What Service Fees Should an IT or MSP Company Charge?

The outcome you should expect

The reason this question matters is not that add-on fees are exotic — it is that managed services is a business where the recurring contract alone almost never funds the whole delivery organization. The per-seat number a client agreed to eighteen months ago was priced against an environment that has since grown, sprouted a second office, added a compliance requirement, and picked up four SaaS tools nobody told you about. Meanwhile your dispatcher, your NOC coverage, your documentation effort, and your vCIO hours all got more expensive. A properly built fee schedule closes that gap without forcing an awkward renegotiation of the base rate every year.

Here is what the arithmetic looks like when you actually run it. The working formula is simple: monthly fee revenue equals the number of clients or tickets times the percentage that trigger each fee, times the fee amount. Take a provider with 40 client companies. Onboarding at $2,500 per new client, signing 3 new clients per month, is 3 × $2,500 = $7,500 per month. After-hours support billed at $175/hour and triggered across roughly 30 incident-hours monthly across the whole base is 30 × $175 = $5,250. A 10% handling fee on $60,000/month of pass-through hardware is another $6,000. Total: about $18,750 per month, or $225,000 annualized, on top of whatever the managed contracts already bring in.

The critical property of that revenue is its margin profile. These fees attach to work the delivery team is largely doing anyway — the onboarding engineer is on salary whether or not you charge for onboarding, the on-call tech is compensated whether or not the ticket bills at premium, the procurement coordinator places the order either way. The incremental cost of the fee dollar is therefore small, and contribution margin typically lands in the 85–95% range once you exclude the direct labor already carried in your fixed cost base. That is fundamentally different from adding another managed seat, where you take on real recurring delivery obligation for every dollar of recurring revenue.

What that margin buys you is capacity. The $18,750/month in the example above funds a dedicated dispatcher — the single highest-leverage hire most 40-client MSPs never make — plus a procurement coordinator, with room left over. Neither of those roles is billable in the traditional sense, which is exactly why they get deferred in a fee-less shop and why the service desk stays chaotic. The correct mental model is not "extra revenue." It is "the mechanism that pays for the non-billable spine of the business."

The second-order outcome is behavioral, and it is arguably worth more than the cash. When after-hours carries a real premium and out-of-scope work carries a real project fee, clients start making better decisions. The "can you just do this at 9pm" requests that were never actually urgent migrate back into business hours. The server migration a client wanted to trickle through the help desk gets scoped as a project with a statement of work, a timeline, and an engineer assigned — which is how it should have been handled anyway. Fees are a pricing signal, and pricing signals shape demand. An MSP with no fee schedule is running an all-you-can-eat buffet and then wondering why the kitchen is on fire.

The outcome you should *not* expect is a client revolt. This is the fear that keeps most owners from acting, and it is largely unfounded when fees are disclosed at contract time rather than sprung mid-relationship. The distinction that matters is between a *fee* and a *surcharge*. A fee is written into the master service agreement, explained during the sales process, and predictable. A surcharge is a line item that appears on an invoice without warning. Clients tolerate the first and fire you over the second.

What drives that outcome

Four levers determine whether a fee schedule produces the numbers above or quietly underperforms, and they are worth separating because MSPs usually get one or two right and leave the others broken.

Scope definition is the foundation. Every dollar of add-on revenue depends on a defensible line between what the recurring contract covers and what it does not. If your MSA says "unlimited support" with no qualifiers, you have no basis to bill a project fee, and your technicians will argue about it internally every week. Good scope language enumerates the covered environment (device counts, user counts, named applications, supported locations), the covered hours, and the explicitly excluded categories: hardware procurement, third-party vendor liaison beyond a stated threshold, major version migrations, office moves, new site standups, compliance audit support, and anything requiring a formal change window. Providers who write this well can bill confidently. Providers who wrote a two-page agreement in 2019 cannot, and no PSA configuration will fix that.

