A jingle cannot play in everyone's head forever, because the brain treats an earworm as a short-lived memory loop: diary studies find that most episodes fade within hours and rarely recur beyond 24 hours, and every model of how a tune spreads between people only reaches 100% of a population if nobody ever recovers. If you forced it anyway, a 10-second jingle looping through 1.46 billion Indian heads would deliver about 8.4 trillion plays a day, worth roughly ₹2.1 lakh crore at YouTube bumper-ad prices, which is more than India's whole 2025 ad market in under 14 hours.
- 90%+ of people get an earworm at least once a week, but episodes rarely last past 24 hours.
- In an SIS epidemic model, a tune reaches 100% of heads only if recovery is zero, which habituation rules out.
- A 10-second loop is 5,760 plays a day per head and about 8.4 trillion across India.
- Priced at a ₹250 bumper CPM, that beats India's whole 2025 ad market in about 14 hours.
- Only about ₹19,000 crore a year of it is useful advertising; the rest is irritation and lost sleep.
The catch: advertising value stops at a handful of exposures a week, so the useful part is worth about ₹19,000 crore a year, and everything beyond it is irritation, lost sleep and a brand people learn to hate.
Mark Twain wrote this scenario in 1876, and it is the best place to start.
In February 1876, The Atlantic Monthly printed Mark Twain's sketch "A Literary Nightmare". Its narrator reads a jingle in the morning paper, a rhyme built from a horsecar fare notice, and it takes over his mind. He loses track of what he is doing, and even of whether he has eaten.
Days later, on a walk with his friend the Reverend, he passes the jingle on without meaning to, and is cured. The Reverend arrives soon after in a terrible state, the rhyme having wrecked his own work with his congregation. He is cured only when he is taken to a meeting of university students and hands it to them.
Twain did not write the rhyme. Isaac Bromley and Noah Brooks made it up that year while riding a tram, finished it with W. C. Wyckoff and Moses W. Handy, and the New York Tribune printed it. Richard Dawkins later used it in Unweaving the Rainbow as an example of a meme.
Wikipedia: A Literary Nightmare (Mark Twain, The Atlantic Monthly, 1876)Psychologists now call the effect involuntary musical imagery (INMI), and they have measured it properly. So here is the jingle-forever premise, solved with the real numbers.
What makes a jingle get stuck in your head?
A tune sticks when it is fast, has a familiar overall shape and contains one surprising interval, and when you heard it recently and often. Those are the measured predictors, not folklore.
The melody
The largest study of what makes a song "sticky" is by Kelly Jakubowski and colleagues at Goldsmiths and Durham. They took the songs most often named as earworms in a large online survey, matched each with a song of similar popularity and chart history that was not named, and compared 83 melodic features.
Earworm tunes differed in three ways:
- Faster average tempos;
- Common global contours, such as the up-then-down arch you find in nursery rhymes;
- Less common gaps between the turning points of the melody.
In plain words: a predictable shape with one odd leap in it. That combination is easy to rehearse and slightly unfinished, which keeps the brain replaying it.
The team compared tunes frequently reported as earworms with matched non-earworm tunes, using 83 statistical and corpus-based melodic features measured against a large pop corpus. Earworm tunes had faster tempos, more common global contours and less common average gradients between melodic turning points, and chart success and recency also predicted which songs got stuck. The limit: the sample was Western pop, so the features may not transfer directly to Indian film music, which uses different scales and longer melodic phrases.
Read the paper (free)The memory loop
The mechanism sits in working memory. Alan Baddeley's model of working memory has a phonological loop, a short rehearsal buffer that holds a brief stretch of speech-like sound and keeps it alive by silent repetition. A jingle hook is built to fit that buffer.
Evidence that the loop is involved comes from a neat experiment at the University of Reading. When people chewed gum, which keeps the mouth and articulation system busy, they reported about 7 intrusions of a just-heard song in three minutes instead of about 10, a drop of roughly 30%. Tapping a finger did less, so the effect is about sub-vocal rehearsal, not general distraction.
