If you're running version 1 of my survey tool, here is the reason to move.
Version 1 gave every respondent one number and a label. Version 2 scores each part separately, and tells them which one to focus on next.
That sounds like a feature. It's closer to a change of purpose, and it's taken me a year to be able to build it.
What one number can and can't do
A composite score is a verdict. It ranks you against a scale and stops there.
For the room, that's genuinely useful: a distribution across an audience says something real, and it says it in one glance from the back of a hall. For the person who just answered twenty questions about themselves, it's a dead end. They know where they stand. They have no idea what to do about it.
Which matters, because the whole point of handing an audience an assessment is what happens after. If the result closes the subject, you've collected email addresses. If it opens one, you've started a conversation with something concrete already on the table.
Version 2 keeps the composite if you want it. But it also reads each facet separately (each distinct part of what you're measuring), reports it on its own scale, and surfaces a single next move. So the result stops being a rank and becomes a direction.
There's a lot more in version 2, and I'll come back to the mechanics properly in a few days. This is the part that changes what the tool is for.
How it started
In April 2025 I was at a mastermind event in New Orleans with Carolyn Strauss.
The new idea Carolyn got from the room was to build a survey: a lead magnet made out of the IP she already had.
She loved it immediately, and was overwhelmed about ten seconds later.
The idea was good and the next step was invisible. Carolyn is not a technologist, and said so that evening when I offered to help: whatever this turns out to be, it has to be very low-tech or it won't work.
I took that as the design brief rather than a limitation.
Built out of what she already owned
Before building anything, look hard at what someone already has. Carolyn already had Google Workspace, so the whole thing could run on tools she was already paying for and already knew how to log into.
The design was deliberately unremarkable. A QR code on the screen at the start of her talk, pointing at a Google Form. Responses land in a Google Sheet. On another tab, a pie chart wired to redraw itself as each submission arrives.
No accounts to create. No platform to learn. Nothing to maintain.
Note what that instrument was for. The chart on the screen is a picture of the audience. Version 1 was built to read a room, and it was good at it.
Fifty-four responses, and then a phone call
She ran the Irresistibility Index for the first time that April. Fifty-four responses, and the chart filling in on screen while the room watched.
Two days later she called and asked whether she could do the same thing later that week with a room of twenty executives.
That question turned out to be the interesting one. Her instrument was the same; her audience wasn't. Most tools solve this by making every respondent type a campaign code: friction paid by the wrong person, in the middle of a live event, to solve a problem the presenter has.
So instead I taught the chart builder to accept a range of rows. One instrument, any number of events, nothing extra for anyone to type. She can chart one group on its own, or everyone who's ever taken it, or a single quarter. The audience does nothing different.
Then it spread. South Africa in May, a mostly American group in June. Carolyn has run it ever since, on her own, mostly without further help from me, and has built a family of instruments off the back of it. This year she shared it with the Washington DC chapter of the National Speakers Association.
Two things I got wrong
It needed me. Read the feedback from those users and the same compliment keeps appearing in different handwriting: patient, good explainer, broke it into steps I could follow, was there when I got stuck.
The feedback was kind and appreciated. It was also a defect report.
If people are thanking you for being patient, what they're describing is a product that required patience, and a person standing in for the part that should have been obvious.
Concretely: to customize the result email, version 1 asked you to open the Apps Script editor, delete what was there, paste in a replacement, and not break the syntax, because a syntax error blocks saving, silently, at the worst possible moment. I documented that beautifully. Screenshots and everything. It was a well-illustrated apology.
And if you tried to make the form your own by adding a question, the formulas in the Sheet quietly stopped matching. The tool punished exactly the instinct you want a user to have.
The score was too general. This was the other criticism version 1 got, and it wasn't about the software at all. One number for a whole person.
The complaint was fair. What I didn't have was a small answer to it. Version 1 was built to read a room: that is what the chart on the screen was for, and against that job a single score does fine. Reading one person properly is a different instrument, not a patch on this one. It needed the next version.
Version 2
You never open the code. Not once, not for anything. There is no script editor step, because the wording now lives on a config sheet and is read fresh every time an email goes out, so an edit takes effect immediately, with no deploy and no held breath.
This was the barrier. I'm fairly certain it cost more adoption than every other problem combined, and it's simply gone.
Good software shouldn't need the person who built it. It should just work.
Every facet is scored on its own. Which is where this article started, and where the year went.
What hasn't changed
It's still free and open source. It runs entirely inside your own Google Drive, which means your respondents' answers sit with you and I cannot see them. Not "we don't look." Cannot. There is no server of mine in the middle, because there is no server of mine at all.
The beta
I'm looking for people who'll use it and tell me where it breaks. Not applaud it. Break it. Where did you get stuck, what did you have to guess, and at what point did you consider messaging me? That last moment, whatever caused it, is a defect. coroin.app/#survey
There's a live demo at backstage.coroin.app: the sample assessment that ships with the tool, running version 2. Take it, see what lands in your inbox.