iPhone Video Audio Sounds Muffled? Check This Before Blaming Your Microphone

21 September 2026 Howard Toh

If your iPhone video audio sounds muffled, distant or strangely poor, you might naturally assume that something is wrong with the microphone.

Before resetting your phone, changing a dozen settings or arranging a repair, however, there is one ridiculously simple thing worth checking — especially if you have installed a transparent protective film on the back or around the camera area of your phone.

I recently helped troubleshoot an iPhone 17 Pro with muffled audio when recording videos. The phone otherwise appeared to work normally, and the usual software updates and troubleshooting didn’t solve the problem.

After researching many reports from people experiencing similar video recording audio problems — with all sorts of different suggested causes — we eventually found one that matched.

A tiny, perfectly circular piece of transparent protective film was still covering one of the phone’s microphone openings.

And it was remarkably difficult to notice.

The problem was the protective film — but not in the way you might expect

Some transparent rear protectors and camera-area protective films are manufactured with precisely cut openings for the phone’s cameras, flash, sensors and microphones.

At first glance, that sounds perfectly fine. The protector isn’t supposed to cover any of those components.

But here’s the catch.

Just because the manufacturer has cut a tiny circle into the film doesn’t necessarily mean that the circular piece will come away when the protector is installed.

The main protective film can be applied successfully while one of those tiny transparent cutout pieces remains stuck in place.

If that happens over a microphone opening, you can end up with a microphone that is physically covered by an almost invisible layer of plastic.

Because the cutout is tiny, transparent and perfectly round, it may not look like an accidental obstruction at all. Once the phone is inside a case, it becomes even easier to overlook.

Why would this make your iPhone video audio sound muffled?

A microphone works by responding to sound waves travelling through the air. Put a layer of plastic directly over its opening and you interfere with how those sound waves reach the microphone.

The confusing part is that the microphone may not stop working completely.

Your video can still have audio.

It may simply sound muffled, distant, dull or noticeably worse than it should.

That makes this problem particularly misleading. If the microphone were completely dead, diagnosing a hardware problem would be relatively straightforward. When it still records sound — just badly — it is easy to start looking for software bugs, camera settings, noise processing problems or microphone hardware failures.

Sometimes the microphone is perfectly fine.

It just has a tiny piece of plastic sitting over it.

What should you check?

If your iPhone video recordings have muffled or poor-quality audio, inspect the area around the rear cameras carefully before doing anything drastic.

This is particularly worth checking if you have ever installed:

A transparent rear phone protector
A camera-area protective film
A camera lens protector
A skin or transparent wrap
A phone case that closely surrounds the camera module

For an idea of the type of transparent rear protector involved, here is a general example of a back protector. Check its camera and microphone cutouts carefully during installation.

Look for the phone’s microphone openings and make sure that every opening that is supposed to be exposed really is exposed.

Don’t assume that seeing a perfectly round circle means you’re looking at an open hole. You may actually be looking at the edge of a transparent cutout that is still attached.

The leftover piece can be extremely difficult to see

The piece we eventually removed was tiny.

Once separated from the phone, it became obvious what had happened: it was the small circular piece created when the protector manufacturer cut the microphone opening.

While it was still attached to the phone, however, it blended into the surrounding protector surprisingly well.

Try shining a bright light across the surface at an angle rather than looking directly at it. Reflections along the edge of the film can make a leftover cutout much easier to spot.

You can also compare the microphone opening against photographs of an unprotected version of your phone to see whether something appears to be covering it.

Don’t poke anything into the microphone

This is important: don’t stick a needle, SIM ejector tool or other sharp object into the microphone opening.

You’re trying to determine whether something is sitting *over* the opening, not clean inside it.

If you discover a piece of protective film covering the microphone, carefully lift the film away without inserting anything into the microphone itself.

Once removed, record another short video under similar conditions and compare the audio.

Ideally, make a simple before-and-after test by speaking from roughly the same position and volume. If the protective film was responsible, the difference should be immediately noticeable.

What if there’s nothing covering the microphone?

This obviously isn’t the cause of every iPhone microphone or video audio problem.

If you’ve inspected the phone carefully and all of the microphone openings are unobstructed, continue troubleshooting normally.

Some sensible next steps include:

Remove the phone case temporarily and test again.
Check whether dirt or debris is obstructing an external microphone

opening.

