"Basically, I'm using data to help tell a good story—whatever it may be... I started looking around at the different sources of data that concerns issues readers have. It's the readers who drive this. And I've always been interested in numbers."
That quote comes from Todd Dills, a senior editor at Connectiv member Randall-Reilly. At the recent Neal Awards, he won for Best Data Journalist and Best Range of Work by a Single Author. At an editorial meeting that morning, he spoke about the difference data was making in their journalism—and equally important in their readership (truckers).
"We started doing a series of state law enforcement profiles; at the end of the day information like this gives the readers a better sense, a better reality," Dills said. "They all have preconceived notions of law enforcement departments" that may not be true. California's turned out to be very different from what readers there thought.
The key, as Dills said, is serving readers—and thus bringing in new revenue. "The big takeaway is that there's data gathered on everything our readers do," he said, adding that the way you use it shows your "priorities as a journalist. I have been able to essentially prove our readers right on some things and wrong on others."
In the past, Russell Perkins, founder and managing director of InfoCommerce Group, has spoken about creating data products. He lists five steps:
1. Identify the opportunity;
2. Validate it;
3. Collect the data;
4. Build the product;
5. Take the product to market.
"The best opportunity is a combination of two elements—organic need and leveraged development and deployment," Perkins said. "That opportunity bubbles up from the industry itself; ...find the pain point that needs to be solved."
The important thing, Perkins has said. is how you structure your data. It's about breaking it down into pieces. A company name and address, etc., in a Word file means nothing. Give it tags in an Excel file and you've got something. He gave an example of a frame wear company that acquired data from every eyewear maker in the world. That actually led to a new automated process of grinding lenses.
Then do not skip the validation step. "Talk to your market, get input from editorial, sales people, customer service," Perkins said. "Once you've identified the opportunities, talk one-on-one to potential customers. You'll talk about concept A, and they'll want concept B. The value to the market may be different from what you assume."
And understand how the data will be put to use. Prospects like being involved in the process. "The more you offer the better, but you don't need it all out of the gate," Perkins said. "Think minimum, viable product but be ready to iterate."
What also helped Dills is that Randall-Reilly has an infrastructure that supports the work he is doing.
Wrote Perkins in a blog post: "[Randall-Reilly] overlaid its basic audience data with a rich public domain database that allowed it to append truck ownership data. Suddenly, it went from offering modestly value truck driver contact information to hugely valuable market intelligence and targeting capabilities as it now knows the exact make, model and year of the trucks its subscribers operate."
So data journalism cannot emerge from thin air. It takes a commitment. "Is there public domain or even licensed data that you could overlay on your own audience database to create new high-value marketing opportunities?" Perkins asks. "All this added intelligence about your audience is also something you can leverage internally as well, getting smarter about who you are talking to and what content they engage with."