Trigger rate is the lever most owners never measure. The formula has three terms and the middle one — the percentage of clients or tickets that actually trip each fee — is the one nobody tracks. If your after-hours fee exists on paper but technicians routinely waive it because they feel awkward, your effective trigger rate is a fraction of what you modeled. Instrument this. Pull the count of tickets opened outside business hours, then the count actually billed at premium, and look at the ratio. A gap between them is not a pricing problem; it is an enforcement and culture problem, and it is fixed with a policy that removes the individual technician's discretion.

Automation determines whether the margin survives contact with delivery. The 85–95% figure assumes the fee attaches to work already being done efficiently. If every after-hours ticket requires a manager to manually adjust the rate on the time entry, you have introduced administrative cost that eats the premium and creates errors clients will dispute. This is where the PSA earns its keep: ConnectWise PSA, Datto Autotask PSA, and HaloPSA all support billing rules that apply a time-of-day or work-type multiplier automatically. NinjaOne and similar RMM platforms attack it from the other direction, reducing the technician-minutes per ticket through automated patching and remediation so the labor behind the fee shrinks.

Collection mechanics close the loop. A billed fee is not a collected fee. Onboarding deposits and project milestones are exactly the charges most likely to sit in accounts receivable while the client "reviews" them. Putting card or ACH on file through Stripe Billing or a comparable processor and charging automatically at defined milestones converts a collections conversation into a receipt. The processing cost — roughly 2.9% + $0.30 per card transaction, materially lower on ACH — is trivial against a 90% margin fee and against the cost of chasing it.

The loop at the bottom of that diagram is the part worth dwelling on. Fee revenue funding a dispatcher lowers the cost per ticket across the entire client base, which improves the margin on the recurring contracts too. That is why fee strategy is a RevOps problem and not a billing problem — the pricing decision, the delivery capacity decision, and the contract structure decision are the same decision viewed from three angles.

Benchmarks and realistic ranges

Channel benchmarks move, and any specific figure should be treated as a starting anchor you calibrate against your own market and client size, not a law. With that caveat, here are the ranges that hold up across the MSP channel.

Onboarding and setup: $1,500–$5,000+ per client, one-time. The spread tracks environment complexity almost linearly. A 15-seat single-site professional services firm with cloud-only infrastructure sits at the low end. A 120-seat manufacturer with on-prem servers, a legacy line-of-business application, three locations, and no existing documentation belongs well above the top of that band — some providers price larger onboardings as a scoped project rather than a flat fee for exactly this reason. A useful sanity check: estimate the engineering hours to inventory, document, deploy agents, configure monitoring and backup, and run a discovery workshop, then price at your standard project rate. If your flat fee is far below that number, you are subsidizing new clients with existing ones. Some providers waive or discount onboarding in exchange for a 36-month term rather than a 12 — that is a legitimate trade, but recognize it as a financing decision, not a giveaway.

After-hours and emergency support: 1.5–2× standard, commonly $150–$250/hour. Define "after-hours" precisely in the agreement: the covered window (say 8am–6pm local, Monday through Friday), the holidays excluded, and the billing increment. Fifteen-minute increments with a one-hour minimum for a callout is a common and defensible structure. The minimum matters more than the rate — a technician woken at 2am has lost the night regardless of whether the fix took eight minutes, and pricing that reality into the contract is honest rather than aggressive. Distinguish between *client-requested* after-hours work and *provider-initiated* remediation of an issue your monitoring caught; billing the client for the latter reads badly and is usually the wrong call.

Project and change fees: $125–$200/hour, or fixed-price by statement of work. Fixed price is generally better for both parties on well-understood work — migrations, refreshes, office moves — because it forces real scoping and gives the client a number they can budget. Hourly is more appropriate for genuinely uncertain work like troubleshooting an inherited mess. Whichever you choose, the trigger is the same: work outside the enumerated scope, initiated by a signed SOW or change order. The signature requirement is not bureaucracy; it is the artifact that prevents a billing dispute four months later.

Hardware and software procurement handling: 5–15% of cost. The band is wide because the work behind it varies enormously. Drop-shipping a monitor is not the same as sourcing, ordering, receiving, imaging, staging, and deploying forty laptops with asset tagging. Providers who charge a flat percentage across all procurement systematically overcharge on the easy items and undercharge on the hard ones. A tiered structure — a lower percentage on simple pass-through, a higher one or a per-unit staging fee on configured hardware — is more defensible and usually more profitable. Software and cloud licensing procured through a marketplace distributor like Pax8, or hardware through Ingram Micro or TD SYNNEX, works the same way: the margin you set is compensation for procurement labor, vendor management, and license true-up work, not a hidden tax.