Three experiments played participants catchy songs and then counted how often the tune came back in a three-minute window. Chewing gum reduced both voluntary and involuntary musical thoughts (around 10 intrusions in the control group against around 7 with gum), changed the experience of actually "hearing" the music, and beat finger-tapping, which ruled out a simple dual-task effect. The limit: these are short lab windows with students, not days of real life.
Read the paper (free)| Factor | What the research found | What a jingle writer does with it |
|---|---|---|
| Tempo | Earworms are faster on average (Jakubowski 2017) | Keep the hook brisk |
| Melodic contour | Common overall shapes stick (Jakubowski 2017) | Use a familiar arch, one unusual leap |
| Recency and repetition | Recent exposure is the commonest trigger (Williamson 2014) | Buy frequency, not just reach |
| Low attention | Earworms rise in low-load states such as chores (Scullin 2021, citing Floridou 2017) | Place audio where minds wander: commutes, kitchens |
| Articulation | Gum cuts intrusions by about 30% (Beaman 2015) | Hooks that are easy to sing beat hooks that are hard to sing |
How long does an earworm actually last?
An ordinary earworm lasts minutes to hours and is usually gone by the next day. That is the single biggest obstacle to "forever".
Victoria Williamson, Lassi Liikkanen and colleagues analysed two large surveys, one in Finland with about 12,000 respondents and one in English with 5,989, and found that over 90% of people get an earworm at least once a week.
Earlier, Philip Beaman and Tim Williams ran questionnaire and diary studies and found that recurrence of the same earworm was uncommon and unlikely to persist past 24 hours. They also found that fighting it works worse than accepting it.
That fits Daniel Wegner's theory of ironic mental control: trying not to think of something keeps a monitor process checking for it, and the checking re-activates it.
The authors coded open-text answers from a Finnish survey (about 12,000 respondents) and an English survey (5,989). More than 90% reported an earworm at least weekly. In Finland, 74.6% said they hummed, sang or talked the tune aloud, 60.2% tried to identify the song and 57.3% went and listened to it, which is exactly the behaviour that spreads a tune to other heads. Recent exposure and memory associations were the commonest triggers. The limit: everything is self-report, and the two surveys asked slightly different questions.
Read the paper (free)
Could one jingle spread to everyone, like Twain's?
A jingle can spread like an infection, but simple epidemic maths shows it can only occupy every head at once if nobody ever gets over it, which the brain does not allow. Here is the working.
Treat the jingle as a disease in an SIS model (susceptible, infected, susceptible again), the standard model for infections that give no lasting immunity, such as the common cold. Assumptions:
- \( N \) people mix randomly. \( I \) of them have the jingle in their head; \( S = N - I \) do not.
- Each infected person hums or sings it near others, infecting at rate \( \beta \) per day. The 74.6% humming figure above is why \( \beta \) is not zero.
- Each infected person recovers at rate \( \gamma \) per day. Beaman and Williams' diary data suggest an earworm rarely lasts past a day, so my estimate is \( \gamma \approx 1 \) per day.
- Set the change to zero at equilibrium. \( \beta S/N = \gamma \), so \( S^*/N = \gamma/\beta \).
- Define the reproduction number. \( R_0 = \beta/\gamma \), the number of new heads one infected head reaches before recovering.
- Solve for the share infected. \( I^*/N = 1 - 1/R_0 \).
- Plug in. To keep 99% of India humming at any moment you need \( R_0 = 100 \): with \( \gamma = 1 \)/day, every infected person must pass it to 100 people a day.
- Take the limit. \( I^*/N = 1 \) only when \( R_0 \to \infty \), which means \( \gamma = 0 \). "Everyone, forever" is the same statement as "nobody ever recovers".
Twain's version is even more hopeless. In his story, passing the jingle on cures the person who passes it. Each transmission moves one infection from one head to another, so \( dI/dt = 0 \) and the number of sufferers never grows beyond the first one.
