From the Cadence team
What people actually want to learn, from our own data
We looked at what people typed into Cadence when a course could be about anything. Speaking dominates, almost every goal is one of a kind, and a course built from a learner's own words gets used several times as often as a ready-made one.
We make Cadence, an app that turns a goal into a daily course. That gives us an unusual view of what people want to learn: not what they search for or what a catalogue offers, but what they type when the course can be about anything. This is an early look at a small dataset, so treat it as a first reading and not a market study.
The patterns below come from our own saved learning state as of mid September 2026. We excluded our own testing, automated checks, and records whose origin we could not verify. We describe goals by category and share, and we never publish an individual learner's wording.
The direction
Start from the learner's own words.
Ready-made courses are useful examples and an easy first tap, but the courses people actually use are the ones they described themselves. In our data, a course built from a typed goal was several times as likely to have a finished lesson as a ready-made one. Cadence should keep the learner's own words as the front door and treat the catalogue as inspiration.
What we looked at
Two doors into the same app.
When you open Cadence you can tap a ready-made course, which we wrote and reviewed, or type a goal and let Cadence build a course around it. We compared the two groups: what people chose or typed, and whether they then read lessons.
Learner origin matters because a new app collects a lot of traffic that is not learning. We removed our own test accounts, automated checks, pre-release testing channels, and older records that predate the point where we could verify origin. What remains is a small set of real learners from the app's first weeks with verified origin.
The typed goals are learner-entered text. To classify them we read each one and assigned a category by hand, then counted. We are publishing only the categories and their rounded shares. Where a goal was written in another language we counted the language and nothing else.
Speaking
Almost half of typed goals are about speaking.
Just under half of the typed goals were about speaking or conversation: staying clear under challenge in a meeting, giving a confident update, presenting without freezing, small talk, speaking in a second language.
The ready-made side agrees. Our most-started ready-made course was Speaking Articulately, by a wide margin. The next two were about finding a new role and getting promoted, and together they did not come close to it.
Speaking appears often through both entry points in this early Cadence sample. Our workplace-focused examples and marketing may also shape who arrives, so this does not establish what adults generally most want to learn.
Specificity
Goals arrive as situations.
Nearly every typed goal was different from every other. People rarely typed a subject name. They typed a situation: answering pushback without going blank, recovering after a mistake at work, negotiating in a specific industry, reaching a named level in a language, preparing for a particular exam.
After speaking, the next largest group was work skills at roughly a quarter: productivity, focus, critical thinking, sales, a first go-to-market plan, a specific industry change. Then came pure curiosity at about one in seven, from how blood works to music theory and psychology. Languages and exam preparation were about one in ten, and money, investing and business basics were the smallest group.
A handful of goals were written in a language other than English: German, Dutch, Turkish and Chinese. We have not marketed Cadence outside English, so those learners found the composer and used it in their own language on their own.
Use
People finish what they asked for.
This is the finding that changes our plans. A course built from a typed goal was several times as likely to have at least one finished lesson as a ready-made course, whether we count by course or by learner. Every course that was finished end to end in this period was a typed one.
The two groups started at different times and may differ in intent, app version and how they found Cadence. This comparison does not isolate the effect of typing a goal. Recorded elapsed time can also include pauses, so it is not a direct measure of attention.
Two cautions. A ready-made course costs one tap to start, so many of those starts are browsing rather than intent, and this is a selection effect rather than proof that typed courses teach better. And the sample is small. We will republish when the numbers are larger, including if they move against us.
What changes
What this changes in Cadence.
The composer, the screen where you type your goal, is already the first thing a new learner sees. This data tells us to keep it there and to make the ready-made library work as examples of what a goal can look like instead of as the main menu.
It also tells us where to invest in teaching quality. Speaking under pressure is the most common goal on both doors, so our review effort goes there first: the practice formats, the feedback on what you actually said, and the lessons that rehearse a meeting instead of describing one.
This analysis does not establish whether the speaking pattern is specific to a role, a region, or a career stage. A learner may share a role during course setup, but we have not validated a demographic breakdown for this comparison.
For you
If you are about to learn something, write the situation.
The pattern suggests a useful starting point: describe a moment, not just a field. Instead of public speaking, try staying clear when a senior colleague pushes back in a meeting. Instead of Japanese, try reaching a named level by a date.
That is also good advice without any app. A situation gives you a way to know when you are done, tells you which parts of a subject to skip, and gives you something to rehearse. A subject name gives you a shelf of books.
If you want to try it, download Cadence, type the situation you are facing this month in your own words, and look at the syllabus it builds before you commit to Day 1. If the plan does not match your moment, edit the goal and build it again.
Questions
Common questions.
Where does this data come from?
From Cadence's own saved learning state as of mid September 2026. We excluded our own testing, automated checks, pre-release channels, and older records whose origin we could not verify, and we publish category shares, never counts.
What do people most want to learn in Cadence?
Speaking. Just under half of typed goals were about speaking or conversation, and the most-started ready-made course was Speaking Articulately by a wide margin. Work skills came next at roughly a quarter, then curiosity topics.
Do typed courses work better than ready-made ones?
They get used more. A course built from a typed goal was several times as likely to have a finished lesson as a ready-made one. A ready-made course costs one tap to start, so part of that gap is browsing without real intent, and the sample is small.
Do you publish individual learners' goals?
No. We read each typed goal to assign a category, then published only rounded category shares and the languages used. Individual wording, account identifiers and email addresses are never part of this analysis.
Will you update these numbers?
Yes. This is a first look at a small dataset. We will republish with the same definitions when the numbers are larger, including if the pattern weakens.
See it in concrete terms
Open a real Cadence course.
The three most-started ready-made courses in this data. They are useful as examples of what a goal can look like, and every one of them can be replaced by a course built from your own words.
Organize your point under pressure so updates, questions, and explanations sound clear without sounding rehearsed.
Focus your search, turn real work into proof, and build an application and interview kit for one target role.
Show senior-level judgment through scope, ownership, communication, and visible leverage instead of hoping good work speaks for itself.
Sources
References.
Continue in Cadence
Turn what brought you here into a course you can keep building.
Use the article now. If you want to go further, Cadence can turn the problem or curiosity behind it into a structured course. Cadence is our app.