Nobody Passes All Five

I've been watching the machine studies come in for about a week now, and one thing has become hard to ignore: nobody passes all five tests.

Fifteen companies have been read so far, everything from Meta and Shopify to Home Depot, the NFL and DocuSign. That's 75 results. Nine of them are "met." Not one company met all five, and the best any of them managed was two.

The company with two is this one. When realprosperity.org was read against its own tests, it met two, partly met one and failed two. I built the site to hold companies to these tests, and it doesn't pass them yet either. The studies I did by hand of my own products find at least one test unmet in every one of them.

Where everyone lands

The most common result by far is "partly." Fifty-one of the 75 results land there. Most companies let you export something, but not everything. Prices usually appear without saying what one unit costs. Privacy policies tend to give examples of what's collected instead of a list. And the records of how a company treats you are almost always its own, with no copy held by anyone outside.

Partly is the honest middle, and it's a decent place to start. A company that's partly there has already done the hard part of admitting the thing matters.

The test almost everyone fails

The fifth test is different. It asks whether the principles live in the charter, where a future owner can't quietly remove them. Ten of the fifteen companies failed it outright, and only one met it.

It's the hardest test because it costs something real. Putting a promise on a privacy page is cheap, and anyone who buys the company later can rewrite the page. Writing it into the charter means giving up the right to change your mind, for you and for whoever comes after you. Most founders, me included, haven't done it yet.

I've started to think this test decides whether the other four last. A company can get every other test right today and lose all of it in one acquisition.

What I make of it

When I wrote the five tests, I half expected a few companies to pass and the rest to fail. That's not what happened, and I've had to rethink what the tests are for.

They work better as a direction than as a verdict. Nobody is finished, so the useful question about a company is which way it's moving. A company that goes from partly to met on one test has done something real, and the study can show it. Every earlier read stays published, so the change is there for anyone to see.

That also changes what I'm asking companies to do. I've stopped waiting for one that passes everything. The ones worth finding are willing to fix one thing and be read again.

What I'm doing about my own

This site failed accountability and structure. Accountability means someone other than me should hold the evidence this site is judged by, and right now I hold all of it. Structure means the principles need to live somewhere a future owner can't quietly change them, and right now they live on a web page I can edit. Both are on my list, and when I've fixed either one, I'll have the site read again and the new study will link to this one.

If you run a company that's been read, pick the test you're closest on and fix it. Tell me when you have, and I'll make sure it gets read again.