A jingle that leaves you when you share it is a hot potato, not an epidemic.
| Reproduction number \( R_0 \) | Share of people humming it at equilibrium | People reached per infected head per day (at γ = 1/day) |
|---|---|---|
| 1 | 0% | 1 |
| 2 | 50% | 2 |
| 5 | 80% | 5 |
| 10 | 90% | 10 |
| 100 | 99% | 100 |
| ∞ | 100% | no recovery allowed |
Why can't \( \gamma \) be zero? Because of habituation: the nervous system turns down its response to a stimulus that repeats without consequence, which is why you stop hearing a ceiling fan.
The 2025 study by Chris Dodds that links earworms to habit systems suggests the loop is a mental habit, and habits respond to context changes. In the real data the loop always ends.
What would a forever jingle be worth to the brand?
Played nonstop in every Indian head, a 10-second jingle would deliver about 8.4 trillion plays a day, worth about ₹2.1 lakh crore a day at bumper-ad prices, but only around ₹19,000 crore a year of that is useful advertising. Here is the calculation.
The worked valuation
- Plays per head. A 10-second jingle across 16 waking hours: \( 57{,}600\ \text{s} / 10\ \text{s} = 5{,}760 \) plays a day.
- Plays across India. The UN puts India's 2025 population at 1,463,865,525: \( 1.464\times10^{9} \times 5{,}760 = 8.43\times10^{12} \) plays a day.
Calculated in this post from the UN population above and the 5,760 plays a day per head, with these example inputs (a 10-second jingle, 16 waking hours): 1.464 x 10^9 x 5,760 = 8.43 x 10^12.
- Price each play. A six-second YouTube bumper ad, the closest paid format to a sonic hook, costs about ₹250 to ₹400 per thousand impressions in India (Mega Digital's June 2026 guide). At ₹250: \( 8.43\times10^{12}/1000 \times ₹250 = ₹2.11\times10^{12} \), or ₹2.11 lakh crore a day.
Calculated in this post from the 8.43 x 10^12 plays above and the low end of the ₹250 to ₹400 CPM, with these example inputs: 8.43 x 10^12 / 1,000 x ₹250 = ₹2.11 x 10^12.
- Compare with the market. dentsu-e4m measured India's total 2025 ad spend at ₹1,21,339 crore. \( 1.213/2.11 \times 24\ \text{h} \approx 13.8 \) hours. By early afternoon on day one, the jingle has matched a year of all Indian advertising.
- Apply effective frequency. Herbert Krugman argued that three exposures do the psychological work (what is it, what of it, a reminder) and later ones only repeat the third. Media planners cap frequency for this reason. Allow a generous 10 useful exposures a week: \( 1.464\times10^{9} \times 10 \times 52 = 7.6\times10^{11} \) useful plays a year, worth \( ₹1.9\times10^{11} \), about ₹19,000 crore. That is 16% of India's ad market, which is still enormous, but it is 0.007% of the uncapped total.
Calculated in this post from the UN population and the ₹250 CPM above, with these example inputs (10 useful exposures a week, 52 weeks): 1.464 x 10^9 x 10 x 52 = 7.6 x 10^11 plays, x ₹0.25 = ₹1.9 x 10^11.
See the data as a table
| ₹ crore a year (₹ crore) | |
|---|---|
| India total ad spend 2025 | 121339 |
| Forever jingle, useful plays only (10 a week) | 19030 |
| Forever jingle, every play priced (one year) | 76950000 |
Source: dentsu-e4m 2025 ad spend via BestMediaInfo; my calculation from UN population, 5,760 plays a day and a ₹250 bumper CPM · The uncapped figure is a paper value: no advertiser would pay for the 5,750th play of the day.
Sensitivity checks
| Change one assumption | Uncapped paper value | Useful value a year |
|---|---|---|
| Base case: 10 s, cap 10 a week, ₹250 CPM | ₹2.1 lakh crore a day | About ₹19,000 crore |
| Jingle doubled to 20 s | Halves | Unchanged |
| Cap cut to 3 a week (Krugman) | Unchanged | About ₹5,700 crore |
| CPM raised to ₹900 | Rises 3.6× | About ₹68,500 crore |
The useful value depends on the cap, not the loop, which is why the jingle's length does not move it. Every version agrees on one thing: almost all of "forever" is waste.