Compare recordings made with the Camera app and other apps.
Test different cameras and recording modes if the problem appears

specific to video recording.

Restart the iPhone.
Install any available iOS updates.
If the problem persists, contact Apple Support or arrange for the

phone to be inspected.

The point isn’t that protective film is always responsible. It’s that this check takes less than a minute and can potentially save you from doing considerably more complicated troubleshooting.

What about the iPhone 17 Pro Max, iPhone 18 Pro and other phones?

Although the problem described here was diagnosed on an iPhone 17 Pro, the underlying issue isn’t unique to that model.

The same principle can apply to the iPhone 17 Pro Max, iPhone 18 Pro, iPhone 18 Pro Max and other iPhone models — as well as Android phones — whenever a protective film, case, skin or camera protector can obstruct a microphone opening.

The exact position and number of microphones varies between phone models, so don’t rely on the location shown in a photograph of a different device. Check the microphone locations for your particular model.

This is especially relevant when using third-party protectors. A protector may have all the correct cutouts and still cause trouble if one of the tiny pieces created during manufacturing doesn’t separate cleanly during installation.

One quick check before blaming the microphone

If your phone records video with inexplicably muffled audio, take a close look at it before diving into software troubleshooting.

Check the case. Check the camera protector. Check any transparent film. And most importantly, make sure that the tiny microphone cutouts are actually open holes rather than transparent circles that are still sitting over them.

In our case, that tiny piece of clear film was the culprit.

Hopefully, if you’ve arrived here searching for why your iPhone 17 Pro video audio sounds muffled — or you’re seeing the same problem on an iPhone 17 Pro Max, iPhone 18 Pro, iPhone 18 Pro Max or another smartphone — this gives you one very quick thing to check before assuming that your microphone needs repairing.

Can an Industry-Run Recycling Scheme Protect Consumer Choice?

20 September 2026 Howard Toh

Singapore’s Beverage Container Return Scheme (BCRS) has a worthwhile objective: increase the recovery and recycling of beverage containers by putting a refundable 10-cent deposit on regulated cans and bottles.

But as the scheme moves towards full implementation, there is another public interest worth protecting alongside recycling: consumer choice and competition.

In particular, will smaller beverage importers — including the parallel importers that bring Singapore consumers cheaper alternatives, niche brands, unusual flavours and limited-edition products — be able to comply with BCRS as easily and economically as the largest beverage producers?

This question deserves particular attention because BCRS Ltd is itself industry-led.

The National Environment Agency describes BCRS as a "producer-led Scheme". BCRS Ltd’s board comprises Coca-Cola Singapore Beverages, F&N Foods, Pokka, Wanin Industries and Chia Khim Lee Food Industries.

There is nothing inherently wrong with an industry-led scheme. Producers have the technical expertise and, under Extended Producer Responsibility, are supposed to bear responsibility for the waste created by their products.

But BCRS is also mandatory.

From 1 October 2026, regulated beverage products supplied in Singapore must carry the BCRS Deposit Mark, and the 10-cent deposit applies to regulated containers.

That makes the accessibility, cost and neutrality of the compliance process important not merely to beverage companies, but ultimately to Singapore consumers.

One scheme, three very different economies of scale

Consider three hypothetical businesses selling exactly the same category of beverage.

The numbers below are illustrative rather than quotations of what any particular company pays. Their purpose is to demonstrate why the same regulatory requirement can have very different commercial consequences at different scales.

Infographic comparing BCRS compliance economics for a large producer of 1 million units, a small importer of 50,000 units and a niche or parallel importer of 1,000 units. — Illustrative comparison: the same BCRS obligation can produce very different per-unit economics depending on scale. The refundable 10-cent consumer deposit is not treated here as a compliance cost.

Producer A: 1,000,000 containers

Imagine a major beverage producer expecting to sell one million units of a product in Singapore.

At this scale, producing Singapore-specific packaging is commercially realistic. The deposit mark and appropriate barcode can potentially be incorporated into the packaging artwork before cans or bottles leave the production line.

There will still be BCRS producer fees, registration requirements, artwork changes and other implementation costs. But crucially, the company can potentially avoid buying and manually applying one million individual stickers.

Its fixed implementation costs are also spread over one million units. This is economies of scale doing exactly what economies of scale normally do.

