The Promise Map: What Music Learning Apps Promise Beginners, and What the Data Says Happens

Map 2 of 4. September 2026. By Raul Rodrigues, founder of Upbeat Studio.

Every music learning app is sold to a beginner at the moment of highest motivation. This map puts the promise made at that moment next to the data on what happens afterwards. The argument fits in one sentence: the industry sells the first song and loses the first year. Four numbers carry it, and the rest of this page shows where each one comes from.

• 8 of 20 platforms name a time to your first result, from two minutes to thirty days.
• 2 in 3 people who learned the piano or the drums no longer play it. More than half of those who learned guitar.
• 1 in 4 annual subscribers to an education app renews, the lowest rate of any app category.
• 0 of 20 platforms publish what happens to their learners after the promise.

Map 1 mapped who owns what in this industry. This one is about the beginner. The rules are the same: promises quoted word for word from the platforms' own pages, checked the week of publishing; outcomes only from primary data; vendor statistics labelled as such; years shown on old figures; corrections welcome and credited.

The promise is a time to your first song

I read the public pages and App Store listings of 20 platforms on 16 September 2026 and kept only the wording that names a time to a first result. Eight do. flowkey says two minutes. Simply Guitar, HDpiano and Playground Sessions say day one. Guitareo says an hour. Fender Play says a few short lessons. Pianote says five days. Drumeo says thirty.

Three more say "fast" without naming a time. Two, Singeo and Melodics, promise a daily habit instead: ten minutes a day. That is a different promise, and a more honest one, so it is counted separately. Five make no speed promise at all, and one of them is Yousician, the biggest earner in the category by its own statutory accounts. The leaders do not sell speed. The middle of the market does.

Most people who learn an instrument stop, and they stop young

In 2022 YouGov asked 3,000 American adults which instruments they had ever learned and whether they still played them. For every instrument, most no longer do: 66 percent of those who learned piano, 63 percent for drums, 57 percent for guitar, 72 percent for violin, 77 percent for flute.

The drop happens early. In the UK, the average child who stops playing has stopped by the age of eleven, and 68 percent of those who stop have done so by fourteen. Among sixteen and seventeen year olds, 39 percent used to play and no longer do. A survival study of 3,303 students in the UK and Germany found that about half drop out of music lessons and other musical activities by seventeen, most between fifteen and seventeen. The decision to stay or quit is made in the first years, mostly before adulthood.

Nobody quits because it was too slow

When the UK exam board ABRSM asked children who had given up their instrument why, half said they lost interest, 28 percent said they were too busy, 21 percent said they had stopped taking lessons, 13 percent had no opportunity at school and 12 percent did not own their instrument. Among children who never started, the answers were being too busy, having nobody in the family who plays, cost, and not owning an instrument. Speed does not appear in either list. In the survival study, the strongest single predictor of sticking with music was a musical home environment, not talent or the type of lesson. Every reason on the list is about the year, not the first song.

Beginners already know it takes years

The biggest survey of new guitar players ever run, 10,644 people aged 13 to 64, asked beginners what it would take to get good. Their own answer was four hours a week for one to two years. Half of them practise two hours a week or less. One in four say they have stagnated. Sixteen million Americans started guitar in the two years to mid 2021, the largest cohort of beginners in a generation.

So the promise of minutes is not aimed at what beginners believe. It is aimed at the moment they buy. The gap that matters is not between the promise and the belief. It is between the first week of practice and the second year, where nobody is holding on to them.

The subscription model has already priced in the quitting

Education apps convert a trial at the moment of the promise and collect most of their money as a year paid up front. In the industry's own benchmark of 115,000 apps, the median education app converts 6.5 percent of trials to paid. Between 59 and 66 percent of the plan mix is annual. Then education renews fewer of those annual plans than any other category: 24 percent. Weekly plans renew at 58 percent, the highest of any category, which tells you the model is built around the first weeks. About five percent of churned annual subscribers ever come back.

A piano app published its own cohort data, which is rare and deserves credit. Across 1.1 million Skoove users, December starters retained 28 percent below the average six months later and January starters 21 percent below, while May starters retained 23 percent above. A year of revenue, collected at peak motivation, from a customer the model expects to lose.

What keeps people playing is known, and none of it is speed

The research on why learners persist is older than the apps and points the same way each time. Tutoring that guides each step is almost as effective as a human tutor: in the standard review of tutoring experiments, step-based systems reach an effect size of 0.76 against 0.79 for humans, while systems that only grade the final answer reach about 0.3. Two thirds of children and 60 percent of adults who play say starting was their own decision. The strongest predictor of persistence is someone at home who plays. People who start when they choose, rather than on a resolution, stay longer. And 84 percent of children who play do so at least weekly; one in eight play every day. Feedback, a reason, a habit. These are design decisions, and a product can be built around them.