Use the calculator to check what you just read with your own jingle length, audience and price.
What is a forever jingle worth?
What would it do to people?
A permanent earworm would cost sleep and concentration, measurably. The two best-studied costs are sleep quality and short-term memory.
Sleep
On sleep, Michael Scullin's team at Baylor surveyed 199 adults and found that people who listened to a lot of music reported more nighttime earworms, and that those earworms went with worse sleep. Baylor's summary puts the odds of poor sleep at six times higher for people with nightly earworms.
In a second study, 50 students slept in a lab under polysomnography after hearing either lyrical or instrumental versions of popular songs. The instrumental versions caused more nighttime earworms and worse measured sleep. Participants reported earworms when they woke during the night, which means the sleeping brain keeps rehearsing melodies.
Study 1 surveyed 199 adults (mean age 35.9); Study 2 randomly assigned 50 young adults (mean age 21.2) to lyrical or instrumental music before lab sleep measured by polysomnography; Study 3 analysed EEG. Frequent listeners reported persistent nighttime earworms tied to worse sleep quality, instrumental music raised earworm incidence and worsened measured sleep, and earworm sleepers showed more frontal slow oscillations, a marker of memory reactivation. The limit: the lab sample was young and small, and the six-fold odds figure comes from survey data, which cannot prove cause.
Read the summary and paper link (free)Concentration
On concentration, the irrelevant sound effect is one of the most replicated findings in memory research: background sound with changing notes or syllables lowers serial recall, even when you are told to ignore it. Perham and Vizard found that liked music (Lady Gaga, Rihanna) disrupted recall of consonant lists as much as disliked death metal, against a quiet baseline.
An inner jingle is not an outer sound, but it runs through the same phonological loop, so the safest assumption is that a permanent loop would behave like permanent background music. That is my inference, not a measured result. For more on that buffer, see Mindblog's piece on the working memory limit, and on what nights of disrupted rehearsal do, sleep and memory consolidation.
Would people love or hate the brand?
They would hate it. Repeated exposure raises liking at first and then lowers it, and a forever jingle lives far out on the falling side of that curve.
Robert Zajonc's mere exposure effect says familiarity breeds liking. A 2017 meta-analysis by Anthony Montoya and colleagues re-examined 268 curve estimates from 81 articles and found an inverted U: liking rises with exposures, peaks, then declines.
Advertisers have known the folk version since Thomas Smith's 1885 guide, which claimed that by the seventh sighting a reader is irritated and by the eighth is cursing the ad. Creative wear-out is the paid-media version of the same curve, which is why the Marketblog guide to how often to refresh Meta ad creative exists.
There is also a legal and commercial nuance. A brand that owned a universal jingle would own a sound mark. India registers them: United Breweries' Kingfisher "Oo La La La La Le O" was granted with effect from 31 January 2025 and published in the Trade Marks Journal in October 2025.
The real goal of sonic branding is the version that works: a short, ownable sound repeated at the right frequency, as Lalitaji did for Surf and the Nirma girl did for Nirma in the detergent ad war, and as Bajaj did with Hamara Bajaj.

So, is a forever jingle possible?
No. Three independent laws stop it: habituation ends the loop inside a person, epidemic maths needs zero recovery to reach everyone, and the exposure-liking curve turns extra plays into negative value.
The SIS model shows a tune can occupy every head only if \( \gamma = 0 \), meaning no one ever recovers, and habituation guarantees \( \gamma > 0 \): real earworms rarely survive 24 hours. Even if you forced it, 99.99% of the 5,760 daily plays would be worthless to the brand; the useful part, about 10 plays a week per person, is worth around ₹19,000 crore a year, and the rest buys irritation and, on Scullin's data, sixfold odds of bad sleep. The achievable version is an ownable sonic logo bought at a sensible frequency, which is how every jingle that lasted decades actually did it.