Producer B: 50,000 containers

Now consider an independent distributor bringing in 50,000 cans.

Creating a Singapore-only production run may no longer be practical. The overseas manufacturer may have minimum production quantities far beyond what the Singapore importer can sell. Stickering therefore becomes considerably more relevant.

BCRS requires scheme stickers to be ordered through BCRS-appointed printing companies. Stickers from other printers are not accepted. BCRS states a printing lead time of approximately two to six weeks depending on quantity. If the importer wants the printer to perform the physical stickering as well, services can include opening cartons, removing products, applying the stickers and repacking them.

There has been a welcome development for some very small producers. Eligible micro-producers of aluminium cans can obtain up to 50,000 pre-serialised stickers annually instead of individually registering products.

That concession suggests that NEA and BCRS recognise the disproportionate administrative burden that product-by-product registration can impose at small scale.

But it also raises useful questions. Who qualifies? What happens to plastic-bottle importers? What happens when an importer carries many niche SKUs rather than large quantities of a few products? And what are the actual all-in sticker and handling costs after the available assistance ends?

Producer C: 1,000 containers

Finally, consider a parallel importer bringing just 1,000 units of a seasonal Japanese drink, a European craft soda or a limited-edition flavour.

Asking the foreign manufacturer to redesign its production packaging for 1,000 Singapore-bound cans is unlikely to be realistic. The importer therefore needs another route to BCRS compliance.

And this is where per-unit economics become important.

Published reporting has put BCRS-approved stickers at approximately 4 to 18 cents each depending on volume, before the labour required to apply them.

At the upper end of that reported range: 1,000 × $0.18 = $180 in stickers alone.

That is before someone opens the cartons, handles every container, applies every sticker and repacks them. It is also before the ordinary producer fee and administrative requirements.

The 10-cent consumer deposit should not be included in this comparison as though it were another compliance expense: the deposit is refundable to the consumer when the container is returned.

The sticker, handling and compliance expenses are different. Those are genuine costs associated with getting the product onto the Singapore market. For an inexpensive beverage with a thin importer margin, an additional ten or twenty cents per unit can matter considerably.

Same requirement, very different consequences

1,000,000 units: Singapore-specific packaging may be incorporated at the factory. Fixed implementation costs are spread across enormous volume.

50,000 units: Stickering may be necessary, although the micro-producer concession could substantially reduce registration friction for eligible aluminium-can producers.

1,000 units: Singapore-specific manufacturing is much less realistic, while sticker, labour and administrative costs are divided among very few units.

The environmental requirement is identical. The economics are not.

This doesn’t mean that BCRS is deliberately disadvantaging smaller competitors. There is no evidence presented here that it is.

NEA has also introduced a Producer Transition Grant of up to $2,500 to help producers with costs including registration fees, producer fees and scheme stickers. That is a positive acknowledgement that implementation costs can disproportionately affect smaller businesses. But temporary assistance doesn’t eliminate the longer-term question.

Why should consumers care about parallel importers?

It would be easy to treat this as an argument between beverage companies. It isn’t.

Parallel importers and independent distributors perform a useful competitive function in a small market such as Singapore. They can source products from other markets when doing so is commercially attractive. They also bring in products that an official distributor may decide are not worth launching locally: regional flavours, seasonal editions, specialist drinks and brands with relatively small followings.

The result for consumers is potentially more choice and more price competition.

If compliance imposes a relatively small cost on a mass-market beverage but a substantial per-unit cost on a low-volume import, some products may simply cease to be commercially worthwhile to bring into Singapore.

The consumer may never see a line item saying "BCRS compliance: 18 cents". Instead, the drink may become more expensive. Or the importer may decide not to import it at all.

That second outcome is arguably harder for consumers to notice. You cannot complain about the price of a product that quietly disappeared from the market.

The question isn’t whether BCRS should exist

Singapore needs better recycling outcomes, and a container-return system is a legitimate way of pursuing them.

Nor is this article suggesting that BCRS Ltd or the beverage companies involved in it have deliberately disadvantaged their competitors.

Singapore’s competition regulator has previously examined the establishment and operation of BCRS and concluded that the arrangement was unlikely to infringe sections 34 and 47 of the Competition Act.

But regulatory compliance and good governance are not quite the same question as public confidence.

When participation in a national scheme is compulsory, and the organisation administering that scheme was established by major participants in the same industry, transparency becomes particularly important.