Not one of the 20 platforms publishes what happens after the promise

I looked for any learner outcome or retention figure on the marketing pages and store listings of all 20 platforms. There is none. What gets published is volume: 180 million downloads for one tuner, 20 million monthly "users" with the tuner included, subscribers when the number flatters, ratings and valuations.

The one famous number about quitting, that 90 percent of first-time players abandon the guitar within a year, comes from a single vendor. It is Fender's internal finding from 2015, made public in 2018, hedged in the same sentence ("if not the first 90 days") and never published with a method. Fender built an app to fix it and has not disclosed a paying-subscriber figure since its 2020 free promotion. The industry measures acquisition. The first year, where it loses most of its customers, has no owner and no number.

Three things this map says

1. Speed is the wrong promise. The first year is the product, and the evidence on how to keep people playing already exists.
2. Retention is nobody's KPI, so it is everybody's cost: the apps that lose three in four subscribers, the makers whose instruments go quiet, the learners who regret quitting.
3. Whoever is paid to keep a beginner playing owns the next decade of this industry. That is where I am building.

Learning to play should not be intimidating. That is the problem I build against at Upbeat Studio with Drum Coach. Next map, in October: the micro-app flood, every music learning tool launched in the last 18 months and who is behind it.

Download the map: The Promise Map, PDF, 10 pages.

Sources, slide by slide

Slide 3: the promise

Twenty platforms checked on 16 September 2026: Skoove, flowkey, Simply Piano, Guitar Tricks, Piano in 21 Days, Pianote, HDpiano, Synthesia, Fender Play, Duolingo Music, Yousician, Simply Guitar, Guitareo, Drumeo, Singeo, PlayBass, JustinGuitar, Playground Sessions, Melodics, tonebase. Where a website blocked automated reading (Guitar Tricks, Fender Play, JustinGuitar, Melodics, Playground Sessions and Simply Piano's speed wording), the platform's own App Store description was used. Two platforms could not be reached (Piano in 21 Days returned an error; PlayBass redirects to a login page).

• Names a time to a first result (8): flowkey, "Get set up in just two minutes and start playing your first songs straight away." flowkey.com. Simply Guitar, "Play songs from day one". hellosimply.com/simply-guitar. HDpiano, "Play your favorite songs from day one". hdpiano.com. Playground Sessions, "hooked and having fun from day one" (App Store listing). Guitareo, "Learn to Solo In An Hour", "Song In An Hour Challenge". guitareo.com. Fender Play, "Begin playing in a few short lessons with learning, practice and feedback activities." (App Store listing). Pianote, "5 Days To Playing Piano". pianote.com. Drumeo, "30-Day Independence", "30-Day Drummer". drumeo.com.
• Says "fast" without a time (3): Simply Piano, "Personalized 5-Min Workouts ensuring you progress fast and always succeed" (App Store listing; the website now reads "progress at your own pace"). Guitar Tricks, "members reported learning 50-100% FASTER than with other guitar apps!" (App Store listing). JustinGuitar Note Trainer, "learn the fretboard fast, in just 5 minutes a day." (App Store listing).
• Promises a daily habit (2): Singeo, "Improve your skills in just 10 minutes a day." singeo.com. Melodics, "Just 10 minutes a day can lead to serious improvement." (App Store listing).
• No speed promise found (5): Yousician, Skoove, Synthesia, Duolingo Music, tonebase ("Finally play the music and pieces you love.").
• Yousician's revenue: EUR 53.0M for 2025 per the Finnish statutory registry, via Finder.fi.

Slide 4: what happens

• "Do you still play the following instrument(s)?", asked of those who learned each one. No: piano 66 percent, electric piano 57, guitar 57, electric guitar 49, drums 63, violin 72, flute 77, clarinet 86, trumpet 73, saxophone 79. YouGov, "Musical Instrument Experiences", 3,000 US adult citizens, 2 June to 4 July 2022, toplines document. yougov.com and toplines PDF.
• Children who stop playing do so on average by eleven; adults report stopping at eighteen; 68 percent of children who stopped did so by fourteen; 47 percent of 16 to 17 year olds currently play, 39 percent used to play. ABRSM, Making Music Report 2021, page 23 (2020 fieldwork; 1,503 children, 1,505 adults). abrsm.org (PDF).
• About 50 percent of students drop out of music lessons and other musical activities by 17, most between 15 and 17. Ruth and Müllensiefen, "Survival of musical activities", PLOS One, 2021; 3,303 students, UK and Germany, 2015 to 2019. journals.plos.org.

