§1The backstory: Asian Journal, no sleep, all worth it
In 2012, I started flipping my math classroom at Barstow High School. I did not ease into it. I flipped Algebra 1, Algebra 2, Pre-Calculus, and AP Statistics at the same time — four preps, four new video libraries to build from scratch, all while I was finishing my doctoral coursework. There is a quote from a profile of me in the Asian Journal from around that time. I said I had no social life, I was finishing my dissertation, and it was all worth it.
I have thought about that quote many times since. The "all worth it" was real — it led to a peer-reviewed randomized controlled trial, a YouTube channel that eventually earned a Silver Play Button, an Apple Book, and more than a decade of data on what the flipped classroom actually does to student outcomes in math. The work was worth it.
But the approach — flip everything simultaneously, sleep later, caffeinate through — was not the only way to get there. It was the way I happened to get there, and I would not recommend it to anyone. The five mistakes on this page are what I learned from doing it the hard way. You don't have to.
This page is a companion to the main flipped classroom guide, which covers how to set up the model, what the research says, and how to start in seven days. What you're reading now is the version I'd give to a colleague over coffee — the honest one, the one with the real mistakes in it.
§2Mistake 1: Flipping everything at once
I flipped four courses in the same semester.
This is the Asian Journal mistake. I recorded videos for all four of my preps simultaneously — new material, every day, before I had any sense of what made a good flipped-classroom video versus a long, tired one. The content was accurate. The execution was rough. And I was running on no sleep while finishing a doctoral degree.
The problem with flipping everything at once is not just the production load, though that's real. It's that you have no feedback loop. When you flip one unit in one course, you see what works and what doesn't within a few weeks. A video that's too long, a concept that needs two videos instead of one, an accountability structure that your students are ignoring — you catch it. You iterate. You get better.
When you flip four courses at once, you produce a semester's worth of videos before you've learned anything from the first ten. You end up with a large library of videos built on your earliest, least-refined instincts about the model. And then you've committed to it, because the library exists and the course is running on it.
Flip one unit in one course. Pick the unit where students historically get most confused — the place where you always know you're going to lose half the room in the live lecture. Flip that. Watch what happens. Adjust the video length, the class-time structure, the accountability loop. After one unit, you'll know more than I knew after one semester of flipping everything. Then flip the next unit. Then, if you want, expand to another course. The pace that feels slow is actually faster in the long run.
§3Mistake 2: Videos that were too long
My first videos were twelve to fifteen minutes long.
This felt natural. In a live classroom, I'd spend about that long on a chunk of new material before pausing for practice. When I moved to video, I replicated the same chunks. Twelve minutes, sometimes fifteen, occasionally more if I was covering something complex like optimization problems or related rates.
The analytics told me immediately that this was wrong. YouTube shows you audience retention — the graph of what percentage of viewers are still watching at each second of the video. My retention curves were dropping fast, and they were dropping in the same place on almost every video: around the four-to-five-minute mark. That's when students were clicking away. Not because they were lazy or distracted. Because I had kept them past the point where they had the one thing they needed, and they didn't know they were supposed to stay for the rest.
A twelve-minute video is not one lesson. It's three or four sub-concepts crammed into one container. When a student gets confused on sub-concept two and rewinds, they have to scrub through the entire video to find the part they need. When they're watching at 11 p.m. on a phone, they don't do that. They stop.
One concept per video. Five to ten minutes maximum. If you get to the eight-minute mark and you're not done, the topic is two videos, not one long one. Name them clearly: "Factoring by Grouping — Part 1: Setting Up the Groups" and "Factoring by Grouping — Part 2: Completing the Factor." Students who get Part 1 on the first watch can skip straight to Part 2. Students who need Part 1 again can rewatch just that without rewinding through Part 2. Short videos also get re-watched. Long videos get abandoned.
If you're using YouTube for your flipped content, check your audience retention graphs in YouTube Studio. The drop-off point tells you where your video stops working. It is the most honest feedback you will ever get on your teaching, because students will not tell you — they'll just stop watching.
§4Mistake 3: Treating the flip as content delivery
I thought the flip was about making videos. It isn't.
For the first couple of months, I measured my success in the model by how many videos I had posted. I had a spreadsheet. I tracked uploads. I got a little burst of satisfaction every time the library grew. What I was not tracking — what I did not think to track — was what was happening to my class time.