Consumers and smaller competitors should be able to understand: How are sticker prices determined? Why must stickers come only from BCRS-appointed suppliers? How competitive was the process for appointing those suppliers? How quickly are products approved? What information about a competing product or importer is visible to BCRS, its management and its board? What safeguards separate commercially sensitive information from companies represented on that board? What happens when a product is rejected? Is there an independent avenue of appeal? And does BCRS monitor whether its requirements are causing low-volume products to disappear from Singapore?

These aren’t accusations. They are questions that become reasonable whenever an industry-led organisation is entrusted with administering a mandatory national scheme affecting every competitor in that industry.

Recycling without sacrificing competition

There need not be a choice between recycling and competition. A successful BCRS should be capable of achieving both.

If technology eventually allows small importers to obtain compliant labels more cheaply, approval becomes faster, administrative requirements are simplified and micro-producer concessions are expanded where evidence shows they are needed, that should be celebrated.

But the measure of the scheme’s success should not simply be the number of containers returned.

We should also ask whether Singapore reached that environmental objective without unnecessarily making it harder for small businesses to compete or reducing the range of products available to consumers.

A million-can producer and a thousand-can importer may both be subject to exactly the same environmental obligation.

Making sure that obligation does not inadvertently become a barrier to competition is precisely why transparency matters.

Sources include the National Environment Agency, BCRS Ltd, the Competition and Consumer Commission of Singapore, and published reporting by CNA. Cost examples in this article are illustrative and should not be interpreted as the actual costs incurred by any named company.

Flipper Zero: 5 Surprisingly Useful and Fun Things You Can Do With It

14 September 2026 Howard Toh

The first time I saw a Flipper Zero, I assumed it was mostly one of those gadgets you buy because it looks cool, play with for a weekend, and then leave in a drawer.

Turns out, it is considerably more useful than that.

Yes, it has acquired a slightly ridiculous reputation online as some sort of pocket-sized hacking superweapon.

It isn’t.

What it actually is, though, is a rather fun collection of radios, sensors and interfaces bundled into one tiny device.

And once you start carrying one around, you begin noticing how many things around you are communicating wirelessly.

Here are some of my favourite practical uses.

1. Replace Half Your Infrared Remotes

This is probably one of the least dramatic Flipper Zero features.

It is also one of the ones I actually find most useful.

The Flipper has an infrared transmitter, so it can behave like a universal remote for things such as TVs, projectors, sound systems, air conditioners, fans, and some air purifiers and other appliances.

Flipper maintains a library of remote-control signals, and its mobile app can download compatible controls directly onto the device. If your appliance isn’t in the database, you can also point the original remote at the Flipper and capture its infrared commands.

That means the Flipper can effectively become a backup remote for appliances around the house.

This has saved me from the classic household mystery: “Where did the aircon remote go?”

Usually followed by somebody insisting they were definitely not the last person holding it.

Flipper’s official documentation covers both its universal infrared remotes and infrared functionality.

2. Copy Your Own Gate or Gantry Remote

This one feels considerably more magical.

Many older remote-controlled gates, barriers, roller shutters and carpark gantries operate using radio signals in the Sub-GHz range.

For compatible static-code remotes, Flipper can receive the signal, save it and later transmit it again.

So instead of carrying another remote around, you may be able to keep a copy on your Flipper.

There is an important catch.

Modern gate systems frequently use rolling codes, where the transmitted code changes every time the remote is used. These systems are specifically designed so that simply recording one transmission and playing it back later doesn’t work.

So this isn’t a universal gate-cloning machine.

But when you come across one of your own devices using a simple static remote, being able to capture it and then watch the gate open from your Flipper is enormously satisfying.

Flipper documents this under its official Sub-GHz documentation.

Important: only record or reproduce remotes for gates, barriers or other equipment that you own or are explicitly authorised to operate. Being technically able to transmit a signal does not grant permission to open somebody else’s gate.

3. Keep Copies of Access Cards You Are Authorised to Use

This is probably the feature most people associate with Flipper Zero.

It supports both 125 kHz RFID and 13.56 MHz NFC, which covers a wide range of cards, tags and key fobs.

Older 125 kHz access cards are particularly interesting because many simply transmit an identification number when presented to the reader.

For supported formats, Flipper can read the card, save it and emulate it later.

