Adaptive Learning Capacity: Learning Only Matters If Something Can Change
This is part 4 of Adaptive Learning Capacity, a five-part series on how organizations learn while the work is happening. New to the series? Start with part 1.
At the end of the last post, the team at Linden Pathways left a meeting with three explanations for why the people they were built to serve were dropping out. Maybe participants faced more barriers than before. Maybe referrals had shifted toward people who needed less help. Maybe the schedule no longer fits the lives of people working night shifts. They didn’t know which was right, but they knew what they needed to find out.
Over the next few weeks, they started finding out. Pulling referral data by source took Marisol an afternoon, and it showed that referrals from the family shelter had fallen by about half over 18 months. Denise called six former participants who had left mid-program. Four of them brought up the schedule without being asked. Childcare and transportation came up too, but mostly as reasons the schedule was so hard to work around.
So the picture was getting clearer. The harder question was about what to do about it.
Adapting is the third move in adaptive learning capacity, after noticing and making meaning, and it’s where I see most organizations stall. A team can notice something important, make real sense of it, and still find that nothing changes. Sometimes that’s because the insight wasn’t strong enough. More often, it’s because the organization has no clear way to turn what it understands into a decision.
Start with the decision you’re facing
One thing helped Linden more than they expected. In January, they would need to submit next year’s program plan to the workforce board, including the cohort schedule. That deadline gave the whole inquiry a shape.
Without a decision on the horizon, a question like “why are people leaving?” can go in a hundred interesting directions. With one, the questions get sharper. What would we need to know to change the schedule with confidence? Which parts of the program need to stay the same, and which could flex? For whom is the current model working, and for whom isn’t it?
Not every learning question needs to connect to an immediate decision. Some of the most important ones stay open for years. Still, ask early and often what decisions your learning could inform, because learning that never meets a decision tends to pile up without changing much.
How much evidence is enough?
When the program director suggested piloting an evening cohort, the executive director hesitated. Six phone calls and a drop in referrals felt thin. She wanted more data before changing something the workforce board had signed off on.
That instinct is a good one, and it deserves respect. Most organizations don’t want to redesign a program every time someone tells a compelling story, and for good reason. A useful rule of thumb is to match your response to the strength of what you know. A single observation is something to notice. A pattern you keep running into is something to explore. Consistent evidence from more than one source might be enough to act on. And the bar should rise with the size and permanence of the decision: rewording an outreach flyer needs far less confidence than restructuring a program.
The program director raised a different question, though, and I think it’s the one most organizations never ask. What’s our evidence that the current schedule is working for people we’re here to serve?
We tend to ask for more and more proof before changing something and very little before continuing it. Staying the course feels like the neutral choice, but it’s a decision too, and it rests on its own assumptions. Linden’s Thursday schedule was set years earlier, when most participants worked day jobs. Nobody had checked whether that was still true.
There’s another thing a simple evidence rule can miss. Waiting has a cost, and it doesn’t fall evenly. While Linden gathered more data, another cohort would start on the same schedule, and the shift workers in it would face the same choices the last group did. And the people whose experience was hardest to see, like single parents on rotating shifts, were exactly the ones who showed up as one or two voices rather than a clear pattern. When you’re deciding how much evidence is enough, consider whose experience a signal represents and who bears the cost while you wait, along with how many times you’ve heard it.
None of this means changing course every time something new comes up. An organization that reacts to every signal ends up exhausted and scattered. What you’re aiming for is judgment: knowing when to stay the course, when to look more closely, and when the evidence is telling you something needs to change.
Adaptation is probably already happening
Around this time, the program director learned something that surprised her. For most of the year, Denise had been running Saturday morning phone check-ins on her own for participants who missed Thursdays, walking them through what they’d missed so they could keep up. Three of the shift workers who did finish had leaned on those calls heavily.
Denise hadn’t mentioned it because it didn’t seem like a big deal. She was just doing her job.
I see some version of this in almost every organization I work with. Frontline staff are adapting all the time, finding workarounds and bending the model to fit the people in front of them. Most of it never reaches the people making decisions, so the organization can’t learn from it. Denise’s Saturday calls were, in effect, a small experiment that had been running for months. Nobody had thought to ask how it was going.
If you want to know where your organization is already adapting, ask your frontline staff what they do that isn’t in the program manual. You may find some of your best evidence there.
In a way, Denise had been doing all three parts of adaptive learning on her own. She noticed the missed Thursdays, made sense of why they were happening, and responded with the Saturday calls. What was missing was the organization doing the same. That’s a good picture of what it looks like when skill is present and capacity isn’t.
Sometimes the right response is a test
Linden’s options weren’t limited to keeping the schedule or overhauling it. There was a lot of room in between.
They decided to run the next cohort with two tracks, the usual Thursday sessions and a Tuesday evening option with the some content, and to make Denise’s Saturday check-ins an official part of the program. They also reached out to the family shelter to find out why referrals had dropped.
