The Pilot Is Not the Point
Workforce development has no shortage of pilots. We know how to launch a cohort, test a partnership, introduce a new training model, and build an initiative around an emerging need. That experimentation is necessary. Industries change, technologies change, employer needs change, and the people entering our programs do not all arrive with the same experiences or needs. Some of the strongest programs I have been part of started with a question and a partner willing to give us enough room to find the answer.
A pilot is supposed to teach us something. It should help us understand what worked, what did not, what surprised us, and what we would design differently the next time. Eventually, it should also tell us what we have learned well enough that we no longer need to treat it as an experiment. The value of learning is not simply that we can describe it. The value is that it changes what happens next.
The RECESS Conference reinforced that for me in a way I did not expect. People from clean energy, workforce development, utilities, education, phila
nthropy, business, and community organizations brought different perspectives to the room, but many of the conversations returned to familiar lessons. Employers need people prepared for the work. Participants may need support beyond technical instruction. Contractors often need business and operational support in addition to understanding program requirements. Community organizations can create trust and connection, but they need real pathways to connect people to. None of that felt especially new, and that was precisely the point.
We have learned a great deal through implementation. The question is whether our next decisions reflect it. Research on the workforce system has made a similar case for directing more investment toward approaches with demonstrated results while preserving pilots for places where experimentation can actually generate new information. When we genuinely do not know the answer, testing is appropriate. When experience has already given us a clearer direction, repeating the same design should not be our default.1
That responsibility does not sit with funders or practitioners alone. Familiar models are easier to support and easier to deliver. Funders know the organizations, measures, a
nd reporting structures. Practitioners know the schedule, curriculum, staffing, and partnerships. Consistency can be valuable when a model is producing strong results, but familiarity should not become the reason a program remains unchanged. The better question is whether the model is still solving the problem it was designed to solve.
This is also where activity and outcomes have to be kept in perspective. Enrollment matters. Completion matters. Credentials matter. They tell us whether people are entering a program, staying with it, and building skills. But a full cohort can still lead to limited employment. A credential can be valuable without creating a clear path to a job. A contractor can attend every technical-assistance session and still leave without being better positioned to compete for work, grow revenue, or hire people. Those measures tell us part of the story, not necessarily the ending.
A U.S. Department of Labor meta-analysis of 46 career-pathway evaluations found substantial gains in educational progress and industry-specific employment, while gains in overall employment and short-term earnings were smaller and average effects on medium- and longer-term earnings were not meaningful. The finding does not diminish the value of training. It is a reminder that workforce outcomes unfold in layers, and that the measures closest to the program are not always the same as the outcomes people ultimately entered the program to achieve.2
Our own work has reinforced how much happens around the training itself. Employer engagement matters because people need somewhere for their skills to go. Participant support matters because transportation, family responsibilities, and financial pressure do not disappear when a cohort begins. Contractor development matters because technical knowledge alone does not necessarily prepare a business to compete, perform, and grow. We did not learn those things because a model was perfect on paper. We learned them through implementation, adjustment, and seeing where people or businesses needed something different.
That is why one discipline feels increasingly important: before the next round begins, be able to say what will be different because of the last one. Sometimes the answer w
ill be that the model worked and should continue. Sometimes recruitment, curriculum, employer involvement, participant support, or the definition of success needs to change. Sometimes an approach should be scaled. Sometimes it should end. All of those are legitimate outcomes of learning.
The same discipline applies when something consistently works. If a form of participant support improves persistence, it should not have to be rediscovered with every new pilot. If stronger employer engagement repeatedly improves the connection between training and opportunity, there com
es a point when it should simply become part of good program design. If contractors continue to show us that business and operational support matter alongside technical assistance, the lesson should eventually move from experiment to practice.
That matters in clean energy because the workforce challenge is moving quickly. The International Energy Agency has emphasized that the availability of skilled workers is increasingly important to the deployment of renewable energy and energy-efficiency technologies, while changing skill needs and barriers within education and training systems continue to put pressure on the pipeline. In a market moving this fast, repeatedly relearning what implementation has already made clear carries a cost.3
None of this means every pilot should produce an obvious answer or that workforce programs should work perfectly. People, employers, businesses, and economic conditions change. Some experiments will disappoint us, and some findings will create more questions than answers. A healthy field should have room for that. What matters is whether the experience improves the next decision.
That is ultimately what a pilot should give us. Its value is not simply that it existed, that the seats were filled, or that the grant period was completed. Its value is in what happens differently afterward. We should be willing to build on what is working, change what is not, and recognize when we already know enough to stop asking the same question. Otherwise, we may find ourselves becoming very good at piloting the same lesson, and the pilot was never the point.