Its NFC capabilities are more sophisticated. Depending on the card technology and security involved, Flipper may be able to read and emulate some or all of the card data.

And this distinction matters: not every access card can be cloned.

A simple RFID card and a cryptographically protected smart card are very different beasts.

Flipper’s own documentation describes its 125 kHz RFID support and separate NFC capabilities.

Personally, one situation where I find this genuinely convenient is Airbnb.

Occasionally an Airbnb gives us only one or two access cards despite several people staying in the apartment.

If the provided access card uses a compatible technology, I sometimes make a temporary copy so we don’t have to coordinate our entire family’s movements around one physical card.

But there is an important boundary here.

I am authorised to access the property during my rental period.

Once that rental ends, so does that authorisation.

So I always delete and destroy any copies I created when I check out.

That seems like a fairly sensible rule for any copied credential: if your permission to enter disappears, your copy should disappear too.

And obviously, don’t copy somebody else’s office, condo, hotel or access card without permission.

4. Figure Out What Technology Your Cards Actually Use

This sounds nerdier than it actually is.

One of the things I like most about Flipper Zero is that it makes invisible technology visible.

Take an access card. Normally you tap it against a reader, hear a beep, and the door opens. That’s about as much thought as most people ever give it.

But put that same card against a Flipper and suddenly you start asking questions.

Is it 125 kHz RFID? 13.56 MHz NFC? MIFARE? Does it expose only a UID? Is there protected memory? Can it actually be emulated?

Even when Flipper cannot reproduce a credential, simply identifying what technology is inside the card is surprisingly interesting.

It also gives you a much better appreciation of just how different two pieces of plastic that look virtually identical can be.

5. Learn How Much Wireless Stuff Is Around You

This may actually be the best reason to own one.

The Flipper Zero encourages curiosity.

You start noticing infrared LEDs. You start wondering what frequency a remote uses. You realise that your condo gate, car key, door badge, TV remote, NFC tag and wireless doorbell are all communicating using completely different technologies.

And because the Flipper gives you one interface for exploring many of them, concepts that previously sounded abstract start becoming tangible.

Infrared suddenly isn’t just something buried inside a remote control. RFID isn’t simply “the thing that opens the office door.” Sub-GHz isn’t merely a frequency range.

You can actually interact with these technologies and see what they are doing.

That makes Flipper Zero much more interesting to me as an educational gadget than as the “hacking device” it is sometimes portrayed as online.

Is Flipper Zero Worth Buying?

If you expect a magical device that opens every door and controls everything around you, you will probably be disappointed.

Modern security systems are deliberately designed to prevent exactly that.

But if you’re the type of person who enjoys figuring out how things work, it is an incredibly entertaining gadget.

It combines infrared, Sub-GHz radio, RFID, NFC, Bluetooth, GPIO and a bunch of other experimentation possibilities into something you can actually carry around in your pocket.

And unlike a lot of hobbyist electronics, you don’t need to build a circuit every time you want to try something.

You simply pull out the little dolphin gadget and start investigating.

If you’re looking to get one locally, you can find the Flipper Zero on Shopee Singapore.

Just remember that owning the tool doesn’t change the rules.

Experiment with your own equipment, your own credentials, or systems where you have permission.

Within those boundaries, there is an awful lot of fun packed into that little orange-and-white box.

Telegram Channel & Group Cloning: Because “Lifetime Access” Sometimes Lasts Until Tuesday

23 August 2026 Howard Toh

You joined a useful Telegram channel years ago. Maybe it contains a paid course, a community’s accumulated advice, a niche hobby archive, family photos, work references, deal alerts, or simply a group of people who became part of your life.

You paid for access. You were told it was permanent. Then one day the invite link stops working, the administrator disappears, or the channel is deleted without warning. No refund, no download button, no polite little “thank you for your custom” email—just a hole where your archive used to be.

That is not a hypothetical concern. It has happened to me repeatedly: eight out of ten channels I had paid to access eventually disappeared. The lesson was not to collect more links. It was to keep an archive of the material I was legitimately allowed to see while it still existed.

This service clones the channel or group into an organised destination you control. Rather than merely handing you a folder of downloads, the aim is to preserve the actual chat-channel experience: the posts, media and accumulated history, in a separate Telegram archive that remains useful when the original does not.

