During an April 2009 visit to the Martinsville branch of the Virginia Employment Commission, I watched as an anxious job seeker learned that the nearest job opening matched to his background was more than 100 miles away. Seeing as the local unemployment rate exceeded 20 percent, the finding itself was not all that surprising—maybe there just weren’t any good jobs to be had in the area. Unfortunately, the information he was given might have been wrong.

In an economy that provides real-time pricing data for bargain-hunting shoppers, it is unacceptable that we don’t have a similar method for matching job seekers’ skills with job-market openings. The problem of matching skills with jobs is particularly difficult for veterans transitioning to the civilian workforce. According to PayScale, an online salary profile database, veterans are 22 percent more likely than the rest of the U.S. workforce to say they are not working in a job that makes use of their education or training.

That finding was one of many resulting from a 2015 open-data collaboration of private and public stakeholders called Veterans Talent. (Disclosure: My firm, Hunch Analytics, is one of the collaborators.) The initiative pooled skills data from public and private sources and published it for researchers to demonstrate how we can improve the workforce-development system for veterans.

One of the research teams at the Corporate Executive Board, which describes itself as a “best practice insight and technology company,” studied Virginia-based job postings that had made a veteran hiring commitment as of April 2014. It mapped those vacancies alongside the skills profiles of veterans under 35 active on the Monster.com platform, and found that every entry-level technology job in Virginia could have been filled by unemployed veterans whose previous experience and skills had indicated that they were “tech-trainable.” But this likely didn’t happen. Instead, like ships passing in the night, too few veterans knew that they might qualify for such jobs and, worse, too many employer HR tools would weed them out even if they did apply.

We can do a better job of bringing such insights to the surface and putting them into action. State-based workforce-development systems are built on outdated technology and incomplete data on employer talent needs, job-seeker skills, and quality-training programs. As Virginia’s state auditor noted in a recent review, “No state agency currently compares the types of education and training offered in workforce development programs to indicators of demand from employers, such as job openings or projected employment growth.”

But there is an innovative way for states to meet these challenges without bearing the brunt of the needed investment that they can’t afford anyway: through the creation of a Virtual Workforce Connector, a web-based digital marketplace that connects various open labor-market data with online talent platforms such as LinkedIn or Monster.com. Coupled with new “pay for success” models that reward programs that generate the best outcomes—more on those below—states could unleash a wave of innovation to ensure that our economy works for everyone, while generating a net savings by avoiding workforce-development investments that don’t work. More broadly, such efforts can help realize an estimated 2 percent boost to global GDP by 2025, according to a 2015 McKinsey Global Institute study.

In Virginia, Governor Terry McAuliffe has taken the first step in such an effort by recently establishing the Commonwealth Consortium for Advanced Research and Statistics to offer the public—including platforms like LinkedIn and Monster.com—free access to “real-time data about human capital, regional skills gaps, local and state wage data, university research and talent, and availability of local and state workforce programs.” A Virtual Workforce Connector could build on this progress by empowering job seekers to choose from a growing number of online talent platforms, utilize high-quality training programs that can close any skills gaps, and enroll in government benefits that might help defray the costs.

How would this work? First, states would publish as much labor-market data as possible in an easily accessible form. Second, states would enable job seekers to use online platforms to access workforce-development programs, including unemployment insurance filings; this would be not unlike the way many Americans file their taxes through services such as TurboTax or H&R Block. Third, states would institute “pay for success” programs that help them more efficiently allocate scarce public workforce dollars and quickly identify which programs are working.

Implementing Connectors will require both technical and business-model standardization. On the technical front, we have already seen similar efforts work in comparable sectors. Green Button, for example, is a data standard that enables consumers to connect energy-efficiency apps to their energy-consumption data held by participating electric utilities. Building to such standards will make it easier for online talent platforms to participate.

On the business model front, requiring platforms to publish an enforceable code of conduct for connected applications to operate as a fiduciary—in other words, putting a system in place that looks out for the job seeker’s interest and not the needs of sponsors—will enable job seekers to use workforce-development services and still be appropriately protected.

A Virtual Workforce Connector is a policy intervention that is pro-growth and pro-government. It can help get citizens back to work faster by increasing access to new features such as recommendations on career pathways. With this program, job seekers like the one I saw in Martinsville could find good careers closer than 100 miles away.