Findings · Fishers AI workshop · October 22 to November 3, 2026
For two weeks we rebuilt one presentation and its offer, and tested almost every change in simulated rooms of local business owners before keeping it. Here's what actually changed how many of them bought, what didn't, and what the tests taught us about testing.
The presentation. A free evening AI workshop for local business owners in Fishers, Indiana. Jimmy presents. It teaches through eight beliefs and ends in a three-tier offer. Buying happens at the back of the room that night.
The tests. Rooms of simulated owners (roofers, salons, pizza shops, dentists, accountants, food trucks, a few skeptics, a few burned by agencies), each with a business, a temperament and a money situation. Rooms of 24, 50, 72, 100 and 120. Each owner read the whole night, bought or didn't, and said why. Along the way: price tests, skeptic studies, burned-owner studies and cold readers grading the offer.
The catch. The rooms buy more readily than real ones, so read the gap between versions, never the level. And the same version, read twice, swings by about 10%: 2 to 5 buyers out of 50, or $10,000 out of $90,000. These are predictions. The live room on November 9 is the judge.
Change in buyers per 100 owners, each change against the version right before it. Different-sized rooms scaled to 100 so they compare.
The door price: $997 / $2,997 / $4,997, from $1,997 / $3,997 / $7,497
Oct 22 · room of 100 · 24 buyers became 43
"How This Works" for owners burned before
Oct 25 · room of 100 · 44 became 53
His words, tested as the whole night, over the polished rewrite
Oct 25 · room of 100 · 32 became 41
The whole bill, and the people doing the work named
Oct 25 · room of 100 · 46 became 54
Screenshots instead of late live demos
Nov 3 · 50 owners, two reads · 14.5 became 17.5
Jim's $673,000 moved to a P.S., values brought into range
Oct 27 · three rooms of 24 · 38 became 42
The marketer's-eye section, two studies first
Nov 3 · 50 owners, two reads · 17.5 became 19.5
"About a day," not "five minutes"
Nov 3 · 50 owners, two reads · 19.5 became 20.5
The money dashboard added to Good
Oct 27 · 120 owners · 65 became 66
A full sweep of every broken hand-off
Nov 3 · 50 owners, two reads · no change
"Not a bread business" slide of real clients
Oct 27 · five rooms, 120 owners · tied 65 to 65
Five "for you" slides inside Jim's numbers
Nov 3 · 50 owners, two reads · 19.5 became 17.5
Every section rebuilt as a lesson
Nov 2 · 50 owners, two reads · 17 became 14.5
Moved buying upMoved buying downInside the noise: the same version read twice swings about 5 per 100
Not shown: four changes that made the cheapest tier feel complete. They barely changed how many bought; they changed what people bought, moving Better buyers down to Good. And "see it in your business first," which added buyers but cost the higher tiers.
One room of 100. Same offer, three price sets.
| Good / Better / Best | Bought | Builds wanted | Sales |
|---|---|---|---|
| $997 / $2,997 / $4,997 | 43 | 24 | $112,871 |
| $1,997 / $2,997 / $4,997 | 32 | 20 | $103,904 |
| $1,997 / $3,997 / $7,497 (the deck then) | 24 | 12 | $92,928 |
The Good price decides how many come through the door ("two grand is my rent"). The build prices decide how many builds you'd owe against a cap of 11. "I have to ask my wife" topped the reasons for waiting at every price. Changed: prices set at $997, $2,997 and $4,997.
Two offers. A cleaner rewrite (B), and the October 9 offer in his own words. Read alone by three cold readers, B won every grade, and every reader bought one tier higher. He called it unfair: B carried every structural fix and his version didn't. So we built A2: his words on B's skeleton, with the same fixes. Then we tested the whole night, presentation into offer, in a room of 100.
| Bought | Good / Better / Best | Sales | |
|---|---|---|---|
| B, the rewrite | 32 | 16 / 9 / 7 | $77,904 |
| A2, his words | 41 | 20 / 16 / 5 | $92,877 |
Nine owners went from nothing to the $997 meeting on one line of his: "Our consultations are $1,997 each." The rewrite's "$20,000 meeting" read as hype. Changed: A2 became the base, B's best concrete bullets went into it, and two rules came out of it: judge the presentation and the offer as one night, and only compare versions that carry the same fixes.