What a Telegram clone gives you

A proper clone is more than a pile of random downloads. The aim is a structured, Telegram-native copy of the material you already have access to, before an admin changes their mind, loses an account or simply evaporates into the internet mist.

A real channel/group copy — content is replicated into a separate destination, not reduced to a folder of loose attachments.
Media retained with the posts — photos, videos, documents, audio and other attachments stay associated with the relevant content where feasible.
Order retained — channel or group content stays grouped in a usable timeline rather than becoming one heroic folder called misc_final_v7_reallyfinal.
A destination you manage — your cloned archive lives under an account/destination you control.
A familiar way to browse — revisit the archive from Telegram in a channel-style format instead of hunting through a hard drive.

Who this is for

This is useful for people who want to protect access to content and conversations that matter to them:

Members of paid communities, courses or membership channels
Community owners and moderators wanting a contingency archive
Hobby, trading, gaming and interest groups with years of accumulated tips and files
Families, alumni groups and communities preserving shared history
Anyone who has already learned that “lifetime” is an oddly elastic word online

Export vs. clone: what is the difference?

Option	What you get	Best for
Standard export	A file-based record of messages and attachments for local storage.	People who are happy to browse content outside Telegram.
Telegram clone	A separate, organised Telegram destination that replicates accessible posts and media in a channel/group-style format.	People who want the content to remain practical to browse and reference inside Telegram.
Clone + file export	A Telegram-native clone alongside a separately delivered file copy, where requested.	The “I have been burned before” plan.

Straightforward pricing

The real workload is driven largely by the size and number of media attachments, not merely how many messages appear in the group. I will assess the channel/group first, confirm the scope and give you a clear quote before the work starts.

Plan	Best for	Included files	Price
Restricted clone	Channels or groups where forwarding, saving or media access needs extra handling.	First 2,000 files	S$50 base, then S$0.03 per additional file
Unrestricted clone	Sources that allow forwarding or saving files, including private channels/groups without those limitations.	First 1,000 files	S$20 base, then S$0.01 per additional file
External file export	Separate disk/file delivery requiring extra storage and physical file handling.	Quoted by scope	Custom quote

Setup fee: Both clone plans have a S$10 setup fee, waived when the confirmed order is at least S$100. Large videos, unusually large source groups and external file exports may need additional handling. There are no surprise calculations after the work begins: the agreed scope comes first.

How it works

Send the channel/group link and a short description. Tell me the channel/group you want cloned and its intended destination.
I assess the source. I check the likely media volume, file sizes, access method and the appropriate clone structure.
You receive a fixed scope and price. You decide before any work begins.
The clone is created. Accessible posts and media are organised in the new Telegram destination.
You receive your cloned archive. It remains available to browse from the destination you control.

Why do this now instead of later?

Telegram channels and groups are convenient living spaces, but access to someone else’s channel is not the same as control of your own copy. Administrators can delete content, lose control of an account, make a channel private, migrate elsewhere, or stop operating altogether. Even well-run communities can disappear for reasons that have nothing to do with you.

The best time to preserve something is while it is still there, your access is active, and the material is complete. After it closes, there is usually no recovery process waiting backstage with a cape.

Frequently asked questions

Can you back up a private group or paid channel?

I can assess channels and groups that you can legitimately access. The practical method, scope and what can be retained depend on the source, permissions, media size and platform behaviour.

Will the clone retain the original media?

Where the source makes them available, the clone is prepared to retain the best practical available version. I will flag any limitations before starting.

Can you clone a source with tens of thousands of files?

Yes. Larger archives are exactly why a structured process and per-file pricing are useful. The scope is assessed first, then you receive a clear quote.

Where does the completed clone live?

It lives in the agreed Telegram destination under your control. A separate file export is not part of the standard clone plans, but can be quoted separately if you need files delivered outside Telegram.

Get a Telegram cloning quote

If there is a channel, group or community archive you would hate to lose, send the link and a short note on what you want cloned. I will first confirm what is feasible, how the clone will be structured and the price.

Contact me on Telegram for an assessment

—

Important: This service is for preserving material you are entitled to access and retain. It does not grant access to restricted content, bypass permissions, or transfer ownership of anyone else’s work. Please respect applicable terms, copyright and community rules.

Welcome to the Museum of Meaningless Metrics

21 June 2026 Howard Toh

Welcome to the Museum of Meaningless Metrics

I recently encountered a comic depicting the “Museum of Meaningless Metrics”.

