The United States spends approximately $3.6 billion annually through the Workforce Innovation and Opportunity Act alone. Add state programs, municipal initiatives, philanthropic grants, and private-sector training partnerships, and the total investment in workforce development easily exceeds $20 billion a year. Meanwhile, 87% of employers report difficulty filling open positions. Only 37% of people who complete a workforce training program are employed in their trained field one year later. These numbers don't describe a system that's failing. They describe a system that's working exactly as designed -- it's just designed to solve the wrong problem.
The American workforce system was built on a reasonable premise: if you train people in skills that the labor market needs, they'll get hired. Supply meets demand. The logic is clean. The problem is that the system's definition of "skills the labor market needs" is set by the people designing the training, not the people doing the hiring. And that gap -- between what training providers deliver and what employers actually require -- is where the entire system breaks down.
How the System Actually Works
Follow the money and you can see the misalignment at every handoff. Federal WIOA dollars flow from the Department of Labor to state workforce agencies. States distribute funds to local workforce development boards. Boards issue contracts to training providers -- community colleges, nonprofits, private training companies. Those providers design curricula, recruit participants, deliver instruction, and report outcomes back up the chain.
Here's where it goes wrong. Training providers design programs based on what they can deliver, not necessarily what employers need right now. A community college with an existing welding lab will offer a welding program regardless of whether the local labor market has open welding positions. A nonprofit with certified medical assistant instructors will run a CMA cohort whether or not the hospitals in the region are hiring CMAs. The incentive structure rewards program completion -- seats filled, certificates issued, credentials earned -- not employment outcomes in the trained field.
The employer, meanwhile, exists in a parallel universe. They have open positions with specific requirements. Those requirements change quarterly -- sometimes monthly -- based on contracts, technology adoption, regulatory shifts, and customer demand. But nobody from the workforce system is sitting in their operations meeting. Nobody is tracking their job orders in real time. The employer posts on Indeed. The training provider runs their cohort. If the two happen to align, someone gets hired. If they don't, both sides report the problem as a mystery.
Training providers design programs based on what they can deliver. Employers hire based on what they need. Nobody built the bridge between those two realities.
Where It Breaks: The Three Handoff Failures
Handoff 1: From labor market data to curriculum design. Workforce boards use Bureau of Labor Statistics data and regional labor market analyses to identify "in-demand occupations." This data is typically twelve to eighteen months old by the time it informs a training program. In a labor market that shifts quarterly, a program designed around last year's data is training people for last year's economy. Detroit's manufacturing sector is a clear example -- the jobs available in advanced manufacturing today require CNC programming, robotics maintenance, and quality systems knowledge that didn't appear in the labor market data from two years ago when current programs were designed.
Handoff 2: From training completion to job placement. Most workforce programs have a "placement" function, but it's typically a job developer with a list of employer contacts and a stack of resumes. The job developer calls employers, asks if they're hiring, and tries to match graduates to openings. This is a manual, relationship-dependent process that doesn't scale and doesn't adapt. In Cleveland, I've watched workforce programs graduate cohorts of twenty trained individuals into a placement pipeline that consists of one staff member and a spreadsheet. The training was excellent. The placement infrastructure was a phone and a prayer.
Handoff 3: From placement to retention. The workforce system measures placement -- did the person get a job within ninety days of completing the program? What it doesn't measure with any consistency is retention. Did they keep the job? Were they placed in a role that matched their training? Did the employer retain them past the probationary period? In Baltimore, a workforce program recently reported an 82% placement rate. When an independent evaluator looked at twelve-month retention, that number dropped to 31%. The system counted placement as success. The workers counted turnover as reality.
The Employer's Experience
Talk to employers -- not the ones on the workforce board, but the ones on the factory floor, in the warehouse, at the healthcare facility -- and you hear a consistent story. They need workers with specific, current skills. They need them on a timeline that matches their production schedule or patient volume or contract delivery date. And they need them to show up with not just technical skills but workplace readiness -- reliability, communication, problem-solving under pressure.
The workforce system delivers people with certificates. Certificates that may or may not map to the employer's actual requirements. A CNC machining certificate from a community college might cover manual programming on equipment the employer retired three years ago. A healthcare certification might not include the specific electronic health records system the hospital uses. The credential says the person is qualified. The employer's job requirements say they're not. Both are telling the truth from within their own system.
This is not a criticism of the workers or the training providers. It's a diagnosis of a system where the inputs are disconnected from the outputs. The training provider is measured on completions. The employer is measured on productivity. Nobody is measured on whether the two connect.
Placement without retention is waste. A 90-day metric that ignores 12-month outcomes is measuring activity, not impact.
Three Shifts That Would Change Everything
Shift 1: Training starts at the employer's door. Instead of designing curricula from labor market data and hoping it matches employer needs, invert the process. Start with the employer's actual job requirements -- the specific equipment, the specific software, the specific workflows -- and design training backward from there. This means training providers spending time on employer floors, not just in advisory board meetings. It means curricula that update quarterly, not annually. It means treating the employer as the customer, not just the destination.
In Detroit, a handful of programs have started doing this with advanced manufacturing employers. The training is co-designed with the operations managers who will supervise the new hires. The equipment in the training facility mirrors what's on the production floor. The result is that graduates walk into their first day already familiar with the environment. Placement rates are higher. Retention rates are dramatically higher. The model works. It just hasn't been adopted at scale because it requires training providers to give up control over curriculum design, and that's where institutional resistance lives.
Shift 2: Placement without retention is waste. The workforce system needs to measure what happens at twelve months, not just ninety days. A person who completes training, gets placed, and leaves within four months is not a success story -- for the worker or the employer. Retention should be a core metric, funded and tracked with the same rigor as enrollment and completion. This means workforce boards need to maintain relationships with employers and workers long after the initial placement. It means building feedback loops that identify why people leave and using that data to improve both the training and the placement process.
Shift 3: Data should flow between providers, employers, and boards in real time. The workforce system runs on quarterly reports and annual reviews. Employers make hiring decisions daily. The mismatch in information speed is a structural failure. If a major employer in a region adds fifty positions or changes its skill requirements, that information should reach training providers within weeks, not months. If a training cohort is graduating in thirty days, every relevant employer in the region should know it -- not because a job developer made phone calls, but because the system automatically surfaced the match.
The technology to do this exists. Customer relationship management systems, automated matching algorithms, real-time labor market feeds -- none of this is theoretical. What's missing is the institutional commitment to build shared data infrastructure across organizations that have historically operated as independent entities. Workforce boards, training providers, employers, and community organizations all hold pieces of the puzzle. Nobody has built the table where the pieces connect.
The Real Question
The workforce system isn't short on money. It isn't short on good intentions. It isn't even short on competent people. What it's short on is alignment -- the structural connections between what gets trained, what gets hired, and what gets retained. Fix the alignment, and the existing resources start producing dramatically different outcomes. Keep the misalignment, and another $3.6 billion will flow through the system next year producing the same results it produced last year.
This is a systems problem, not a people problem. And systems problems have systems solutions -- if someone is willing to redesign the wiring instead of just adding more programs on top of the ones that already aren't connecting. The resources exist. The need exists. The gap between them is architecture, and architecture can be rebuilt.
