Pricing, offer, guarantee, downsell, deliverables and the roadmap artifact. Each decision traces to a retrieved Hormozi passage, or is flagged as inference where the corpus is silent. Section 07 is the honest accounting of which is which.
What it has to respect: the current $4,000 a month for 3 pages, a willingness to go to 4 pages or down to 2 with bigger variation between them to make four tests, a client spend range of roughly $30k to $3M a month, the free landing page design as the existing front end, and the three rung definition of working (Signal, Winner, Scaled).
Clients turn $4,000 into $1,333 a page because pages are the only thing they have been given to count. Hormozi's instruction is to fix the frame before touching the number.
He makes the point hardest with an IT provider whose entire value was that nothing broke. The correction: you are not selling that nothing bad happened, you are selling that the plane lands on time, and then you guarantee around the bad thing responsibly. Decouple the elements of value first.
He is equally blunt that averaging what competitors charge, or undercutting it slightly, is how you guarantee you stay a commodity, and that cost-plus pricing fails because it is about you rather than what the buyer will pay.
LandersRX does not sell pages, and does not sell iteration. It sells at-bats against a named control, each with a read at the end. The page is the instrument. The read is the product.
That is the only framing under which iterating a winner costs the same as building a new page, because both are one at-bat.
A test counts when all three are true. It runs at a separate URL against a named control, being the incumbent page or the PDP where there is no incumbent. It changes at least one load-bearing variable. It receives enough spend to produce a read inside its window.
Colour, section order, image swaps, copy polish and button text are not load-bearing. Those are revisions inside a test, not tests.
It is not a preference, and the account decides. A test needs a minimum daily spend pointed at it to resolve inside a two week window. Call that floor S dollars per day per concurrent test, placeholder $250, needing calibration against real reads.
concurrent tests = floor( daily budget pointed at landers / S ) tests per month = concurrent tests x (30 / window length in days)
At the floor of the client range, with maybe a third of budget pointed at landers, that is $300 to $550 a day, so one to two concurrent tests. Four distinct pages in one window would each be starved and none would read. The instinct to go down rather than up is correct at the floor, and it now has a reason.
This also matches existing company doctrine: under roughly 1,500 views per window, bundle the changes into one big-swing challenger rather than splitting single variables. At low spend, two pages with obvious variation between them is not a concession, it is the only structure that can produce a read.
At $150k a month and up the same formula allows four or more concurrent, so higher bands get width instead of sequence. The deliverable shape changes with the account. The unit does not.
Asked almost exactly this question by a data firm whose clients got more complex as they got bigger, Hormozi's answer was tiers on the client's own value metric, then a waterfall of declining percentage bands, not retroactive, so the blended rate lands well under the top band while absolute dollars rise. He said to keep the existing fixed fee as the bottom and put the bands above it.
On picking the metric he is specific. It has to be the thing that, when it rises, proxies how much value the client is getting and by extension how much they can pay. For ecommerce he names revenue directly.
For LandersRX the right metric is monthly paid ad spend, not revenue. Spend is what a landing page multiplies, it is already the number the buyer watches, it is verifiable in the account being read anyway, and it is the same number that sets how many tests the account can feed. One metric drives both price and delivery.
| Monthly ad spend | Rate on that band | Tests per month |
|---|---|---|
| First $50,000 | $4,000 flat | 4 |
| $50,000 to $150,000 | 4.00% | 6 |
| $150,000 to $400,000 | 2.50% | 10 |
| $400,000 to $1,000,000 | 1.25% | 14 |
| Above $1,000,000 | 0.60% | 20 |
Bands are marginal and never retroactive · the base band is 8.0% effective
| Client spend | Monthly fee | Blended rate | Tests | Cost per test |
|---|---|---|---|---|
| $40,000 | $4,000 | 10.0% | 4 | $1,000 |
| $100,000 | $6,000 | 6.0% | 6 | $1,000 |
| $150,000 | $8,000 | 5.3% | 6 | $1,333 |
| $400,000 | $14,250 | 3.6% | 10 | $1,425 |
| $1,000,000 | $21,750 | 2.2% | 14 | $1,554 |
| $3,000,000 | $33,750 | 1.1% | 20 | $1,688 |
Two things in that table are deliberate.
