How to Remember Everything You Hear: A System for Active Listening

You’re going to forget most of this article by next Tuesday.

That’s not a slight at me or you. It’s just how memory works. Hermann Ebbinghaus, working in the 1880s, found that humans forget about 50% of new information within an hour and about 70% within 24 hours. His research is now called the forgetting curve, and the curve has held up remarkably well across 140 years of follow-up research.

Now consider how much of your day is spent listening to podcasts. If you listen to ten hours a week, common for many of us in 2026, the forgetting curve says you’re meaningfully retaining maybe 90 minutes of it. The rest is gone by morning.

This essay is about how to keep more of it. Not all of it; that’s impossible. But meaningfully more, perhaps double or triple what you currently retain, with a manageable amount of work.

I’ll explain the science, then the system, then a one-week challenge you can try.

The forgetting curve, and why audio is uniquely bad

Ebbinghaus’s curve looks like a steep drop, then a long tail. Most forgetting happens in the first hours. After about a week, what’s left tends to stay.

What he showed, and what later research has refined, is that retention isn’t passive. It’s intervention-sensitive. Specifically, what you do in the hour after hearing something has an outsized effect on what you remember a week later. Active engagement (writing, re-explaining, applying) is the biggest lever.

For audio, the forgetting problem is uniquely bad for three reasons:

1. You can’t ctrl-F. With text, the words sit still. You can scan back, search, find. With audio, you can’t search, you can only scrub, and scrubbing is slow and unreliable.

2. You usually multitask. Most podcast listening is paired with another activity, walking, driving, cooking. The dual-task condition is well-studied in attention research, and the consistent finding is that retention drops by 20–40% under split attention. (See: Speier, Valacich & Vessey on multitasking; later work by Wood et al. on media multitasking.)

3. There’s no surface for marking. With a book, you can underline. With a podcast, the act of marking requires unlocking a device and using an app. The friction is the killer; most people skip the marking, and the marking is what would have saved the memory.

So you’re working against three things: the forgetting curve, split attention, and a missing physical surface. No wonder we forget.

The four-step system

Here is the system. It’s a synthesis of what the retention research suggests, what readers have done for centuries, and what I’ve learned building a podcast notes app and watching people use it.

Step 1: Capture (in the moment)

The single biggest lever is capture in the moment, within 30 seconds of when the thing struck you.

Why does this matter? Because what you want to keep isn’t a paraphrase of what the host said. It’s the reaction you had to what they said. That reaction has a half-life of about 30 seconds. After that, you remember roughly what struck you, but the specific quality of the strike is gone.

Capture can be:

• Voice (best): speak a sentence into your phone. Lowest friction, highest fidelity.
• Text (acceptable): type a phrase. Higher friction; risk of taking too long.
• Paper (good if you’re sitting still): scribble in a notebook. Doesn’t work when walking.

The right tool is the one that doesn’t require unlocking your phone. (This is why I built Margin, press and hold from the lock screen captures a voice note in three seconds.)

Step 2: Review (within 24 hours)

The research on spaced repetition (popularized by Anki and shown in studies going back to Ebbinghaus) consistently finds that early review dramatically slows forgetting.

Within 24 hours of taking a note, look at it again. Not to study it. Just to see it. The review reactivates the memory and resets the forgetting curve.

For podcast notes, this can be as light as:

• Opening your notes app once before bed and scrolling through what you marked
• Five minutes the next morning while drinking coffee
• A weekly “Sunday scan” if you can’t do daily

The effort is minimal. The retention payoff is large.

Step 3: Connect (within a week)

The third lever is integration with what you already know.

The research on this is rich, see Hyde and Jenkins on the “depth of processing” effect, but the gist is that information you process at a deep, integrative level (connecting it to other ideas, evaluating it, contradicting it) is retained dramatically better than information you process at a surface level (just hearing or reading).

For your podcast notes, this means: within a week of taking a note, do one of the following:

• Write a sentence connecting the idea to something else you’ve encountered
• Disagree with it (writing why you disagree forces engagement)
• Tag it to a topic in your broader notes system
• Send it to a friend with a comment

Any of these turns the note from a passive record into an encoded memory. The act of doing something with the idea is what makes it stick.

