All-You-Can-Eat vs. Tiered Managed IT Plans: Which Fits Your Book
Published by
Throne of Profit EditorialReviewed by
William Hassell
Founder & Chief Editor, Throne of Profit
Most MSPs sell some version of "all-you-can-eat": a flat per-seat fee, unlimited support, everything covered. It's easy to explain and easy to buy. The trouble shows up on the profit report, not the pitch. A handful of high-touch clients — the ones who open ten tickets a week, run brittle line-of-business apps, never let you standardize anything — quietly consume the hours you priced for everyone. When every client pays a flat rate but some clients cost three times as much to serve, your light-touch accounts are silently subsidizing your heavy ones — and you can't see it in the top-line revenue.
The fix isn't to punish the heavy clients or cheapen the light ones. It's to design plans so each tier carries its own cost to serve, and to gate the features that drive that cost behind the tier that pays for them. Done well, a tiered structure lets a demanding client pay for the demand instead of the account next door doing it for them.
FLAT PLAN (hidden cross-subsidy)
light client ▇▇ pays same → MARGIN +++
avg client ▇▇▇▇▇ pays same → MARGIN +
heavy client ▇▇▇▇▇▇▇▇▇▇▇▇ pays same → MARGIN ─── (eats the others)
TIERED PLAN (cost follows the client)
heavy client → higher tier / gated features → pays for the loadOwner symptoms
Revenue looks healthy, but a few named accounts feel like they eat all your labor.
Your best margins come from clients who barely call — and you're afraid to lose them.
Every plan is "unlimited," so there's no lever to pull when a client's usage balloons.
Why this happens
Flat, all-you-can-eat pricing averages every client's cost of service into one number. That works only while clients cluster around the average. The moment your book spreads out — a few very light accounts, a few very heavy ones — the average hides a cross-subsidy. The heavy clients don't pay more because nothing in the plan is tied to how much they consume: no cap on hours, no gate on the services that generate the load, no tier that reflects the difference between a standardized 20-seat office and a chaotic 20-seat shop running legacy software. Usage varies wildly; price doesn't move at all. The gap between the two is margin leaking out of your light accounts.
Common mistakes
Pricing on seat count alone, when two clients with the same headcount can differ threefold in support hours.
Making everything unlimited, which removes every lever to match price to load.
Gating nothing by tier — after-hours support, project work, and non-standard apps all included at the base rate.
Never measuring cost to serve per client, so the subsidy stays invisible until a renewal feels wrong.
Grandfathering heavy clients forever on the plan they signed three years and two acquisitions ago.
Business consequences
An MSP that can't see cost to serve keeps signing more of whatever it's underpricing, because those clients are easy to close and feel like growth. Meanwhile the profitable light-touch accounts carry the load, and if one of them leaves, the margin math gets worse fast. The owner who tiers deliberately gets the opposite: each plan roughly covers its own cost, heavy clients self-select into (and pay for) the tier that fits them, and a lost light account doesn't take the whole book's profit with it. The difference between the two isn't revenue — it's whether the revenue keeps anything after labor.
How experienced operators think about it
They stop thinking in seats and start thinking in cost to serve. The mental model is simple: every client sits somewhere on a load curve, and price should rise with load — not smoothly for every ticket, but in steps that map to real drivers. What actually drives load in an MSP is usually a short list: standardization (or the refusal of it), after-hours and response-time expectations, non-standard or legacy applications, and project versus break-fix work. Experienced operators design tiers around those drivers and put the expensive ones behind gates, so a client who wants unlimited after-hours or refuses to standardize is choosing a higher tier, not getting it for free at the base rate. The plan does the sorting.
Practical actions
Measure cost to serve per client. Pull ticket volume and hours by account for a few months. Rank the list. The heavy tail is where your margin goes.
Name your real load drivers. Standardization, after-hours coverage, response-time targets, non-standard apps, project work. Price follows these, not seat count alone.
Design three tiers, not one. A standardized base plan, a middle plan, and a high-touch plan — each with a cost to serve you've actually estimated.
Gate the expensive features. Put after-hours support, guaranteed response times, and non-standard app support in the tiers that pay for them, not the base.
Tie the base tier to standardization. Make the cheapest plan require the practices that keep it cheap; a client who won't standardize belongs a tier up.
Re-tier at renewal. Review each account against its actual load once a year and move the ones that have drifted onto the wrong plan.
Questions every owner should ask
Do I actually know which accounts cost the most to serve, or am I guessing from memory?
If my three lightest-touch clients left tomorrow, would the rest of the book still be profitable?
Which of my "included" features are really being paid for by the clients who don't use them?
Frequently asked questions
Won't tiered plans annoy clients who are used to "unlimited"?
Most clients don't object to tiers; they object to feeling nickel-and-dimed mid-relationship. The way to avoid that is to sell tiers as clear packages up front, with the expensive drivers — after-hours, guaranteed response times, non-standard apps — visible and named, so the client chooses the level of service they want. Light-touch clients often prefer a leaner, cheaper plan once they see they're not paying for coverage they never use. The friction comes from re-pricing a heavy client who was underpriced for years, which is why it's better to tier from the start and re-tier at renewal, not mid-term.
How many tiers should an MSP offer?
Usually three is enough — more than that and clients can't tell them apart, fewer and you're back to averaging. The point isn't the number of tiers; it's that the tiers map to your real load drivers and that the features driving cost are gated to the plans that cover them. Start with a standardized base, a middle plan, and a high-touch plan, and only add complexity if a genuine segment of your book doesn't fit those three.
Related articles
Running a Profitable Managed IT Services Business — the pillar.
Onboarding a New Managed IT Client Without Losing the First 90 Days — set the standardization that keeps a tier cheap.
Building a Ticket Triage Workflow That Stops the Chaos — where you first see load by account.
Why Jobs Take Longer Than You Quoted — the general cost-to-serve problem.
Where Time Leaks on a Typical Job — finding the hidden hours.
Every business has more decisions than time
Whether you need help solving one problem, evaluating a major opportunity, or making a company-changing decision, Throne of Profit gives you consulting capacity on demand.
Purchase only the consulting capacity you need and use it across Weekly Focus, Strategic Focus, Financial Focus, and ThinkTank engagements.