Blended learning

Most definitions stop at "a mix of online and face-to-face instruction". Accurate, and almost useless, because it names the ingredients rather than the recipe. Two programmes can share identical ingredients, the same video modules and the same day in a room, and one changes how people work while the other changes nothing.

Why the simple definition misleads

The word blend does the damage. Blending implies mixing, and mixing implies the proportions are the decision, so organisations argue about the split. Sixty-forty. Two days down to one plus e-learning. None of that is design. What matters is sequence and dependency: what each part is for, what it hands to the next one, and what would break if you deleted it.

That last question is the honest test. If you can remove the online module and the live session runs exactly as before, you did not build a blend. You bought two things on one invoice and stapled them together.

A blend is not a ratio of online to offline. It is a chain, where each link is hard to use without the one before it.

It also gets confused with its neighbours. Hybrid learning normally means one session delivered simultaneously to a room and to people on a call, a delivery problem rather than a design. Flipped classroom is one specific blended pattern: content before, application during. And blended learning is not a polite name for cutting classroom days, though that is often what gets bought.

What it looks like in practice

Take feedback skills for first-time managers. A real blend might run six weeks. Week one: a fifteen-minute video on a feedback structure, plus a task where each manager writes down one conversation they have been avoiding, and submits it. Week two: a three-hour live session opening with those real cases on screen, most of it spent rehearsing in threes. Weeks three and four: they hold the conversation, then report back on a peer call, including the ones that went badly.

The bolted-together version: a ninety-minute e-learning module completed by a third of the group in the lift, then a day of slides covering the same material because the facilitator cannot assume anyone did it, then nothing.

The observable signal is simple. In a working blend, nobody in the live room is being told what the model is, because they arrived knowing. The room is spent on the hard part: saying the difficult sentence out loud while colleagues watch. That only works where people will risk being visibly bad at something, which makes psychological safety a design input.

Blended, hybrid, flipped: what the terms actually mean

How to design a blend that works

• Start from the behaviour, not the format. Name what someone should do differently on an ordinary Tuesday, then decide which format carries which part. Format-first design produces content libraries nobody opens.
• Give live time only to what needs to be live. Practice under observation, disagreement, rehearsal, feedback. Information transfer is the cheapest thing to move online and the most expensive to do in a room.
• Build a real dependency between the parts. Prework that participants bring in as their own material, put on screen and worked on. Optional prework is not done, and everyone knows it.
• Space it out. The spacing effect, documented by Hermann Ebbinghaus in the 1880s and confirmed many times since, is one of the sturdiest findings in learning research. Four contacts across six weeks usually beat one long day, even at fewer total hours.
• Complete the cycle rather than starting it. David Kolb's experiential learning cycle runs experience, reflection, conceptualisation, application. Most blends deliver the concept and abandon application, which is where behaviour change lives.
• Give the line manager a job. Two questions inside an existing one-to-one do more for transfer than another module.
• Change the facilitator's brief. In a blend, facilitation is not presenting. It is running practice and refusing to re-teach content people already had.

How to tell whether it is working

1. Can people demonstrate it, not describe it? Ask for the conversation itself, live, not a summary of the model.
2. Does it survive week eight? Anything visible on the day is partly adrenaline. What remains two months later is learning.
3. Do the digital parts get reopened? Material that became a genuine reference gets used again. Material that satisfied a completion target never does.
4. Would someone who was not in the room notice? Donald Kirkpatrick's four levels run reaction, learning, behaviour, results. Most measure the first and report as though they measured the third.

Be honest about the limits. Completion rates measure compliance, not capability. End-of-session ratings largely track how much people liked the facilitator. Self-reported confidence often rises fastest in those who learned least, because they acquired a vocabulary rather than a skill. You rarely have a control group, so clean attribution is not available. Triangulate: observed behaviour, the manager's account, one business signal agreed in advance.

Myths and common mistakes

• "Blended is cheaper." Cheaper to deliver at scale, more expensive to design well. The cost moves, it does not disappear.
• "Match the format to each person's learning style." The idea that people learn better when taught in their preferred style has been tested repeatedly and has not held up. Match the format to the task: reading for concepts, practice for skills.
• "People remember 10% of what they read and 90% of what they do." Those percentages were invented, usually traced to a distorted version of Edgar Dale's Cone of Experience, which carried no numbers. Practice really does beat reading for skills; the statistics are fiction, and quoting them weakens a sound argument.
• "70:20:10 is a rule." It came from work at the Center for Creative Leadership on how executives said they had developed. A reminder that most learning happens at work, not a budget formula.
• "We recorded the session, so remote people got the same thing." A recording is a reference, not a substitute. Nobody rehearses a hard conversation in front of a video.

Blended learning is not a mixture of formats, it is a chain of them. Decide the behaviour you want, spend the live hours on what only happens live, leave everything else where people can reach it twice, and check in week eight whether anything moved.

Tour De Force runs this as live, experiential training for teams worldwide — online and in person. Talk to us, or play Gamified learning appsThe Weekly Challenge to see the method in ten minutes.