‘This yields more votes’: DNC touts sweeping change to voter data hub

The Democratic National Committee is unveiling a sweeping system-wide change to its central hub of voter information that party officials hope will make their campaigns more efficient at contacting and persuading voters in the November elections.

It’s the DNC’s latest step to refurbish the party’s data and technology operation, which officials say fell into disrepair during the 2016 election. Democrats have said that improving their data operation is necessary to compete with President Donald Trump’s campaign.

DNC officials say that in years past, its voter file — a vast compilation of information that includes details about nearly every voting-age citizen in the country — has been plagued by wrong or partial information. When voters moved, for example, their relocation would create a new file in the system but not erase the old one, causing headaches for campaigns in need of up-to-date information.

But a new algorithm, developed by DNC Data Science Director Jody Heck Wortman, seeks out, identifies, and merges the duplicate files, in what committee officials describe as a first-of-its-kind effort within the party. The new National Record Linkage system will affect about one-fifth of the total voter file, or about 60 million individual records.

“It’s generally not the kind of work that campaign can spin up in a few months,” Heck Wortman, who joined the committee two years ago, said in an interview. “It’s really more of a several months to years-long problem that can’t be solved in an election cycle.”

DNC officials say that although humans can recognize relatively easily when two separate files refer to the same person, the data file is far too vast to have them combing through it. And until now, they say, algorithms hadn’t yet been sophisticated enough to do it themselves.

“This is obviously a known issue for the voter file and it’s not like we just discovered it,” said Nellwyn Thomas, the DNC’s chief technology officer. “But it’s such a thorny problem and it requires so much specialized brain power and so much time and energy that no one has done it well.

“This level of exhaustive work is really pretty groundbreaking,” she added.

DNC officials hope that as a practical matter, the new algorithm will help the campaigns for de-facto presidential nominee Joe Biden and down-ballot candidates avoid time-consuming mistakes like reaching out to voters who have moved to a different location. And in the case of voters who have just moved, it will help make sure a campaign has a complete list of the issues and concerns that motivates them.

“Campaigns are resource constrained,” Thomas said. “And so, the more efficiency we can build in for the campaign, the more they can stretch their money and volunteer time to cover more people. Basically, this yields more votes.”

DNC Chairman Tom Perez has made retrofitting its data operation a priority since assuming control of the committee in 2017. Last year, the party unveiled a new data exchange operation led by former DNC Chairman Howard Dean that party leaders hoped would make the electoral information more current and accessible for its party’s campaigns.

Correction: DNC Data Science Director Jody Heck Wortman’s name was misspelled.

This story was originally published April 24, 2020 at 5:00 AM.