The public funnel, walked.

1. The quiz door is the barest page in the funnel.

Every CTA on every page lands on welcome.mindbloom.com, and it is the only captured page with zero proof elements: no stat, no testimonial, no time estimate, no "no payment info required," nothing but a warm sentence and a Start button above a legal acknowledgment. The visitor is asked to commit at the exact moment the site goes quiet.

The test: one reassurance line under the button. Time estimate, no-commitment language, and what happens after. Cheap to ship, measured in start rate, and my first A/B as a lifecycle hire.

2. Cost never appears unprompted, and the first number shown is the hardest one to get.

No dollar figure exists on the homepage, the program explainer, the science page, or the reviews page. A visitor meets cost only by seeking out pricing, where the headline "as low as $165/session" requires the 18-session, nine-month commitment; the actual entry program is $215 a session. Meanwhile the homepage's insurance-logo strip with "reimburse over 50%" reads as "insurance accepted" until an FAQ several scrolls down says it is not, and the pricing page's own footnotes make reimbursement conditional.

People fill silence with worst cases. My own review research (piece 01) suggests cost surprise, not cost size, drives the stall. The test: a plain cost module on the primary path, "programs from $1,290, monthly plans, full refund if you are not a medical fit," against the current hide-and-seek. I would bet on candor.

3. The pricing page justifies the price with endorsements labeled "Paid Partner."

Directly under the cost comparison sits a large grid of client videos, and most cards carry a Paid Partner tag that no tooltip or link explains. This is the one page where a visitor is explicitly deciding whether the price is worth it, and the social proof offered for that decision is compensated content presented without context.

This one I would not A/B test; I would just fix it. Explain the tag where it appears, and let the organic stories carry this page. There are 334 testimonials on the reviews page; the pricing page deserves the unpaid ones.

4. The proof stack travels everywhere; its fine print stays home.

The same figures, 89 percent improvement, 95 percent no side effects, 4.7 across 850,000 sessions, appear on essentially every page. They trace to one peer-reviewed but open-label, company-run study, and that limitation is disclosed exactly once, on the results page, never where the stats are reused. Similarly, the fact a cautious reader most needs, that ketamine is Schedule III and prescribed off-label for these conditions, sits roughly twenty sections deep in a state page, past every CTA. The skeptics in the public reviews say the same thing in their own words: they want the unvarnished version before the impressive one.

The company's first stated value is intellectual honesty. The test: carry a one-line caveat with the stats and surface an "honest facts" box higher on state pages, measured against skeptic-segment conversion. I think the honest version wins on revenue, not just principle, and Sam's thread in Tend is built on that bet.

5. Smaller frictions, quick wins.

The homepage's "How our program works" link goes to a differentiation page that never walks through the steps; the actual walkthrough lives at /program, one label swap away. The "27% stronger than talk therapy" chart offers no visible methodology, which reads as a claim to exactly the visitor it is meant to persuade. And FAQ content is split across three homes with duplicated answers, so the nervous visitor's questions have no single address.