Mid-term adds: $25–$75 per user or device. This covers provisioning, account setup, agent deployment, and the administrative work of amending the agreement. Many providers skip it and simply prorate the recurring fee — that is fine and arguably cleaner, but it means the provisioning labor is unfunded. Pick one and be consistent.

Early termination: 50–100% of remaining contract value. This one deserves care. It is legitimate when you have front-loaded real cost — a discounted or waived onboarding, subsidized hardware, a longer amortization on tooling. It is indefensible as a pure retention lock, and courts in some jurisdictions treat unreasonable liquidated damages as unenforceable. The cleanest version ties the fee explicitly to unrecovered onboarding investment and steps down over the term.

Per-seat and per-device base rates vary too widely by market, vertical, and stack depth to give a single number responsibly, but the relationship to fees is worth stating: a provider with a robust fee schedule can competitively price the base contract, because the base does not have to carry every cost. A provider with no fees must load everything into the per-seat rate and then looks expensive on the comparison spreadsheet. That is the quiet competitive argument for building this properly.

Adjacent industries land in the same place, which is a useful signal that the structure is sound rather than an MSP quirk. Property management companies charge a base management percentage plus lease-up, maintenance-coordination, and renewal fees. Commercial cleaning charges a recurring contract plus deep-clean and emergency-callout fees. Equipment maintenance contracts charge a service plan plus parts markup and after-hours dispatch. The pattern — a predictable recurring base plus disclosed event-triggered add-ons — recurs everywhere recurring service meets variable demand, because it is the only structure that survives both slow months and crisis months.

Risks, edge cases, and failure modes

The unenforced fee. By a wide margin the most common failure. The fee exists in the MSA, appears in the PSA's rate table, and is waived in practice because a technician did not want an awkward conversation or a manager traded it for goodwill during a tense week. The revenue model assumed a trigger rate that never materialized. The fix is structural, not motivational: automate the rate application so it is applied by default, require a manager's explicit approval to *waive* rather than to *charge*, and report waived fee dollars monthly as a visible number. What gets measured as a leak stops leaking.

Scope creep disguised as goodwill. Related but distinct. Here the fee is not waived — the work is simply never classified as out-of-scope, because classifying it would require a conversation. Over a year this quietly converts a fixed-fee contract into unlimited-consumption, and it shows up as declining margin on a client whose contract value never changed. Watch cost-to-serve per client, not just contract value. Any client whose delivery hours have drifted 30%+ above their cohort is a scope conversation waiting to happen.

Surprise billing. The genuine reputational risk. An after-hours fee the client's office manager did not know existed, appearing on an invoice three weeks after the incident, generates a dispute that costs more in relationship damage and admin time than the fee was worth. Mitigation is procedural: state the fee at the moment of the request ("we can absolutely handle this tonight — that engages the after-hours rate at $175/hour with a one-hour minimum, want me to proceed?"), and get an affirmative yes in the ticket. This single habit eliminates most fee disputes.

Pricing yourself out at the top of the funnel. If your onboarding fee is materially above your market and you cannot articulate what it buys, you will lose deals you should have won. The defense is not a lower fee — it is a deliverable. Onboarding that produces a documented environment, a network diagram, a risk assessment, and a 90-day roadmap is worth $5,000 and demonstrably so. Onboarding that produces "we installed the agents" is not.

Margin illusion on procurement. The 85–95% margin claim breaks badly if you commingle pass-through hardware cost with handling-fee revenue in your books. A provider running $60,000/month of hardware through the P&L as revenue looks like a much larger, much lower-margin business than it is, and it distorts every metric downstream — gross margin, revenue per employee, and any valuation conversation. Book the pass-through cost and the handling fee separately in QuickBooks or whatever ledger you run, and know whether you are reporting gross or net revenue on resale. This one has real consequences at diligence time.

