Scan time: 2-4 minutes / Read time: 4-6 minutes
Howdy rebel ideapreneurs 🦸♂️🦸♀️
🦉 The company that paid to be tested
Duolingo put a link on its own website asking Spanish learners to come and be tested.
It was 2012, and the app had opened to the public a few months earlier.
Duolingo is a free language-learning app.
Nobody knew whether it actually worked, and that included Duolingo.
Every language app said the same three things about itself.
Fun.
Effective.
The best way to learn.
Not one of those is something a person can go and check.
So Duolingo paid two university statisticians to measure it instead.
Roumen Vesselinov taught statistics at Queens College in New York.
John Grego ran the statistics department at the University of South Carolina.
They took a hundred and ninety-six volunteers who had never studied Spanish anywhere before, so everybody started from zero.
Anyone who finished all eight weeks got a twenty-dollar Amazon gift certificate.
Each volunteer sat a Spanish test on day one.
The test was WebCAPE, built at Brigham Young University in Utah.
American colleges already used it to work out how much of a language a new student knew, and put them in the right class.
So Duolingo never touched the scoring.
Then those people studied Spanish on Duolingo and nothing else for eight weeks.
Eighty-eight of them came back and sat the same test again.
The other hundred and eight never came back, and the report says so.
Roumen and John wrote up what the two scores showed, in a report anybody can read.
Get this — the first page says Duolingo paid for the study.
It also says the two of them collected and analyzed the data on their own.
The finding was that somebody starting with no Spanish needed between twenty-six and forty-nine hours on the app to cover a first college semester.
Thirty-four hours on average.
Turns out nobody can check whether an app is fun.
Anybody can check whether it took them thirty-four hours.
Duolingo has led with thirty-four hours ever since.
⛳️ The problem
Duolingo paid two outsiders to put a number on how well its app actually taught Spanish.
But right now, we are writing the line that describes our newsletter, and we reach for the same three words every time.
Thorough, practical, genuinely useful.
They feel true to us, so we use them, and so does every other newsletter on the topic.
So a reader cannot check a single one of them, and next week's issue says the same three things again.
Duolingo stopped using words like that and started using a number anyone could check: thirty-four hours.
There is one measurement that does more for you than any of those words, and you probably already have the data for it.
A nurse used hers, one number from the people she had taught, to sell $1 million of courses in seven months, so meet Sarah.
🩺 The tooth that cost seven thousand dollars
Sarah Michelle Boes was a nurse in Kentucky, in the United States, and she had just passed her board exams.
Passing those exams gives a nurse the license to diagnose patients and prescribe medicine.
Months of studying had left her grinding her teeth in her sleep, hard enough that she broke one of them.
She needed a dental implant, and it was going to cost seven thousand dollars.
"I figured if I put together a little review course and even just earned $500, that would help pay for it."
Five hundred dollars was the whole plan.
So in May 2020, about a week after her own exam, she opened Zoom and recorded a three-hour crash course.
She had tutored nursing students before, so she built it as questions and answers instead of a lecture.
Then she gave it away.
She already ran a Facebook group where nurses studying for the same exam traded questions.
So she sent the course free over Facebook Messenger to the thirty classmates in her cohort, and asked one thing back.
"I'll give you this course for free if you invite someone else into the group when I give it to you."
Every new person brought another new person, so the group kept growing.
She gave the course away free like that until a hundred nurses had it.
Then she started charging twenty-five dollars for lifetime access, collected through Venmo and Cash App.
Her first buyer got angry that a friend had gotten it free, and haggled her down to fifteen.
She took a thousand dollars on her first day of selling.
Then the exam results started coming back.
A nurse would pass the boards, post the photo of her passing score in the group, and say what had helped.
Nobody asked her to post it.
She wasn't the only one.
Week after week, more nurses posted that they had taken her course and passed.
Nobody had to argue that the course worked.
Anyone could scroll and watch it working.
Three months in, sales passed a hundred thousand dollars.
Seven months in, a million.
She had called the business Sarah Michelle NP Reviews, and in June 2022 the test-prep company Blueprint Prep bought it in an eight-figure deal.
Go to her website today, and the first thing on it is not about her.
It's one number: over ninety-nine percent of her students pass.
Not her own exam score, not the hours of video, not how friendly the group is.
A nurse shopping for a course hasn't taken the exam yet.
So she has no way of knowing whether the teaching is any good.
But she knows exactly what ninety-nine percent means, and she buys the course.
🔥 The recipe
➡️ Lead With the Documented Outcome
Duolingo paid two statisticians to measure how much Spanish its app actually taught.
The answer was thirty-four hours to a college semester, and Duolingo has led with that number ever since.
Sarah Michelle's students posted their results in her group, and she leads with their pass rate instead of her credentials.
This hack is called Lead With the Documented Outcome.
It means measuring what happened to the people you helped, and putting that number in front of everything else.
Documented just means somebody counted it.
It doesn't have to be a paid study.
Duolingo paid outsiders for its number.
Sarah Michelle added hers up herself, for nothing.
Neither of them asked anyone to take their word for it.
They both went and observed what happened to the people they taught.
Then they put that number first, and said nothing else about themselves.
So the next time you write the line that describes your newsletter, don't call it thorough and practical.
Report what happened to the readers who read it.
Someone who can't judge your expertise can still judge a number.
Go and count your thirty-four hours.
That's it, my fellow renegades!
Yours 'helping you gain years of experience without putting in the years' Vijay Peduru 🦸♂️
🏄♀️ The prompt
The prompt below finds the one outcome you could measure and turns it into the line you lead with.
Save this prompt as a Skill or add to Project in your favorite AI tool — build once, use often.
Run it as is. Your AI sidekick will use your own files if it can see them, and the example values if it cannot.
CONTEXT:
- Use my own material first. If you can see my drafts, my sent issues, or my subscriber dashboard, pull the real details from there instead of the examples below.
- If my Voice Profile exists, write in that voice. Otherwise, write in a clear, warm, no-jargon voice — first-person, plain words, no startup-speak.
- If my ICP / Audience doc exists, target that reader. Otherwise assume finance analysts sitting CFA Level 1 while working.
- Anything I leave as an example, treat as real. Anything I delete, fill in from what you already know about me, and mark it as your assumption.
- Adjectives are disqualified. Every claim must be a number I could actually collect, and you must say how I would collect it. Never invent a figure — if I do not have one yet, tell me how to get it.
=============== INPUTS — edit these, or just run as is ===============
What I make, for whom, and how I currently describe it:
{e.g. a newsletter on passing the CFA Level 1 for people working full time, described as "clear, practical, built by someone who has been there"}
What actually happens to people who use it, and how many have been through it:
{e.g. a lot of them pass but I have never asked systematically, and maybe 300 people}
What I could ask them in one email:
{e.g. did you pass, and how many hours did you put in}
======================== END OF INPUTS ========================
Outputs:
1. My adjectives — every unmeasurable word in my current description, listed, with what each one is standing in for.
2. The one number — the single outcome worth measuring, and why that one rather than the others.
3. How to collect it — the exact email or form to send this week, written out, that gets me the figure.
4. What to do while I wait — the honest interim line I can use that is true today.
5. Where it goes — the three places the number replaces an adjective once I have it.
Then tell me the number I am afraid to collect, and whether that fear is telling me something.