The United States spends $3.6 billion annually on the Workforce Innovation and Opportunity Act. Add in state-level programs, community college workforce tracks, nonprofit training initiatives, and corporate-sponsored bootcamps, and the number climbs significantly higher. And yet, according to the Government Accountability Office, only 37% of people who complete federally funded workforce training are employed in their trained field one year later. Meanwhile, there are 8.8 million open positions nationally, per the Bureau of Labor Statistics' JOLTS data. The training is happening. The jobs exist. The connection between them does not.
This is not a criticism of training programs themselves. Many of them are well-designed, staffed by dedicated people, and genuinely effective at teaching skills. The problem is that a training program is one node in a chain that requires at least six. A career doesn't begin when someone learns a skill. It begins when that skill connects to a credential that an employer recognizes, which connects to a placement process that actually reaches the right employer, which connects to an onboarding system that retains the person past ninety days, which connects to advancement pathways that keep them growing. Most workforce programs handle steps one and two. The other four have no owner.
The Six-Link Chain
Think about what actually has to happen for a person to go from unemployed or underemployed to building a stable career. There is a sequence, and every link in it matters.
First, skill development. The person needs to learn something market-relevant. This is where most programs focus their energy, and many do it well. Welding, CNC machining, healthcare certifications, CDL training, IT fundamentals -- the catalog of available programs is extensive.
Second, credentialing. The skill needs to translate into something an employer can verify. This means industry-recognized certifications, not just certificates of completion from a program most hiring managers have never heard of. There is a meaningful difference between a credential that opens doors and a piece of paper that sits in a folder.
Third, placement. The trained individual needs to be connected to an actual employer who is hiring for that specific skill, in a geography they can reach, at a wage that justifies the transition. This is where the system starts breaking down in most cities. The training provider and the employer exist in parallel universes. The workforce board might have a job matching portal, but the employer posts on Indeed and the candidate applies on their phone. Nobody is brokering the connection in a way that accounts for timing, location, and fit.
Fourth, onboarding. Getting hired is not the same as being retained. The first ninety days determine whether someone stays. If the employer has no structured onboarding, no mentorship, no clear expectations -- the person leaves or gets let go. The training program counts them as a successful placement. The employer counts them as a failed hire. Both are right, and the system produced both outcomes.
Fifth, retention. After onboarding, can the person sustain the role? Do they have reliable transportation? Childcare? Housing stability? If any of these fall apart, the job falls apart, regardless of how good the training was. Retention is not a workforce problem -- it is a systems problem that workforce programs are not designed to solve.
Sixth, advancement. A career is not a job. It is a trajectory. If the person has no path from entry-level to something better, they are not building a career -- they are occupying a position until something pulls them away. Without advancement infrastructure, training programs are producing temporary outcomes and calling them permanent ones.
A training program that ends at placement is like a hospital that discharges patients without follow-up care. The intervention happened. The outcome is still uncertain.
Where the Chain Breaks
In Detroit, I have watched this pattern play out across multiple industries. A training program graduates forty people in advanced manufacturing. Twenty-five get interviews. Fifteen get hired. At the six-month mark, eight are still employed. At twelve months, five. The program reports a 62% placement rate, which is technically true on the day the metric was captured. But the number that matters -- the one nobody tracks consistently -- is how many people are still employed, still advancing, still building something eighteen months later.
Cleveland has a similar pattern. The city has invested heavily in healthcare workforce pipelines, which makes sense given the concentration of hospital systems. But the training programs produce certified nursing assistants at a rate that exceeds the onboarding capacity of the hospitals themselves. There is a bottleneck between completing certification and starting a position that can stretch weeks or months. During that gap, people take other jobs, lose momentum, or face financial pressures that redirect them entirely. The pipeline exists. The flow through it is inconsistent.
In Baltimore, the challenge is different but the structural failure is the same. Workforce programs there have struggled with a credential recognition problem. Training providers issue certificates from programs that local employers do not weight heavily in their hiring decisions. The skill was taught. The credential was earned. The employer does not consider it sufficient. Nobody mapped backward from what employers actually require to what the training program delivers. The system produced a qualified person and an unsatisfied employer simultaneously.
The Measurement Problem
Part of what keeps this system stuck is how we measure success. Federal workforce programs are typically evaluated on completion rates and initial placement rates. Those are the numbers that determine future funding. So programs optimize for completions and placements -- which are the easiest parts of the chain to control. They are also the least meaningful indicators of whether someone's life actually changed.
If we measured workforce programs the way we measure business investments -- by sustained return over time -- the picture would look very different. An investor does not celebrate a company that has revenue in month one and is bankrupt by month twelve. But that is essentially what we do with workforce outcomes. We celebrate the placement and ignore the trajectory.
The question that should drive workforce investment is not "did they complete the program?" It is not even "did they get a job?" The question is: are they still employed eighteen months later? Are they earning more than they were before? Do they have a path to the next role? Those are career indicators. Everything else is activity measurement.
We measure workforce programs by completions and placements. We should be measuring them by what happens in month eighteen.
What a Functional System Looks Like
A career system -- not a training program, but a system -- would work differently. It would start by mapping backward from employer demand. What positions are open today, what will be open in eighteen months, what credentials do those positions require, what does onboarding look like, and what retention support does the employer provide? That mapping determines what training gets built, not the other way around.
The training itself would be co-designed with employers, not handed to them after the fact. Credentialing would be aligned to what hiring managers actually screen for. Placement would be active and brokered, not passive and portal-based. Onboarding support would extend past the first day. Retention would be tracked as a shared metric between the training provider, the employer, and the workforce board. And advancement pathways would be visible from day one, so the person entering the system can see where the road goes.
None of this is theoretical. Elements of it exist in scattered pockets across the country. Some community colleges have employer-embedded training. Some workforce boards track twelve-month outcomes. Some employers invest heavily in onboarding and retention. But these are exceptions, not systems. They work because someone locally decided to build the connections. They don't scale because the infrastructure to support them doesn't exist.
Where This Goes Next
The workforce conversation in this country is stuck at the training level. More programs, more certifications, more bootcamps. That is step one of a six-step process being treated as the whole solution. The cities that will build functional workforce ecosystems are the ones that stop measuring completions and start measuring careers. That means tracking people over time, holding multiple institutions accountable for shared outcomes, and designing systems that carry a person from skill acquisition to career trajectory.
The question is not whether we can train people. We are good at that. The question is whether we can build the connective infrastructure that turns training into employment, employment into stability, and stability into growth. That is a systems problem. And it will take systems thinking to solve it.