Slide 5: why they stop

• Top five reasons children gave for giving up their instrument (239 children who no longer play): lost interest 50 percent, too busy 28, stopped taking music lessons 21, no opportunity to learn at school 13, did not own the instrument 12. Top reasons for never learning (370 children): too busy 36, not from a musical family 26, cost 25, no instrument of their own 23, not interested 21. ABRSM 2021, page 53, figures 51 and 52.
• Home environment as the strongest predictor: PLOS One 2021, above.

Slide 6: what beginners expect

• 7 percent of Americans aged 13 to 64, about 16 million people, started guitar in the last two years; 72 percent of new players are 13 to 34; 62 percent cite the lockdowns; 53 percent of beginners practise two hours or less a week; beginners believe four hours a week for one to two years is what it takes to "get good"; one in four have stagnated. Fender "New Player Landscape" study, fieldwork by YouGov, 8 May to 13 June 2021, 10,644 US respondents. Commissioned by Fender. PRNewswire, October 2021.

Slide 7: the model

• Education apps: trial-to-paid conversion median 6.5 percent, top quartile above 11.6 percent; Education, Travel and Shopping favor annual plans at 59 to 66 percent of the plan mix; annual renewal 24 percent ("a firm last for annual") against 58 percent first weekly renewal (the highest category); about 30 percent of education apps' annual-plan cancellations happen in month one; churned annual subscribers reactivate at about 5 percent across categories. RevenueCat, State of Subscription Apps 2026, 115,000+ apps, 2025 data. revenuecat.com and renewal rates by category.
• December starters retain 28 percent below the cross-year average over six months, January 21 percent below, May 23 percent above; 1,137,446 Skoove users. Skoove with DataPulse Research, 2026. First-party, single app, not peer reviewed. skoove.com and datapulse.de.

Slide 8: what works

• Effect sizes against no tutoring: human tutoring 0.79; step-based intelligent tutoring systems 0.76; answer-based tutoring systems about 0.3. VanLehn, "The Relative Effectiveness of Human Tutoring, Intelligent Tutoring Systems, and Other Tutoring Systems", Educational Psychologist 46(4), 2011. ERIC EJ946764, DOI 10.1080/00461520.2011.611369.
• 66 percent of children and 60 percent of adults who have played say it was their decision to start (1,120 children, 949 adults); 84 percent of children and 59 percent of adults who can play do so at least once a week; one in eight children play every day. ABRSM 2021, pages 22 and 53.
• Home environment: PLOS One 2021, above. Starting month: Skoove and DataPulse 2026, above.

Slide 9: nobody measures it

• None of the 20 platforms publishes a learner outcome or retention metric on its marketing pages or store listing, as of 16 September 2026.
• GuitarTuna: 180 million downloads, self-reported. guitartuna.com/about. Yousician's "20 million" monthly users includes the tuner. Musora: 100,000+ paid subscribers, Globe and Mail, 2025. Simply: 1 billion dollar valuation, NoCamels, 2021.
• The 90 percent: Fender press release, 16 October 2018, quoting CEO Andy Mooney on a 2015 internal finding, "in 12 months, if not the first 90 days". spotlight.fender.com. Fender Play subscriber disclosures: about 110,000 paying before COVID and about 930,000 at the 2020 free-promo peak, Forbes, March 2022; no paying figure disclosed since, checked through August 2026.

Slide 10

• Drum Coach: 200,000+ drummers across 180+ countries. upbeat.studio, 2026.

Corrections log

• 16 September 2026, first revision on the day of publication: the map was rebuilt around one argument. The first version gave a whole slide to the Fender story and counted "ten minutes a day" as a speed promise; both were wrong weightings. Fender is now one evidence row, habit promises are counted separately, and four datasets were added (YouGov per-instrument "still play" figures, ABRSM reasons for stopping, the Fender and YouGov New Player Landscape survey, VanLehn's tutoring review).

If you find a figure that is wrong, outdated or missing its context, message me on LinkedIn. Verified corrections are folded into the map and credited here.

About the author

I am Raul Rodrigues, a drummer, a former music-school founder and the founder of Upbeat Studio in Valencia, where we build Drum Coach and Drum Notes. I map the music learning industry because nobody else publishes it, and because the thing I care about most, keeping beginners playing through their first year, is the thing this industry measures least. Find me on LinkedIn.