Here is the thing about flipping: the video is the easy part. Or at least it becomes the easy part once you've made a few hundred of them. The hard part — the part the model actually lives or dies on — is what you do with the classroom time once the lecture is gone.
In those first months, I was running class the same way I always had. I'd start with a review of what they'd watched. I'd do a few examples on the board. I'd let them practice. The structure was the same. The only change was that the initial explanation had moved online. Which means, if I'm being honest, all I had done was move the boring part of class into the evening and kept the boring part of class in the morning. The students who came prepared from the video sat through review they didn't need. The students who hadn't watched the video were lost. The class time was not doing new work.
Design the class session from scratch. Don't start with review — start with the thing the lecture was always the barrier to. For math, that's usually a harder problem set, a conceptual discussion, a peer-teaching exercise, or a group task that requires the knowledge from the video. A good flipped session assumes preparation and moves immediately into application. The warm-up problem at the start is not review — it's a diagnostic that tells you who's ready and what's still soft. The session is where the difficult thinking happens. Design it that way.
The significant performance gains in the 2016 Springer randomized controlled trial (n=91, p<.05) came from a model where in-class time was fully redesigned for collaborative problem-solving and teacher-facilitated application — not from video production alone. The control group had a competent teacher. The difference was in what happened during the session.
Esperanza, P., Fabian, K., & Toto, C. (2016). Flipped Classroom Model: Effects on Performance, Attitudes and Perceptions in High School Algebra. Lecture Notes in Computer Science, Vol. 9891. Springer.§5Mistake 4: Assuming students would watch
Some students didn't watch. I didn't know how to handle that.
This is the mistake that catches every teacher who flips for the first time. You post the video. The class session is designed around the assumption that students have seen it. A student arrives having not watched. The session starts. They're immediately behind. You have two choices: slow down the whole class to re-teach what the video covered, which defeats the purpose; or leave that student stranded while everyone else moves ahead, which feels wrong and is wrong.
I reacted badly to this in the early months. I made it punitive. I gave a daily quiz on the video content. Students who hadn't watched failed the quiz. They resented the quiz. They didn't watch anyway, or they faked the watching — played the video in the background while doing something else. The quiz was not solving the problem. It was measuring the symptom and calling that accountability.
The real issue was that for most of the students who weren't watching, the barrier wasn't attitude. It was competing demands — a job, younger siblings to take care of, a phone with no data plan in an area without reliable home wifi. Treating that as a discipline problem was a misdiagnosis. The barrier to watching is almost never laziness. It's almost always a structural constraint that I hadn't designed around.
Make the video notes part of the grade. Not the quiz score — the notes. In a traditional classroom, teachers often grade worked assignments by counting how many items students got right. That is too much work in a flipped model, and it doesn't address the right behavior. What I grade instead is completeness of notes. A quick scan tells me whether the notes deserve zero, five, or ten points. The grading is fast, the expectation is clear, and students are accountable in a way that actually changes behavior.
At the high school level, I would walk around the classroom during the warm-up activity and check students' handwritten notes on the spot. I could cover the whole room in five minutes. In college, I require students to upload a photo of their handwritten notes to Canvas before class. That makes it even easier — the record is already there, and the grade is a two-second decision per student. Zero, five, or ten.
Here is what makes this work: students who take notes while watching a video are processing the content. They aren't just pressing play and scrolling their phone. They are accountable, they receive points, and they are learning without quite realizing how much the routine is helping them. Making the notes gradeable is, in my experience, the single most effective way to ensure that most students actually watch and engage with the pre-class video. This is the secret.
§6Mistake 5: Teaching the math the way I learned it
I made videos for myself, not for my students.
This is the mistake that took me the longest to catch, and it's the one I'm most embarrassed about in retrospect — because it was hiding behind competence. My videos were accurate. The math was correct. The explanations were clear to me and to students who had already halfway gotten the concept. What they weren't was clear to the student who was genuinely lost from the beginning.
I was teaching the way I would teach myself the math. I skipped the verbal setup steps because they were obvious to me. I used notation shortcuts because they save time and I knew what they meant. I said "and as you can see" at points where a struggling student could not see, and I moved on. The students who'd already gotten the concept halfway through the practice problems were fine. The students who arrived at the video with a genuine foundational gap got lost in the same places they'd gotten lost in the live lecture — I had just moved the loss of them to the evening instead of the morning.