The $4,000 base is the most expensive band in the whole schedule, at 8% effective, and every band above it is cheaper. Hormozi's note on why a waterfall sells well is that prospects anchor on the first band and assume it is the rate, so a high small band acts as a reverse anchor and everything after reads as a shoe-in. The current price point is not a floor to apologise for. It is the anchor the rest of the schedule discounts from.
Tests grow slower than fee, so cost per test rises from $1,000 to $1,688 across the range. That is where margin comes from, and it is honest rather than arbitrary. The second test inside an account is cheaper to run than the first, because the account read, the angle map, the tracking and the offer box architecture already exist. The client is paying for compounding insight, not for labour hours.
Yes, and the schedule says so out loud. Hormozi's answer on the rich buyer and the poor buyer receiving an identical deliverable is that price is set by willingness to pay, not by the deliverable, and that the same service genuinely is not worth the same to both.
Same page. The waterfall is the mechanism that lets price follow value without manufacturing extra deliverables to justify it. He also argues for being the premium price leader in any service business involving humans, on the grounds that the same money from 10 clients beats the same money from 100, and the 10 client version is the one you would rather scale. That is an argument for raising the floor over time, not for adding a cheaper tier beneath $4,000.
Hormozi is against linear pricing wherever it can be avoided, and specifically for onboarding fees and upfront charges, because signup is the moment of maximum deprivation and maximum willingness to pay, and because service work is front-loaded anyway. His gym example: stop selling a trial at the moment the buyer is most motivated, charge for the first transaction and go cheaper after. He adds that the reason people buy is not the reason they stay, so the two should be priced separately.
They buy to get a winner. That is Phase 1, maximum pain, maximum willingness to pay, and the work is genuinely front-loaded.
They stay to keep winners winning and to widen. That is Phase 2 and 3, lower pain, ongoing, and correctly priced on spend.
So a one-time build-out fee of $2,500 on top of month one, covering the account read, angle map, offer box architecture, tracking setup and the first four builds. Waived on a three month commitment.
Two passages, same conclusion. Do not present a price and then change it, and replace discount requests with speed or with risk removal. On retention structure: use the onboarding fee as the anchor, waive it for a longer commitment, and have an early exit pay back what month to month would have cost.
Applied: the $2,500 build-out fee is the only negotiable item, and it trades for commitment length, never for cash. A client who leaves before month three pays the waived $2,500 on exit. Nobody ever gets $4,000 reduced to $3,200.
Hormozi's formula is "if you don't X in Y time, we will Z", and the part most people get wrong is the remedy. People do not want their money back, they want the roof. The guarantee exists to pay down the risk of not getting the thing, so the remedy should be getting the thing.
In the same conversation he proposes a profit guarantee rather than a money back guarantee, on time and on budget or the provider hands over its 20%, because it puts visible skin in the game while leaving the provider solvent. On uncontrollable outcomes he is clear that you can still guarantee, you just scope and condition it the way insurance does, and you size it as a controlled burn you have done the math on.
If no page in the first round reaches Signal within the round window, LandersRX keeps testing at no additional fee until one does, up to four further tests.
X is Signal, the first rung, which is the only rung reachable inside one round and the only one partly inside LandersRX's control. Winner and Scaled depend on the client's budget decisions and belong to the downsell and the roadmap, not the guarantee.
This is the 3 of 5 and 1 of 5 problem, and it needs a pre-written answer so it is never negotiated live.
The guarantee is satisfied when any one test in the round reaches Signal.
