Leaving No One Behind: Making Your Learning Systems More Inclusive

What Happens When No One Learns?

Years ago, I worked on an evaluation for a community program designed to support low-income women entrepreneurs. The data looked promising—participation rates were high, and many women reported feeling more confident in business. The final report was glowing, full of numbers that seemed to prove the program’s success.

Months later, I had the chance to visit one of the communities again. I spoke with a few of the women who had attended the program, and their experiences painted a different picture. Many struggled to access the financing they were promised. Some had attended the sessions but couldn’t apply what they learned because of ongoing childcare and transportation challenges. A few had even dropped out early because the training materials weren’t in their native language.

None of this made it into the final report.

That’s when it hit me—the programme had been evaluated, but it hadn’t been learned from.

This is where many Monitoring, Evaluation, and Learning (MEL) systems go wrong. The L—Learning—should be the part that helps organisations reflect, adapt, and improve. But when learning fails to capture the realities of marginalised communities, we end up celebrating programs that may not be working as well as we think.

Here’s why this keeps happening—and what we can do about it.

1. The Data Looks Good, But What’s Missing?

Many MEL systems focus on outputs:

• How many people participated?
• What percentage reported increased knowledge?
• Did the project meet its targets?

These numbers can give a surface-level picture of success but often fail to capture who is still being left behind. If 50 percent of participants are women, but none of them feel empowered to actually start a business, what did we really achieve?

What to Do Instead

Ask different questions

Instead of just tracking participation, explore barriers. Who dropped out? Why? Who benefited the most, and who struggled?

Use diverse learning methods

Stories, focus groups, and follow-up interviews often reveal what numbers alone cannot.

Challenge success narratives

If the data looks too good, ask: Who wasn’t included in the story?

2. Who Controls the Learning?

In many MEL systems, learning is something that happens at the top—between donors, programme managers, and evaluation teams. The communities being studied often never see the insights gathered from them.

This creates a power imbalance: those closest to the problem have the least control over the learning process.

What to Do Instead

Bring learning back to the community

Instead of just producing reports for funders, create feedback loops where community members can engage with findings.

Involve marginalised voices in reflection

Don’t just collect data from them—give them a role in interpreting what it means.

Use accessible learning formats

Not everyone will read a technical report. Use storytelling, visual reports, or community discussions to make learning meaningful.

3. Learning That’s Too Safe

Many organisations focus on positive lessons—what worked, what was achieved, what should be scaled up. But real learning happens in discomfort—when we look at what didn’t work and why.

Ignoring hard truths doesn’t make them disappear. It just means mistakes get repeated.

What to Do Instead

Make space for failure

Encourage open discussions about what didn’t go as planned and what adjustments are needed.

Normalise adaptation

Learning isn’t just about reflecting—it should lead to course corrections along the way.

Encourage radical honesty

If MEL systems only tell donors what they want to hear, learning isn’t really happening.

4. When Learning Becomes Extractive

One of the biggest flaws in traditional MEL systems is that learning is often taken from communities but never returned. Reports are written, findings are shared in conferences, but does any of it make a difference for the people most affected?

Marginalised groups shouldn’t just be subjects of learning—they should be partners in the process.

What to Do Instead

Make learning reciprocal

Findings should be shared back with communities in ways that are useful to them.

Design MEL as a tool for empowerment

The best evaluations don’t just measure impact—they help communities push for change.

Shift from extractive to participatory learning

Co-create evaluation questions and learning processes with the communities involved.

MEL systems don’t fail marginalised communities because of bad intentions—they fail because of bad design. When learning is centralised, sanitised, or selective, it does little to improve the lives of those it’s meant to serve.

But when done right, learning becomes a tool for change.

If we want MEL systems that truly work, attend this Saturday's webinar to learn how. Only 5 spots left!

Sign up quickly: https://www.annmurraybrown.com/webinar