I didn't catch this from my own re-watches. When you watch your own video, you're filtering for correctness and completeness, not for comprehension. You know what you mean. The filter you need is a student who doesn't already know what you mean.
Find a student who typically struggles with the course — not your weakest student, but someone you know gets lost on new material the first time through. Ask them to watch your next video before you post it. Ask them one question: "Tell me the first moment you got confused." Record that moment. Re-record the video from that point, without the shortcut, without the notation jump, with a slower verbal setup. Then ask a different student to watch it. Do this once more. By the third iteration, the video will be noticeably shorter than what you started with — because you'll have cut the parts that only made sense if you already knew the math. Those cuts are the work.
I've been using this approach since around year three of Numberbender, and it's the single change that had the biggest effect on the quality of the channel. The videos got shorter and clearer at the same time — which are usually in tension, but aren't when the length was caused by explanatory shortcuts that didn't actually explain anything.
§7The smaller mistakes
There are things that don't make the top five because they fix themselves eventually — but they're worth naming so you don't lose a week to them.
File naming. I spent an embarrassing amount of time looking for "Algebra Unit 3 Video 4 final FINAL v2 USE THIS ONE.mp4" on my hard drive. Name your videos with the date, the course, the unit, and the concept. Every time, from the beginning. Something like 2026-09-14_A1_U3_factoring-gcf.mp4 will save you more time than you can imagine.
Thumbnail consistency. Thumbnails don't need to be fancy. They need to look like they belong together so students can scan a playlist and know they're in the right place. I spent a semester with inconsistent thumbnails — some with text, some without, some with my face, some without — and the playlist looked like it had been assembled by six different people. A simple template you apply to every video is enough.
Posting the wrong video. I posted the wrong version of a video the night before a quiz exactly once. The version I posted was a rough take with a math error in it. I caught it the next morning when a student asked about the error in the warm-up. The lesson: keep a simple log of what you've posted and verify the link, not just the file name, before you finalize the assignment.
These are the mistakes that teach themselves. The five above are the ones I'd want a colleague to know before they start. The model works — the research supports it, the student outcomes are there, and the teaching life it creates is genuinely better. Getting there doesn't require doing it the hard way. Thirteen years in, that's what I'd go back and say to myself in 2012: flip one unit, keep the videos short, redesign the session, build a gentle accountability loop, and make the video for your struggling student, not for yourself.
§8FAQ
What is the biggest mistake teachers make when flipping their classroom?
Trying to flip every course at once. I made that mistake in 2012 and paid for it in sleep and in video quality. The fix is to flip one unit in one course, watch what happens, and iterate. The pace that feels slow is actually faster — because you spend a semester getting good at the model before you scale it, instead of scaling before you're good at it.
How long should a flipped classroom video be?
Five to ten minutes, covering one concept. If you need more than ten minutes, the topic is two videos. Check your audience retention in YouTube Studio — the drop-off point on your retention graph is telling you exactly where your video stops working. That's the cut point for the next version.
What do I do if students don't watch the video before class?
Make the video notes part of the grade. That is the answer. A quick completeness check — zero, five, or ten points — is enough. At the high school level, walk around during the warm-up and check handwritten notes on the spot. At the college level, require students to upload a photo of their handwritten notes to your LMS before class. The check takes seconds per student, the expectation is clear, students are accountable, and they are actually learning by taking the notes. This single change has more impact than any quiz or entry ticket I ever tried. Don't make non-watching a discipline issue — make note-taking a rewarded habit.
What is the most common misunderstanding about the flipped classroom?
That the flip is about the video. The video is the easy part — or becomes the easy part. The model lives or dies on what you do with class time. If you move the lecture to video and run class the same way you always did, you've added homework without changing anything. The redesign of the session is the whole thing.
How do I know if my video is actually clear enough?
Have a student who typically struggles with the topic watch it before you post it. Ask them one question: "Tell me the first moment you got confused." Re-record from that point. Repeat once more with a different student. By the third iteration, the video will be shorter and clearer than what you started with — because you'll have removed the parts that only made sense if you already knew the math.