The compliance and regulatory edge case. Clients under HIPAA, PCI DSS, CMMC, or similar frameworks generate support work with a fundamentally different cost profile: evidence collection, audit response, attestation support, security questionnaires from *their* clients. Folding this into a standard managed contract is how MSPs end up doing hundreds of unbilled hours. Either price a compliance uplift into the recurring fee for those clients or define audit support as a fee-triggering category. Do not leave it ambiguous.

Contract and jurisdiction risk. Fee language that is vague, one-sided, or automatically escalating can be challenged. Auto-renewal clauses in particular are regulated in some jurisdictions and require specific notice. Have counsel review the fee sections of your MSA — this is a few hours of legal spend against a document that governs your entire revenue base.

The transition problem. Introducing fees to an existing book is harder than launching with them. Grandfathering everyone forever means the strategy never reaches most of your revenue; imposing fees at the next invoice cycle generates churn. The workable path is to apply the new schedule at each client's renewal, bundled with a broader value conversation — a business review, a roadmap, sometimes a modest base adjustment — so the client experiences a repositioned relationship rather than a bill increase. Expect this to take a full contract cycle, typically twelve to eighteen months, to work through the base.

Over-engineering the schedule. A fee list with fourteen categories is unbillable in practice and unsellable in a sales conversation. Four to six well-defined fees that everyone in the company can explain from memory will outperform a comprehensive schedule nobody applies consistently.

A practical rollout plan

Sequence matters here. Providers who start by configuring their PSA end up with beautifully automated billing rules for fees their contracts do not support.

Weeks 1–2 — model the math. Before touching software or contracts, run the arithmetic on your actual numbers: your client count, your realistic new-logo rate, your after-hours ticket volume from the last six months, your annual pass-through hardware spend. Compute the monthly and annual revenue per candidate fee and rank them. Most MSPs discover one or two fees carry the overwhelming majority of the opportunity — usually onboarding and procurement handling — and that clarity prevents building six fees when two would do. PULSE publishes a free Service Fees Calculator that runs this model in the browser if you would rather not build the spreadsheet.

Weeks 2–4 — write the scope. Rewrite the covered-services and exclusions sections of your MSA and your standard SOW template so each fee has a defined trigger. Name the covered hours, the covered environment, the excluded work categories, the billing increments, and the minimums. Have counsel review it. This step is the one people skip and the one everything else depends on.

Weeks 4–6 — configure the systems. Now build it in the PSA: rate tables, work types, time-of-day billing rules, agreement types, and the invoice presentation. Configure the automatic after-hours multiplier so no human decides it. Set up payment collection with card or ACH on file for onboarding deposits and project milestones. Create separate income accounts in your accounting system for each fee category and keep hardware pass-through cost segregated from handling-fee revenue.

Weeks 6–8 — train the humans. Every technician and account manager needs to be able to state each fee, its trigger, and its rationale without reading from a document. Role-play the awkward conversations — the client who asks you to waive it, the one who says the last provider never charged for this. Make the disclose-at-request habit explicit: the fee is stated when the work is requested, confirmed in the ticket, and never a surprise on the invoice.

Weeks 8–12 — launch on new business, then migrate. Every new proposal carries the full schedule immediately. Existing clients transition at renewal, paired with a business review. Track the transition as a pipeline with named clients and dates so it does not stall.

Ongoing — measure and correct. Monthly, report fee revenue by category against model, the trigger rate for each fee, waived fee dollars, and contribution margin. Quarterly, revisit the rates against market and against your own cost changes. Annually, re-scope: your service catalog will have drifted, and the exclusions list needs to keep pace.

The feedback edge from the quarterly review back to scope is deliberate. A fee schedule is not a document you write once. Your service catalog changes, your cost base changes, and the market moves — the providers who treat this as a standing operating rhythm rather than a one-time project are the ones whose margins hold up over a decade.

Related questions

Should an MSP charge for onboarding if the client signs a longer term?

Discounting or waiving onboarding in exchange for a 36-month rather than 12-month term is a reasonable trade. Recognize it as financing: you are amortizing real cost over more months. Pair it with a termination clause that recovers the unrecovered portion if the client leaves early.

