Founders & Small BusinessFree
Service Offer and Packaging Design
Share your client list and what delivery costs you, and get service packages priced where your costs jump, tested against the clients you already have.
Marlowe Systems, a fictional IT support firm, has 34 managed IT clients and 34 different prices, because every one was negotiated. Seven of those clients buy materially the same thing: a server on site, business hours only, somewhere between 28 and 46 endpoints. They pay between $1,180 and $1,480 a month for it, a spread of 25.4%. River finds the points where the cost of delivering actually steps, puts the package boundaries there, and then runs the whole historic book back through the tiers it proposed.
Every guide to this arrives at three tiers priced off each other, the middle at one and a half times the entry and the top at two and a half. Applied to Marlowe's own cost structure, that puts the middle tier at $937.50 against $1,019 of delivery cost at the top of its own range. It clears a healthy margin on a 30-endpoint client and loses money on a 60-endpoint one, and nothing about the tier tells you which you just signed.
Written for the owner whose every quote turns into a negotiation, and who suspects the largest clients are the worst ones. It needs a real cost per job underneath it, which is what the profitability by job or service works out. What belongs in each package comes partly from the discovery interview synthesis and partly from the competitor and market scan. Moving the clients you already have onto a new structure is a price increase communication job. Which of them is quietly leaving anyway is the customer feedback and churn review.
The boundary is a cost fact, not a price ratio
The middle-tier advice is not wrong about buyers. The work behind it found that an option becomes more attractive simply for being the intermediate one in the set, and less attractive for sitting at either end. That tells you which tier gets picked. It says nothing about whether that tier's price covers the cost of delivering it, and the ratios people bolt onto the idea are pure convention. A boundary drawn at 1.5x is a guess about a buyer. A boundary drawn at a cost step is a fact about you.
Managerial accounting already names the thing. A step cost holds flat across a range of activity and then jumps to a new level, and the span it holds flat across is the relevant range. Marlowe has three such jumps and one smooth cost. A server on site adds $310 a month. The on-call rota adds $640, because it needs two engineers on standby whether anybody calls or not. The four-hour on-site guarantee adds $890 for a held slot and a stocked vehicle. Endpoints cost $9.40 each, on top of $145 per client.
Three switches allow eight combinations, and six of them appear in Marlowe's book. Packages drawn on those switches price at $625, $1,595 and $4,250, holding 39.2%, 36.1% and 26.8% at their own ceilings. Tested against the 34 current clients, 28 fit and 6 do not. That is 17.6% of clients but $16,580 of $63,880 a month, so 26.0% of revenue, because the clients that do not fit are the big ones. Three of them share one combination, and a fourth package at $2,450 takes all three.
How it works
Send the book
Share your client or job list, what each one pays, and what makes clients bigger or smaller.
Name the switches
Name what changes how you deliver: a server on site, after-hours cover, a response guarantee.
Give the cost side
Say what an hour of delivery costs, and which standing costs exist because some clients asked.
Review the packages
Get the packages, the margin at both ends of each, and the clients that none of them fit.
What you get
- Delivery cost split into what scales per unit and what jumps at a threshold
- Boundaries placed at those thresholds, so no tier straddles a jump in your cost
- Margin computed at both ends of every tier, not at one average client
- Every historic client run back through the proposed tiers, one row each
- The share of clients and of revenue that no package fits, reported separately
- The price spread you currently charge for service that is materially identical
Common questions
How many packages should I have?
Not a number you pick. Count the distinct combinations of cost switches in your own book. Marlowe has three switches, so eight combinations are possible and six actually appear. Three packages leave 26.0% of revenue with no tier that fits, and a fourth drops that to 14.8%. Stop when another package no longer fits many more clients.
What does it mean when a client does not fit?
It means no proposed package fits that client cleanly: either they need something every package leaves out, or the only package that covers them makes them pay for a service they never asked for. River reports this share twice, by client count and by revenue, because the two usually differ. At Marlowe it was 17.6% of clients but 26.0% of revenue.
My costs scale smoothly. Nothing steps.
Then look for what you hold in reserve rather than what you consume. An on-call rota, a held appointment slot, a stocked vehicle, a licence sold in blocks of ten, a second person needed for lifting. All of those cost the same whether one client uses them or five. If genuinely nothing steps, price per unit and skip the tiers.
Should I build a package for clients that don't fit?
Only where several fail to fit in the same way. Three of the six Marlowe clients that do not fit need a server and after-hours cover but not the on-site guarantee, so a fourth package at $2,450 fits all three. The other three do not fit for unrelated reasons, including one at 156 endpoints. Those stay bespoke, and naming them as bespoke is itself a result.
Do I have to move existing clients onto the packages?
No, and the backtest is partly there to tell you what it would cost. Marlowe's 34 clients pay $63,880 a month, and under four packages the same book bills $67,720, which is 6.0% more or $46,080 a year. Whether that survives contact with the renewal is a price increase communication question.
What if a tier loses money at its own ceiling?
Then the ceiling is in the wrong place, or the price is. Margin gets computed at both ends of each tier precisely so you can see it. Marlowe's entry package holds 75.3% at one endpoint and 39.2% at 25. A tier whose top end goes negative is a discount you have agreed to give the largest buyer in that tier.
Sources
Built by
Chandler Supple
Co-founder & CEO, River
Chandler has spent the last six years working in AI, building tools for millions of users and Fortune 100 companies. In 2025, he co-founded River to build a new kind of operating system for AI agents.
Service Offer and Packaging Design
Describe what you need, add any files you have, and your workspace opens with the work already underway.