Step 4: Apply (within a month)

The deepest level of retention comes from use.

The generation effect, first documented by Slamecka and Graf in the 1970s, shows that information you produce (write, speak, teach) is remembered far better than information you only receive. This is why teaching something is the best way to learn it.

For podcast notes, “apply” means:

• Writing a tweet or post that uses the idea
• Making a decision differently because of it
• Bringing it into a conversation
• Using it in a memo at work

Most podcast notes never reach this step. The ones that do are the ones that genuinely change how you think.

Putting the four steps together

A typical loop looks like this:

Total weekly time investment: about 30 minutes for the average heavy listener. Retention gain over the no-system baseline: roughly 2-3x, based on my own informal testing with friends.

Tools that help

You don’t need any special tools to do this. A paper notebook works. Apple Notes works. But certain tools reduce friction at each step:

• For capture: Margin (press-and-hold voice notes attached to Spotify episodes), Snipd (AI clips), or Voice Memos + manual timestamp.
• For review: any notes app you’ll actually open daily. The best app is the one you’ve already opened today.
• For connect: Notion or Obsidian for building a tagged library; Readwise for spaced-repetition review of highlights.
• For apply: Twitter, your blog, conversation. The tool here is the audience.

Don’t over-engineer. The system fails most often not because of bad tools but because of unclear commitments about when each step happens.

The one-week challenge

If you want to try this seriously, here’s a one-week protocol you can run starting Monday.

Monday-Friday (capture week): - Every time you hear something on a podcast that makes you stop, mark it. Voice memo, app, paper, anything. Aim for at least 3 marks per day. (Don’t force it; if you don’t hear three good things, that’s fine.)

Each evening: - Spend 2 minutes scrolling through the day’s notes. Don’t process. Just see them.

Saturday morning: - Spend 15 minutes with all the week’s notes. For each one, do one of three things: - Keep + connect: write a sentence linking it to something else - Discard: delete it (it didn’t survive) - Promote: copy it to a separate “this matters” list

Sunday: - Pick one promoted note. Write a paragraph or send a thoughtful message to a friend about it. That’s the apply step.

At the end of the week, look at your “this matters” list. Five to ten notes that survived the filter. That’s roughly five times more keepers than the no-system version.

What the science actually says (citations for the nerds)

For anyone who wants the receipts:

• The forgetting curve: Ebbinghaus (1885), Memory: A Contribution to Experimental Psychology. Replicated extensively, including by Murre & Dros (2015) in PLoS ONE.
• Spaced repetition: Cepeda et al. (2006), “Distributed Practice in Verbal Recall Tasks: A Review and Quantitative Synthesis,” Psychological Bulletin, strong meta-analytic evidence.
• Depth of processing: Craik & Lockhart (1972), Journal of Verbal Learning and Verbal Behavior. Hyde & Jenkins (1969). Foundational.
• Generation effect: Slamecka & Graf (1978), “The Generation Effect: Delineation of a Phenomenon,” Journal of Experimental Psychology. Effect size is large and well-replicated.
• Multitasking and audio retention: Wood et al. (2012) and various follow-ups on media multitasking and comprehension.

The system isn’t novel. It’s a packaging of well-established findings into a workflow appropriate for the way we listen in 2026.

A note on what you can’t keep

I want to end with a small honest thing.

You cannot remember everything you hear. Not even close. The four-step system above will let you retain maybe 10% of the highlights of what you listen to, which is roughly 5x your no-system baseline, but is also nowhere near “everything.”

This is fine. It’s how human memory works. The point is not to keep everything; the point is to keep the things worth keeping. The act of choosing what to mark is part of what makes the marked things matter.

The forgetting is also valuable. Most of what we hear isn’t worth remembering. The forgetting curve is, in a sense, doing you a favor, it’s clearing out the noise so the signal can stand out.

What the system above gives you is a way to make sure the signal doesn’t get lost with the noise. That’s the goal. Not perfect retention. Just better than the default.

Selinay P.S. If the capture step is the one that’s hardest for you (it is for most people), it’s the one I tried to solve with Margin. Press and hold; everything else is automatic. Get on the list if it sounds useful.