Its exhibits included:

Lines of Code
Story Points
Pull Requests
Tokens Spent

The newest and proudest exhibit was a towering token counter, presumably recording enough artificial intelligence activity to heat a modest apartment block.

The comic is funny because it captures a recurring organisational habit: when something important is difficult to measure, we measure something nearby instead.

Productivity is difficult to measure. Value is difficult to measure. Quality, good judgement, avoided mistakes and long-term maintainability are all difficult to measure.

Counting things, however, is wonderfully easy.

And once a number appears on a dashboard, it acquires an air of authority. It becomes a target, a performance indicator and eventually a presentation slide with an upward-pointing green arrow.

The problem is not that these metrics contain no information. Most of them do. The problem begins when they are promoted from operational signals into proof of productive work.

Exhibit One: Lines of Code

To a non-programmer, lines of code can look like a straightforward measure of output.

One developer writes 500 lines. Another writes 5,000. Clearly, the second developer has done ten times as much work.

That conclusion is roughly as reliable as judging an author by the weight of the manuscript.

A large codebase may genuinely represent a substantial amount of work. A complex application cannot always be produced in 12 elegant lines and a motivational comment. But the number of lines tells us very little about whether the work was necessary, correct or well designed.

Those 5,000 lines could represent:

a major new system;
duplicated logic;
an unnecessarily complicated solution;
generated boilerplate;
poorly understood requirements;
several fixes for problems introduced by earlier fixes.

Meanwhile, an experienced developer might spend a day studying the problem and submit a 20-line change that eliminates the need for hundreds of additional lines.

That does not mean fewer lines are always better. Code can be compressed until it resembles an ancient curse. Readability, maintainability and testing matter more than winning a code-golf tournament.

The meaningful questions are not “How many lines were written?” but:

Does the software solve the intended problem?
Is it reliable?
Can somebody else understand it?
Can it be changed safely later?
Did it reduce or introduce complexity?

Lines of code can help estimate the size of a codebase. They are poor evidence of an individual developer’s productivity.

Exhibit Two: Story Points

I have participated in sprint-estimation exercises and have also been asked to assign story points.

In theory, story points provide a relative indication of a task’s size. They can incorporate effort, complexity, uncertainty and risk. A task assigned eight points is not necessarily expected to take eight hours or eight days. It is simply considered substantially larger or less predictable than a task assigned two points.

The Fibonacci-like sequence—1, 2, 3, 5, 8, 13 and so on—is intended to discourage fake precision. As tasks become larger, our confidence in estimating them usually decreases. The gap between the available numbers forces teams to acknowledge that uncertainty.

That is the theory.

In practice, the exercise can sometimes feel like a group of people cautiously plucking numbers from the sky while discreetly checking what everyone else has plucked.

Developers and requestors may view the same task very differently. One sees a minor form change. The other sees legacy code, undocumented dependencies, six integration points and a database that becomes emotionally unstable after 4 p.m.

Ideally, disagreement exposes hidden assumptions. Someone who estimates a task as two points may know something the person estimating eight does not. The discussion is supposed to produce shared understanding.

But the process breaks down when:

people anchor their estimates to the first number mentioned;
participants follow the majority to avoid defending a different view;
developers agree on the answer beforehand;
management treats points as hours in disguise;
velocity becomes a performance target;
teams inflate estimates to make their output appear larger.

At that stage, story points stop being a conversation aid and become ceremonial arithmetic.

The points themselves were never meant to be objective units. An eight-point task for one team cannot reliably be compared with an eight-point task for another. Even within the same team, the meaning may drift as the people, technology and work change.

Story points can help a stable team forecast its own capacity. They become dangerous when outsiders mistake them for a universal measurement of productivity.

Exhibit Three: Pull Requests

Pull requests are an important part of modern software development. They allow proposed changes to be reviewed, discussed, tested and improved before being merged into the main codebase.

That does not make the number of pull requests a meaningful productivity score.

A developer who submits many pull requests may be working in a disciplined way, breaking changes into small and reviewable units. Another developer might submit fewer pull requests because the work involves longer-term research, architecture or debugging.

Or the high count could come from:

repeated corrections;
failed implementations;
unnecessary fragmentation;
automated dependency updates;
trivial formatting changes;
fixes for problems introduced by earlier pull requests.