We pulled the 28 hardest owners (10 skeptics, 18 burned by an agency) and asked what stopped them. The walls: "Show me it gets built before you show me a price" (12). "One client isn't proof" (8). "Another monthly bill" (7). And one the presentation itself created: Jimmy's story said he once "promised more websites than he could build," and then the offer sold websites.
| Room of 100 | Bought | Sales |
|---|---|---|
| A2 | 44 | $81,868 |
| A5: adds "How This Works" (everything we agree on gets built, in a signed agreement, yours if you leave) | 53 | $116,841 |
| A5, run again as the new base | 46 | $107,862 |
| A6: adds the whole bill ("The price is the price. Nothing gets added without your yes."), the four people doing the work named, a track record, and the "promised more websites" line cut from Jimmy's story | 54 | $123,838 |
"How This Works" moved 13 owners and nobody away. Among the burned owners, buyers went from 8 to 12. One owner dropped from Best to Better because a promise became "could": "Could means nobody's sent it yet." One caution: the A6 test was run by a simulator that had already seen the burned-owner study, so part of that gain is confirmation, not a fresh read.
Two tests that day were dropped. A "see it in your business first" picture added Good buyers (28 to 36) but cost Better ones, and he cut it because it "made Better buyers cautious" and pulled in weaker customers. A6.3, which made the meeting sound more finished, sold the same as A6 (57 against 56) and moved two Better buyers down to Good. He kept A6: "we're really splitting hairs."
A new room of 24 read the old night and the newly locked one side by side. The new one sold 13 against 10. The biggest loss in both: values that read as made up. Jim's bread company, losing $4.58 a box, was shown making $673,000 a year. That was the main target.
So Belief 5 was rebuilt from the bottom: the math per box, $140 a box as a marketing budget against the other guy's $7, then "say it's only 30%, still six times the other guy." The $673,000 dropped to a P.S. The offer's values came into range at the same time ($35,640 total, the story at $1,997 against a video company's $5,000).
| Owners doubting each number | Before | After |
|---|---|---|
| Jim's $673,000 | 17 | 5 |
| The website's value | 8 | 3 |
| "Priceless" | 7 | 0 |
| The meeting at $10,000+ | 3 | 11 |
| $140 a box | 1 | 13 |
| Found money at $9,997 | 0 | 11 |
The doubt moved, and buyers still went from 38 to 42 across three rooms of 24. The new doubts were about numbers people grumbled at and bought past anyway. Changed: nothing about those numbers, and a rule: read a simulation by what owners did, not what they felt. Fix only what the no-buyers stopped at.
The same morning, two more tests went nowhere. A slide for "not my kind of business" owners ("Not A Bread Business? Neither Were These," eight real clients) moved none of the four owners it was built for: three had already checked out during Jim's bread numbers, before the slide arrived. And repeating "this offer lives here, in this room" cost two owners who won't spend without their spouse. Changed: that line came out of the written offer, to be said live. Across five rooms (120 owners) the two finalist nights tied: 65 buyers each.
Adding the money dashboard to the cheapest tier also tested even (66 against 65). It gave Good buyers another reason, not new buyers, and Good can't actually deliver it. Not adopted.
