Every person on Bob's list has their own three-page PDF report with their real scores in all six areas, sitting at a private URL on bobmanthy.com. The email that carries it is written and personalized to how each person actually scored.

Nothing has been sent. This is still for you and Bob to read first.

What changed on August 6

Bob sent two rounds of edits and a shorter list. Both are applied and live.

His six wording changes, applied everywhere they appear:

The "Rebuilding" swap was done by hand rather than by find and replace, because a blanket replace would have produced "Rebuilding Seminars Seminars" in the places that already said it, and would have rewritten web addresses. Only the words a reader actually sees were changed.

His shorter list. Bob re-sorted the shared sheet himself, took out the remote leads and the do-not-contact people, and left 58. Annette's own record came off it on August 7, so it is 57 now. That list is what drives everything, instead of the tag rules I was using. If someone isn't on Bob's list, no report gets built and no email goes out.

Every one of the 57, on one page

Open the full list of all 57

Everyone on Bob's list, lowest score first, with each person's score, band, and subject line, and a link to read their email and their PDF. You can search a name, and filter to the nine who were previously held. Their rows are tinted amber, since the group column came off on August 7.

That page lives at an unguessable address inside the reports folder, so it carries the same noindex and no-store protection as the reports themselves. The address is the only thing protecting it, so it's fine to send to Bob and not fine to paste anywhere public.

See a few of them side by side

If you'd rather just spot-check, here are four spanning the whole range. The lowest and the highest side by side are the point: the old automated email told a perfectly well adjusted person that "we understand this time can be very challenging."

The big image at the top of each email is page one of that person's own report, showing their own numbers. That's what earns the click. Clicking it opens the full PDF.

These four sample links are stable now. They used to carry the sample person's first name, which meant they changed every time the list changed and broke whatever link Bob had open.

Who gets what

The 9 who were held are now released. They're the people who answered the suicidal thoughts question with Sometimes or Almost Always. From August 6 to August 8 they were kept out of the campaign entirely, in a separate file, so that no import could reach them. On August 8 Bob decided they should be emailed after all, and since he's the licensed counselor that's his call to make. They're now ordinary rows in the send file alongside everyone else.

The history is kept, not erased. That matters because this was a clinical decision that got reversed, and a future reader shouldn't have to dig through git to find out why nine people were treated differently for two days:

• The roster carries a Previously held column recording why each was held and when they were released.
• previously-held.csv lists all nine with their answers, whether Bob already wrote to them personally, and links to their report and draft. It's a record, not an import file.
• Each of the nine has an amber note on their own draft page saying they were held from August 6 to 8, what they answered, and whether Bob already replied. Bob sees that context at the moment he's reading their email, without opening a spreadsheet.
• On the all-57 list, their rows are tinted amber and there's a "Previously held" filter.

Two of the nine already got a personal reply from Bob on August 5, so they'd be receiving this on top of that. Worth him knowing before it goes.

The two-file split that used to protect them is gone, because they're no longer held and keeping a second import file around would just invite a mistake. If anyone is ever held again, the build keeps them out of ghl-import.csv automatically and writes the separate file back.

The 51 who are off the list are mostly Bob's own call: he removed the remote leads and the do-not-contact contacts when he re-sorted the sheet. Annette's own record is the fifty-first, removed on August 7.

Of the 57 who'd be emailed: 19 scored under 40, 19 scored 40 to 59, 12 scored 60 to 79, and 7 scored 80 or above. The copy changes by band, which is the whole point of doing this.

Four things worth your attention

Annette's own record was on the send list, and is now removed. She had taken the test on March 31, 2026, scored 14, and was tagged Boulder area and consult-interest, so she was queued to receive the under-40 email, the one that opens "You are still raw." Removed from the roster on August 7 at her instruction. Her report and draft email are deleted, locally and from the site, and the roster went from 58 to 57.

The remote lead problem solved itself. Yesterday 29 of the people on the list were out of state, and their copy had to avoid promising them a class near them. Bob removed all of them, so nobody on the current list is a remote lead. The open question about whether a referral network exists outside Colorado no longer blocks this send. It's still worth an answer eventually.

One of the nine still doesn't match Bob's stated policy. He wrote that he replies to Almost Always and Usually and deliberately doesn't reply to Sometimes. Two of the nine have his reply date of August 5 against them. But one person who answered Almost Always is marked "n/a" in the reply column, meaning no reply. He also mentioned replying to three people by name, and only two have dates. That's the question you asked him on August 6 and he hasn't answered. Releasing all nine doesn't settle it: it's still worth knowing whether that person got a personal reply, because they'd now be receiving a campaign email either way.

Bob said he's not finished. His words were that the letters are "very good, but I'll make them sound a little less AI-like." So more edits are probably coming, and it's worth waiting for them before importing anything into GHL.

Privacy

The reports contain named people's psychological scores, so the URLs are long random strings that can't be guessed, and guessing one returns a 404. Robots.txt blocks that whole folder for every crawler individually, including the AI crawlers, and the server sends a noindex header on every file.

The 33 reports belonging to people Bob removed have been deleted from the site. Nothing was ever emailed, so no link is sitting in anyone's inbox. One loose end: Cloudflare is still serving a cached copy of those files to anyone who has the exact address, and clearing that needs a permission your API key doesn't currently have. Nobody outside the three of us has ever had those addresses, so this is tidiness rather than exposure, but it should be cleared.

What I need before anything sends

1. Bob's final wording. He has more edits coming. These are worth waiting for.
2. Bob has released the 9, and that's done. Worth him seeing once more before it goes: what each of them receives is the standard campaign email for their score band. It's honest and it carries the 988 crisis line, but it is still a marketing email, and two of them already got a personal note from him on August 5. Their draft pages carry that context. Still unanswered from your August 6 question: why one person who answered Almost Always shows no reply sent.
3. The GHL step. The 57 rows in ghl-import.csv need importing as custom fields, and the email template needs pasting in as a raw HTML block. GHL can't attach a per-person PDF, which is why the report is a private link behind the preview image. That's better for deliverability anyway.
4. A test send to you first. If a merge field is empty the email looks broken, and that's much better to find on your own inbox than on 57 of Bob's leads.