Timothy Taylor has a post on the Beveridge Curve, which plots employment rate against number of job openings.
To put it another way, employers in the years after 2009 seemed more reluctant to fill their job openings, or as economists say, it appeared to be harder for employers to find a match when they listed a job among the workers who were applying for those jobs. The "matching efficiency" of the US labor market had declined.
According to the article he references, this generally seems to have worsened since the great recession. This fits with a few trends around concentration of work - that demand (and pay!) is present and growing, but only given people that have the right skills, and location. There is also a question about how many of these openings are for badly paid or otherwise not-very-good jobs: leading to the rise in "joblessness" in some sectors where possibly folks could work, but not conditions that are significantly better than their alternatives. It would be interesting to see the data around what rate of jobs are filled with changing from one employer to another versus those joining the workforce or returning from unemployment.