be validated with data. Many companies have intuitions about how to penetrate markets and opportunities for their products. Data analysis can verify -- or challenge - such assumptions.

"I always find when you start with those things you probably already know, you will probably confirm most of what you know but will start to learn things that you didn't know," Mason says. "Once you start asking questions and getting useful answers the process grows -- once you've got a few good questions you will end up with many more."

Looking ahead, Mason says the path for Fast Forward Labs is not to simply add more staff, even though there are big market research companies, consultancies and vendors that are jumping on the machine intelligence and data science bandwagons.  "There's a great energy when you have a team that's less than 15 where you sort of know what everyone's up to and there's not a lot of bureaucracy," she says. So far the company business model has worked: After putting in some of her own money and not taking a salary for a while, Mason now pays herself and says the company is profitable.

Mason has tweaked the company business model a bit, offering options other than the annual subscription. Fast Forward Labs just made its natural language generation report available on a tiered-pricing basis, from $250 for a short, basic version, up to $5,000 for a version that includes the prototype and consulting time with the company.

Mason's ambition, however, is to rethink applied research on a broader scale -- how people bring new technical capabilities into organizations and how to find growth opportunities around those capabilities. "I believe there's something valuable in our model of applied research and advising that can be applied even outside of the world of strictly data -- but that's an experiment we will be doing in a couple of years."

Previous Page  1  2  3  4  5  Next Page