The question they were asking had shifted from are we sure this is the right answer? to what’s the smallest responsible step that would help us learn whether we’re right? The evening track also helped tell their explanations apart. If schedule was the main issue, shift workers on the evening track should stay at noticeably higher rates. If it was mostly other barriers, the difference would be small, and they’d know to look elsewhere.
Who gets to decide?
When the program director first raised the evening track, she assumed the executive director would need to approve it. The executive director assumed the workforce board contract locked in the schedule. Nobody had checked. When someone finally read the contract, it specified total instructional hours and said nothing about which days.
This happens more often than you’d think. An organization can learn something important and still not act, because nobody is sure who’s allowed to make the call, or everyone assumes someone else is. If information lives in one place and the authority to act on it lives in another, with no clear path between them, learning gets stuck in the gap. So ask plainly: When we learn something important, who can actually do something about it?
Nothing here was about anyone’s skill. The team had noticed, made sense of what they saw, and knew what they wanted to try. What almost stopped them was a condition of the organization, and conditions like this one are exactly what adaptive learning capacity is made of.
The conversation with the funder
The executive director’s biggest worry was still ahead. If Linden did a better job of reaching and keeping the people it was designed for, its completion rate would probably dip, at least for a while. Completion was the number the workforce board watched most closely.
This tension is real, and I don’t want to wave it away. Grant-funded organizations feel accountable for delivering what they proposed, and changing course can feel risky. It helps to separate fidelity to an outcome from fidelity to a particular activity. Linden had promised to help adults with significant barriers get into healthcare careers. Holding onto a schedule that pushed those adults out wasn’t keeping that promise, even if it kept the completion numbers steady.
When the executive direct met with their program officer, she framed it simply: here’s what we’ve been noticing, here’s why we think this change better serves the outcome we’re all working toward, and here’s how we’ll know whether it’s working. The program officer was more receptive than she expected. Two other grantees had raised similar concerns that year, and the workforce board had started wondering about it too.
Not every funder conversation goes that well. Still, many funders would rather hear about a thoughtful adjustment than find out later that a program drifted away from the people it was meant to serve.
What are you willing to stop?
The conversation with the family shelter turned up one more thing. Two years earlier, trying to improve completion rates, Linden had added a short “readiness assessment” to intake. Shelter staff noticed that their clients rarely passed it, so they’d mostly stopped referring people.
The assessment had worked, in a sense. Completion went up. It had also screened out many of the people Linden was created to serve.
Stopping something is often the hardest kind of change. The readiness assessment was somebody’s good idea, it had produced results, and it was tied to numbers the whole organization was proud of. Deciding to drop it meant accepting that an earlier fix had caused part of the problem. That’s hard for any team, and go easy on yourselves about it. Every organization has a few practices that made sense when they started and outlived their purpose.
Changes like this come with trade-offs, and name them before you act. Who is likely to benefit? Could the change create new barriers for someone else? What are you giving up? And do you have the capacity to do it well? For Linden, that last question was about Denise and Marisol, whose caseloads would get heavier with more participants who needed more support. A good change with no one to carry it out can leave things worse than before.
A decision starts the next round of learning
Once a decision is made, most of us slip back into execution mode, and that’s understandable. From a learning perspective, though, a decision is where the next round begins.
Before launching the evening track, Linden wrote down what they expected to happen, how they’d know whether it was working, and when they would come back to it (after the first eight weeks, and again at the end of the cohort). They also wrote down the questions they were still unsure about, including whether the Saturday check-ins would hold up once they became official and more participants used them.
That short page of notes turned the decision into something they could learn from, rather than something they’d forget they had decided.
Putting this to work
Your decision point might look different from Linden’s January program plan. At a foundation, it might be a strategy refresh or the next grant cycle. In a coalition, it could be the annual workplan. In a small organization, it might simply be the next cohort, the next event, or the next budget. Whatever it is, a few questions can help you respond with more confidence.
Before you decide, ask:
What’s our evidence that the current approach is working, especially for the people it’s meant to serve?
Who bears the cost if we wait for more certainty?
Who actually has the authority to make this call? (check, rather than assume)
What are staff already doing differently that we haven’t asked about?
If a funder needs to know, you can keep it simple: Here’s what we’ve been noticing. Here’s why we think this change better serves the outcome we share. Here’s how we’ll know whether it’s working.
After you decide, write a short note that answers: What do you expect to happen? How will we know? When will we come back to this? What are we still unsure about? Half a page is plenty and it turns the decision into something you can learn from.
And the next time your team finishes a conversation about what you’re learning, try adding one more question before everyone leaves: What, if anything, does this change? The answer might be a decision, a small test, a need for more evidence, or more confidence in what you’re already doing. If the answer is “nothing” over and over, it may be worth looking at what’s standing between your learning and your decisions.
The Adaptive Learning Capacity series:
Learning Only Matters If Something Can Change (you are here)
Building the Habits That Make Learning Last (coming 11/3)
