User feedback

I asked users what's wrong with it.

I emailed every subscriber who worked more than 100 questions in September the same six questions. Every reply is below, complete and unedited, in the order asked, including the typos. If more come in they get added here with the date received.

None of them would give a full name. So you cannot verify any of this, and you should treat it accordingly. I am publishing it because the alternative is publishing nothing, and because several of them told me something I did not want to hear.

The status under each response says what I am doing about it, including the one I am refusing. Come back and check whether the rest happened.

What I sent, verbatim

  1. What did you find most useful?
  2. What frustrated you?
  3. What would prevent you from recommending it?
  4. What other course are you using?
  5. May we publish your exact response?
  6. May we identify you fully, partially or anonymously?

Questions three and two are there because I wanted the answers to them. A survey built to harvest praise does not ask what would stop you recommending it.

The responses

RESPONDENT 1 · NO OTHER COURSE · PUBLISH: YES, NO PERSONAL INFO · IDENTIFY: NO

What did you find most useful?

Its fine

What frustrated you?

Getting to 2,000 questions feels impossible. I’m over 700 in and still have so far to go.

What would prevent you from recommending it?

The explanations can actually be too much. Sometimes I just want the answer key and to move on to the next, rather than 5 more minutes of rabbit hole

What other course are you using?

None now

Already doing itThis is the worst thing anyone said, and it came from the one person using no other course. Word counts were already coming down before this arrived and the passes are logged in the changelog. This told me to keep going.

Not changingThe 2,000-question target stays. It is not supposed to come easy, and somebody at 700 who feels the distance is feeling the right thing. I would rather show a number I believe in than one that flatters the dashboard. This is a real disagreement with a paying user and I am leaving it on the page as one.
RESPONDENT 2 · ALSO USING BECKER · PUBLISH: YES · IDENTIFY: ANONYMOUS

What did you find most useful?

The questions are totally different than Becker.

What frustrated you?

I have to upload my tracking between phone and laptop, I dont think the login communicates between devices or something. It works, just an extra step.

What would prevent you from recommending it?

Nothing, it's such a good value. Maybe if someone learned better from the Becker lectures, I wouldn't recommend this.

What other course are you using?

Becker

FixingProgress should sync automatically between devices. Reproducing the “manual upload” path you hit; if confirmed it goes away. Sign-in will refresh your latest work across phone and laptop without you doing anything.
RESPONDENT 3 · ALSO USING NINJA · PUBLISH: YES · IDENTIFY: NOT FULLY

What did you find most useful?

Teach back

What frustrated you?

Can you make it so my mock exam results get emailed to me?

What would prevent you from recommending it?

If I end up failing.

What other course are you using?

Ninja

QueuedEmailed mock results make sense and the mailer already exists for progress emails. Adding this to the next batch.
RESPONDENT 4 · ALSO USING GLEIM · PUBLISH: YES · IDENTIFY: PARTIAL

What did you find most useful?

On the sims theres a helper for how to approach if youre stuck, I like that.

What frustrated you?

Some of the questions have a lot of words, sometimes important, sometimes filler. Also seems clunky on mobile.

What would prevent you from recommending it?

idk

What other course are you using?

Gleim

MeasuringLong stems are a defect, not a style. Tracking stem vs explanation length, reading the worst decile, cutting filler.

Update, October 10We rewrote 9,218 multiple-choice questions: stems are 16% shorter, answer options 53% shorter. See the October 10 entry in the changelog.

FixingMobile clunk: tightening touch targets and scroll behavior.
RESPONDENT 5 · NO OTHER COURSE · PUBLISH: YES · IDENTIFY: NO

What did you find most useful?

nothing

What frustrated you?

unhelpful questions, your questions need to be a lot clearer. no one can understand these. not helpful for passing.

What would prevent you from recommending it?

dont recommend

What other course are you using?

nothing

RESPONDENT 6 · NO OTHER COURSE CURRENTLY · PUBLISH: YES · IDENTIFY: PARTIAL

What did you find most useful?

Teach-back thing. Whenever I thought I was good it put me in my place lol.

What frustrated you?

You guys need some short videos for topics that need more. Even a few brief ones would help.

What would prevent you from recommending it?

What other course are you using?

None currently. But I tested several others.

ThinkingNo lectures planned, but 60–120 second micro-explainers for tricky topics may help. I’ll test a few and keep them only if they move results.
RESPONDENT 7 · ALSO USING BECKER · PUBLISH: YES · IDENTIFY: NO

What did you find most useful?

Good on phone.

What frustrated you?

Nothing yet.

What would prevent you from recommending it?

Don't know.

What other course are you using?

Becker.

Not changingNothing to do here, but thank you.

What this is not

These responses are not a sample. They are the people willing to reply, at one point in time. Most of them also use another course.

It is not evidence the product works. There is still no published pass rate, because no outcome data has been collected, here or anywhere else in this category.

And it is not verifiable. Nobody agreed to a full name, so every word above could have been written by me. The only thing I can offer against that is the changelog: if the things listed above show up there on the dates they show up, that is at least consistent with real people having asked. If I were inventing users I would not have invented one I then publicly disagree with.

Asked in early September 2026 of every subscriber who had worked more than 100 questions that month. Latest response received 14 September 2026. The survey is still open; any further response gets added above with its date. Responses reproduced in full.