Conversely, a low number could represent careful, high-impact work—or a single 18,000-line pull request that causes every reviewer to suddenly develop an urgent dental appointment.

The count does not tell us whether the changes were useful, safe or even necessary.

Better questions include:

How quickly are useful changes delivered?
How often do changes introduce defects?
Are reviews substantive or merely ceremonial?
Are pull requests reasonably sized?
Does the team learn from incidents and review feedback?
Are customers or internal users seeing meaningful improvements?

Pull requests are a workflow mechanism. Counting them is like measuring a restaurant by the number of plates that passed through the kitchen.

Interesting, perhaps. Conclusive, no.

Exhibit Four: Tokens Spent

Now we arrive at the museum’s newest attraction: tokens spent.

As organisations adopt generative AI, token counts have become the latest impressively large numbers available for dashboards.

Someone used ten million tokens.

Excellent.

What happened?

Did those tokens produce a working application, complete useful research, resolve customer cases or automate a tedious process?

Or did an AI agent repeatedly inspect the same repository, misunderstand the task, rewrite its own plan six times and confidently announce that it had completed work for which it did not even possess the required credentials?

The token count cannot tell us.

Tokens are units of text processed or generated by an AI model. They matter for practical reasons:

expenditure;
capacity planning;
latency;
model selection;
efficiency optimisation;
detecting unexpectedly wasteful workflows.

But tokens spent are an input cost, not an output measure.

Using more tokens does not necessarily mean that more reasoning occurred. It may mean that the task was complex, the context was large or extensive validation was performed. It may also mean that the prompt was inefficient, the agent became stuck in a loop or the system generated vast quantities of polished nonsense.

Similarly, minimising tokens is not automatically desirable. Spending additional tokens to verify an important result may be entirely justified. A cheap but incorrect answer is not efficient. It is merely an error delivered at a discount.

The more meaningful measures depend on the objective:

Was the task actually completed?
Was the result accurate?
How much human correction was required?
How much time or money was saved?
Was the outcome independently verified?
Did the system avoid inventing actions, identifiers or results?
What was the cost per successful outcome?

“Tokens spent” might earn bragging rights among people with generous compute budgets. It does not, by itself, demonstrate useful work.

When the Metric Becomes the Mission

The deeper issue behind all four exhibits is not measurement itself. Measurement is necessary.

The trouble begins when a proxy becomes a target.

When developers are rewarded for lines of code, codebases grow.

When teams are rewarded for completing story points, estimates mysteriously expand.

When pull-request counts become visible performance measures, work gets divided into more pull requests.

When token usage is presented as proof of AI adoption, systems find remarkably creative ways to consume tokens.

People adapt to the measurement system placed around them. This is not necessarily dishonesty. It is often rational behaviour. If an organisation repeatedly signals that a number matters, employees will naturally optimise for that number.

Unfortunately, the number may become healthier while the actual outcome deteriorates.

What Should We Measure Instead?

There is no single perfect replacement metric. That is precisely why organisations keep returning to convenient counts.

Useful measurement normally requires a combination of quantitative data and human judgement.

Depending on the work, better indicators might include:

successful outcomes delivered;
user or customer impact;
reliability and defect rates;
lead time from request to usable result;
frequency and severity of production incidents;
rework required;
maintainability;
cost per successful task;
quality of documentation and knowledge transfer;
whether the original problem was actually solved.

Even these can be manipulated or misunderstood. No dashboard eliminates the need for context.

A metric should begin a conversation, not end one.

The Museum Is Still Expanding

Lines of code, story points, pull requests and tokens spent are not entirely useless. Each can answer a narrow operational question.

How large is this repository?

How does this team perceive the relative size of upcoming work?

How are changes flowing through the review process?

How much AI capacity are we consuming?

Those are legitimate questions.

The absurdity begins when the answers are repurposed to answer much larger questions:

Who worked hardest?

Which team is most productive?

Who created the most value?

Is our AI programme successful?

Numbers cannot answer those questions simply because they are available, countable and aesthetically compatible with a bar chart.

So perhaps the comic’s museum label is slightly unfair.

These are not always meaningless metrics.

They are metrics from which we frequently demand meaning they were never capable of providing.

And judging by the rate at which dashboards are being built, the museum gift shop should be extremely profitable.