The marketer's lesson is frequency discipline. Budget for the first few exposures per person, then spend on reach, not repetition. India's ad market is now 59% digital, as the breakdown of ad spend by medium shows, and digital lets you cap frequency per person, which radio in 1975 could not.
If your sound needs thousands of plays to work, the sound is the problem.
What's new
- Kingfisher's jingle registered as a sound mark, one of the few in India's register S.S. Rana & Co., Aug 2026
- Accurate singers have more vivid earworms, but not more frequent ones (226 participants) University at Buffalo, May 2025
- Earworms track with habitual behaviour and preference for routine in 883 UK adults PsyPost, May 2025
What if next?
- What if everything cost one rupee? Supply and demand empty India's shelves in about two hours, and the queue eats the discount.
- What if a billboard were a kilometre tall? Readable from 100 km away on paper, invisible in a Delhi winter, and the wind wants it down.
- What if a marketing team ran only on AI agents? The output is infinite; the judgement is not.
Questions people ask
Why do jingles get stuck in your head?
Because they are short, fast and predictable with one surprising leap, which fits the short rehearsal buffer of working memory. Jakubowski's 2017 study found earworm tunes have faster tempos and common melodic shapes, and recent, repeated exposure is the commonest trigger.
How do you get rid of an earworm?
Chewing gum cut intrusions by about 30% in Beaman's 2015 experiments. Listening to the whole song, or doing a moderately demanding verbal task, also helps. Actively suppressing the tune tends to make it come back, as Beaman and Williams found.
Are earworms bad for sleep?
They can be. Scullin's 2021 studies linked nightly earworms to worse sleep quality, and in the lab, instrumental music before bed raised nighttime earworms and worsened measured sleep. Avoiding music in the last stretch before bed is the practical advice.
How many times does an ad need to be heard to work?
Planners still start from Krugman's idea that about three exposures do the psychological work, with frequency caps of a few per week. Beyond that, the mere exposure curve turns down and wear-out begins, so extra plays cost money and goodwill.
Sources
- Jakubowski et al., Psychology of Aesthetics, Creativity, and the Arts: Dissecting an earworm (2017; source for tempo, contour and the 83 melodic features)
- Williamson et al., PLOS ONE: Sticky tunes (2014; source for 90%+ weekly prevalence, survey sizes and the 74.6% humming figure)
- Beaman and Williams, British Journal of Psychology: Earworms (stuck song syndrome) (2010; source for recurrence rarely beyond 24 hours and suppression backfiring)
- Beaman, Powell and Rapley, QJEP: Want to block earworms from conscious awareness? B(u)y gum! (2015; source for the gum experiments)
- Association for Psychological Science: Scullin et al., Bedtime music, involuntary musical imagery, and sleep (2021; source for study sizes and sleep findings)
- Baylor University: Music listening near bedtime disruptive to sleep (2021; source for the six-times poor-sleep figure)
- Montoya et al., Psychological Bulletin: A re-examination of the mere exposure effect (2017; source for the inverted U from 268 curve estimates)
- Perham and Vizard, Applied Cognitive Psychology: Can preference for background music mediate the irrelevant sound effect? (2011; source for liked and disliked music both impairing recall)
- Wikipedia: A Literary Nightmare (source for Twain's 1876 sketch, the jingle's authors and Dawkins's use of it)
- Wikipedia: Effective frequency (source for Krugman's three-exposure argument and Thomas Smith's 1885 sequence)
- Wikipedia: Demographics of India, UN World Population Prospects 2024 (source for the 1,463,865,525 population in 2025)
- Mega Digital: YouTube ads cost in India (June 2026; source for ₹250–400 bumper CPM and ₹500–900 non-skippable CPM)
- BestMediaInfo: dentsu-e4m, India AdEx 2025 and 2026 (Feb 2026; source for ₹1,21,339 crore in 2025)
- Wikipedia: Nirma (source for the 1975 radio and 1982 TV dates of the Nirma jingle)
- S.S. Rana & Co.: Kingfisher jingle registered as a sound mark (Aug 2026; source for the registration dates)