One of five satisfies it. Three of five satisfies it. Zero triggers the remedy. Said up front this reads as confidence rather than as a loophole, and it removes the only question that currently gets asked after delivery. What it deliberately does not do is promise a hit rate. A round producing one Signal out of five is a normal, healthy round, and the pricing assumes exactly that.
Separate from the guarantee, and fully inside LandersRX's control. Advising an SEO agency with the same invisible-outcome problem, Hormozi moved the promise onto a countable thing the agency could always deliver, at least seven revenue opportunities found, and told him to drop the word audit because it sounds boring.
Applied: every round ends with a written read naming at least seven page-level opportunities in the account, each with the behavioural evidence behind it and the page it implies. Not called an audit. This is the floor of delivered value that exists even in a round where nothing reaches Signal, and it is what makes the guarantee safe to offer, because free continuation is never the only thing standing between LandersRX and an angry client.
Hormozi on upsell timing: selling at the end does not work, people do not sign up again when they are done. Sell at the halfway point, and credit what they already paid toward the next, longer commitment, as an either-or. And on why iterating a winner is the right use of the slot rather than a waste of it: once you have a control that wins, more of what works is the highest risk-adjusted move available.
Your first $4,000 is credited. Either another $4,000 locks the next two months at this rate with iteration included, or another $8,000 locks the next five months and moves the account to six tests.
The objection disappears because the sentence is never "pay a page slot to iterate". It is "Signal is not revenue, Winner is, and iteration is the only thing that converts one into the other". They are buying the rung, not the asset.
The open question was what performance means. The three rung ladder answers it, and the two rungs do different jobs so they cannot collide. Hormozi's continuity logic: the front end liquidates acquisition cost, then you down-sell the upsell as more for less to keep them. His pricing survey is the reason to expect this to convert at all, since the identical service framed as pay-per-show drew roughly four times the willingness to pay of a flat fee, and moving a retainer toward pay for performance is itself an implied guarantee.
Trigger. A completed round, zero pages at Signal, client signalling exit, and the six conditions met on their side.
Offer. The next round carries no monthly fee. LandersRX keeps testing. When a page reaches Winner, the client pays a one-time success fee of 1.5x their monthly rate. Either side can end it with fourteen days notice.
Cap. One downsell round, extendable once at LandersRX's discretion. After that it is a fit problem and the right move is to end the relationship.
Signal is the guarantee rung and has already been spent by the time the downsell is on the table, so the downsell pays on Winner. That is a bar the client will accept as real, and it keeps the two instruments from overlapping. The risk is a controlled burn with arithmetic behind it: if a saved client continues six months at $4,000, one free round pays for itself at better than a one in six save rate.
Hormozi's position on churn is that it is a value-match and qualification problem rather than a price problem, that LTV is king, and that volatile small accounts churn regardless, so messaging should name the ICP and repel everyone else. He also notes that marketing vendors specifically get churned left, right and centre, so the job is finding what clients do not churn out of.
Qualification gate, applied before a free design is built: at least $30k a month in sustained paid spend rather than a launch spike, enough budget to point S dollars a day at two concurrent tests, headroom to scale meaning the account is spend-constrained rather than demand-constrained, and a decision maker who can approve a page in three days.
A prospect who reduces $4,000 to a per-page number after the framing in section 01 has been delivered is failing qualification, not negotiating. That is avatar signal, and the answer is to disqualify rather than to discount.
Hormozi's decoupling instruction is the reason to list these separately rather than bundling them into "3 pages". Each line is a distinct element of value.
| # | Every round, at every band |
|---|---|
| 01 | Ad account read, with the wasted angles named |
| 02 | Angle and persona map for the account |
| 03 | Offer box architecture: what the buy box promises, how price is framed |
| 04 | N page builds at separate URLs, each against a named control |
| 05 | Ad split instructions written for the client's media buyer |
| 06 | Tracking and read setup, verified pre-launch |
| 07 | Midpoint read with the rung called out loud |
| 08 | End of round written read, at least seven named opportunities with evidence |
| 09 | Ranked iteration queue for the next round |
| 10 | Roadmap update, stage marked per pillar |
Social proof blocks, timeline and transformation, FAQ built off real objection data, shoppable destination, FOMO mechanics, image to GIF to video progression, expanding the reason count.