How do you introduce fees to clients who have never been charged them?

Transition at renewal, not mid-term. Pair the new schedule with a business review, an updated roadmap, and a clear statement of what changed in scope. Expect the full base to migrate over twelve to eighteen months. Never impose new fees on an existing invoice without notice.

Do service fees hurt client retention?

Disclosed, scope-defined fees generally do not. Surprise surcharges do. The differentiator is whether the client knew about the fee before the work happened. Providers who state the fee at the moment of request and confirm it in the ticket see very few disputes.

What percentage of MSP revenue should come from add-on fees?

There is no universal benchmark worth quoting, and the right answer depends on your mix of project work and procurement volume. The more useful target is coverage: fee revenue should at minimum fund your non-billable delivery overhead — dispatch, NOC, documentation, and vCIO time.

Does a procurement handling fee apply to cloud and software licensing too?

Yes, and it works the same way. Licensing procured through a marketplace distributor carries real labor — provisioning, true-ups, renewal management, vendor escalation. Whether you express it as a markup on the license or a separate management fee, the underlying work is what you are charging for.

FAQ

What is the typical range for an MSP onboarding or setup fee?

Most providers charge a one-time onboarding fee between $1,500 and $5,000 per client, scaling with environment complexity. It covers discovery, documentation, agent deployment, monitoring and backup configuration, and integration into the service desk. Complex multi-site or heavily on-prem environments frequently price above that band as a scoped project rather than a flat fee.

How much should we charge for after-hours or emergency support?

After-hours rates commonly run 1.5–2× the standard hourly rate, landing around $150–$250 per hour, typically billed in 15-minute increments with a one-hour minimum for a callout. Define the covered-hours window and excluded holidays explicitly in the agreement, and distinguish client-requested after-hours work from provider-initiated remediation you caught yourself.

What is a reasonable hardware-procurement handling fee?

Five to fifteen percent of hardware cost is the standard band, covering sourcing, ordering, receiving, staging, asset tagging, and vendor management. Consider tiering it — a lower percentage on simple drop-ship, a higher percentage or per-unit staging fee on configured equipment — since a flat rate overcharges on easy items and undercharges on the labor-heavy ones.

Should project work be billed hourly or fixed price?

Fixed price is better for well-understood work like migrations, refreshes, and office moves, because it forces real scoping and gives the client a budgetable number. Hourly, commonly $125–$200, suits genuinely uncertain work such as troubleshooting an inherited environment. Either way, require a signed SOW or change order before work begins.

Is an early-termination fee enforceable?

It depends on jurisdiction and how the clause is written. Fees tied explicitly to unrecovered onboarding investment, subsidized hardware, or documented front-loaded cost — and which step down over the term — are far more defensible than a flat percentage of remaining contract value with no stated rationale. Have counsel review the language.

How do we keep procurement pass-through from distorting our margins?

Book pass-through hardware cost and handling-fee revenue in separate accounts. Commingling them inflates reported revenue and depresses apparent gross margin, which distorts revenue-per-employee, benchmarking, and any diligence conversation. Know whether you are reporting resale gross or net, and be consistent about it.

Sources

• https://www.connectwise.com/platform/business-management
• https://www.datto.com/products/autotask-psa/
• https://www.atera.com/pricing/
• https://www.ninjaone.com/pricing/
• https://halopsa.com/pricing/
• https://stripe.com/pricing
• https://www.pax8.com/
• https://quickbooks.intuit.com/pricing/
• https://www.ingrammicro.com/
• https://www.tdsynnex.com/

Related on PULSE

• [How Many Sales Reps Do I Need to Hire for My IT Managed Services (MSP) Business?](/knowledge/q15550)
• [How should a 2027 partner team segment VAR vs MSP partner archetypes?](/knowledge/q12530)
• [What Service Fees Should a Property Management Company Charge?](/knowledge/q16169)
• [What Service Fees Should a Tree Service Company Charge?](/knowledge/q16157)
• [What Service Fees Should a Septic Service Company Charge?](/knowledge/q16156)
• [What Service Fees Should a Junk Removal Company Charge?](/knowledge/q16155)