A new room of 50, each version read twice.
| Version | What changed | Buyers | Sales |
|---|---|---|---|
| 15 | The starting point for this room. | 17 | $43,949 |
| 16 | Every section rebuilt as a lesson (tell, show, try, hand off). Offer reshaped into three steps. | 14.5 | $39,457 |
| 16.1 | Belief 6: one live moment (a call to the pizza shop's AI), screenshots instead of live builds. | 17.5 | $42,448 |
| 16.2 | Belief 7: "through the eyes of a marketer." Two outside studies first, then what we did to one handyman's estimate. | 19.5 | $48,442 |
| 16.3 | "About a day," not "five minutes." Belief 4 rebuilt around what your customers love. | 20.5 | $48,439 |
| + "for you" | Five slides translating Jim's money ideas to every business, plus a full flow sweep. | 17.5 | $43,448 |
| Final | The sweep kept. The five "for you" slides out. | 19.5 | $47,442 |
The marketer's-eye section was the clearest win here: 14 of 100 readers named it as what moved them, and 12 of those bought. The "for you" slides were the clearest loss: owners who named Belief 5 as what held them back went from 6 to 14, all with the same complaint, that Jim's numbers are estimates. More minutes on projections meant more minutes to doubt them. The flow sweep cut the number of owners who got lost (32 to 27) and didn't move sales at all.
The two biggest gains of the whole build came from the same place: the owners who'd paid for promises before. "How This Works" (+9 buyers, about +$35,000 per 100), then the whole bill and the named people (+8 buyers, about +$16,000). Neither one added a deliverable. They took the fear out of saying yes.
Before the stack, show how it works, what it costs in total, and who does it. Transparency sells to the people who've been burned, and it costs nothing with everyone else.
Dropping the entry tier from $1,997 to $997 and the top from $7,497 to $4,997 nearly doubled buyers (24 to 43 per 100). The top tiers set how much building you sign up for. The bottom tier sets how many people start at all.
Price the first step for the owner whose rent it would be. Make money on the climb, not the door.
Read alone, the polished rewrite beat his words with every reader. Run inside the whole night, his words sold 41 against 32. A grader judges an offer's logic. A buyer reacts to what the night has already made them believe, and to lines like "our consultations are $1,997 each."
Never pick an offer by reading it alone. Run it at the end of the presentation it lives in.
The $673,000 went from the most doubted number in the room to a P.S. most owners bought past. The per-box math they could check in their heads did the selling.
Lead with the number they can check. Let the big one ride underneath.
Two outside studies first (one option, 10 of 100 buy; a choice, 66 of 100; a third option, 43% more spent), then "So now you can see what I saw that Jamon didn't." Buyers rated "I need someone with a marketer's eyes, not just the tool" 4.3 out of 5. Non-buyers, 2.9.
The sale lives in the gap between the tool and the eye. Borrow outside proof that the eye is real, then show yours at work on a real client.
A meeting stripped of everything it previewed. A "see it in your business first" picture. A meeting described as "ready," with the new site "already picked." The dashboard put into Good. Each time, Better buyers stepped down or Good buyers just got another reason. "My new site's already picked and ready." "I'll build it myself."
Good should be valuable and clearly the start. When it sounds like the finish, it eats the tier above it.
"This offer lives here, in this room," said twice, stopped two owners who never decide without their partner: "I didn't love being told it disappears tonight." A free 24-hour hold would have won them back on paper (about 20 buyers instead of 16), but only if someone calls every hold the next morning. He kept "tonight," said once, live, and no hold: a pile of holds in the real world becomes phone chasing.
Say the deadline once and mean it. Repeating it doesn't add buyers. It only costs the careful ones.
Twice we tried to bring back owners who felt "that's bread, not my kind of business." A slide of eight real client businesses afterward moved none of them; they'd checked out during Jim's numbers. Five "for you" translation slides inside Jim's numbers cost 3 buyers. The one thing that worked was a single line on the total: "For you, that's every order. Every job."
A fix placed after the moment of doubt never reaches the person. One line in the right spot beats five slides.
Rebuilding every section as a clean lesson read easier and sold the same or less. A full sweep of every broken hand-off cut the number of lost owners and didn't move sales. Telling the truth ("about a day" instead of "five minutes") cost nothing either.
Fix clarity and honesty because a confused or misled room can't buy. Don't expect them to sell for you.
Next: the live night, November 9. What to watch: how many stop at Jim's numbers, whether "How This Works" and the marketer's-eye section move a real room the way they moved the simulated ones, and how many say "I need to talk to my spouse."