New products, pillars, angles and personas, plus the harder page types (expanded PDP, quiz, advertorial) once listicle and sales page are cracked.
An account does not have one stage. Each pillar, meaning a product, angle or persona, carries its own stage. So a brand sits at stage 3 on its hero product and stage 0 on a new persona at the same time, and the roadmap shows both without contradiction.
Phase 2 and Phase 3 work then run concurrently by construction rather than by exception, which is the thing the current phase model could not express.
| Stage | Name | Bar to graduate |
|---|---|---|
| 0 | Unproven | No page on this pillar has held spend |
| 1 | Signal | A page holds baseline efficiency at readable spend |
| 2 | Winner | A page holds baseline at real spend share for two weeks |
| 3 | Width | Winners live on two or more angles or products |
| 4 | Engine | Pages are a top three spend destination with a standing test cadence |
Highlighted rows are the three rungs · stages 0 and 3 are the transitions between them
The existing phases map on directly. Phase 1 is stage 0 to 2. Phase 2 is the depth work that gets 1 to 2 and holds 2. Phase 3 is 2 to 4.
The source roadmap's personalisation is fake. It is templated by quiz answer with nothing merged in, no name, no revenue, no headcount. Real account data in the scorecard is the whole advantage, and it is cheap to produce because the read already exists.
Binding constraints on the artifact, from existing company policy: third person by brand name throughout with no "you" or "your", no pricing or packaging or terms anywhere in it, every roadmap item must imply a page, and check whether a recorded walkthrough accompanied it before building.
Worth stating so none of the above reads as more sourced than it is.
It is MoreMozi podcast era material: consulting calls, hiring, scaling, pricing, guarantees, upsell timing, churn, all in real depth. The guarantee formula arrives secondhand, through the roofing call where he references page 125 rather than reads it. Treat the formula as reliably reported and the surrounding mechanics as directly sourced.
How many landing page tests a paid account can actually resolve at once is the central operational constraint in this business, and he never touches it. The per-test spend floor S, the concurrency formula, and the whole two-versus-four answer in section 02 are not his. They need calibration against real reads before anything is quoted to a client on them.
His guidance is that offer tiers should sit roughly 5x apart rather than 20% apart, and that each new tier should aim to double total revenue at 20% to 30% uptake. That assumes one audience choosing between tiers. LandersRX clients do not choose, they self-sort by ad spend, so a waterfall replaces the ladder. The 5x spread survives where it matters, between the 8% effective base rate and the 0.6% top band, which is a 13x spread in rate.
Across the pricing material he pushes almost every operator up rather than down, and his view is that the common failure is not going high enough at the top or low enough at the bottom. This schedule keeps the $4,000 anchor intact because it is live and working. A harder reading raises the base and lets the free design carry more of the conversion load.
| # | Parameter | How to set it |
|---|---|---|
| 01 | S, minimum daily spend per concurrent test. Placeholder $250. | Derive from actual reads: what daily spend has historically produced a callable result inside two weeks. |
| 02 | Whether the bands hold against delivery cost. | Cost per test rises $1,000 to $1,688. Check against real hours per test at the higher bands, where page types get harder. |
| 03 | Success fee multiple on the downsell. Placeholder 1.5x monthly. | Break-even is roughly a one in six save rate at six month continuation. |
| 04 | Build-out fee and waiver threshold. Placeholder $2,500 at three months. | Six months is the more aggressive version. |
| 05 | Where Signal's baseline efficiency bar sits, per account. | The guarantee is unbackable until this is written down per client before launch rather than argued after. |
Whether the free landing page design stays free once the waterfall is live, or becomes the qualification step in front of it. Hormozi is firmly pro free front end for proof, so the current answer is keep it free.