Kit 01: managed services
Managed Services Pricing and Service Design Kit
Price a managed service from what it costs to serve, package it in tiers, write the promises and fair use down, and keep every agreement profitable after it is signed.
This kit is how I would price and design a managed service today: build the price up from what each client costs to serve, package the service in tiers, and write the promises, the fair use and the terms down, so an agreement is still profitable a year after it is signed. It is for managed service providers running IT for small and mid-market businesses (users, devices, servers, sites, security, Microsoft 365 and backup), and for consultancies, agencies and IT services firms adding a recurring service beside their project work.
The price comes first. The playbook sets out the models in use (per user, per device, per site, good-better-best tiers, all-inclusive, a la carte, value-based, and the retainer for a professional services firm) and one rule for choosing: price on a unit that moves when your effort moves and that the client can count. For most small and mid-market clients that is the user, with servers and sites as add-ons. The fee is built from the cost to serve (labour at your real cost per productive hour, the tool stack, overhead and a small buffer), turned into a floor price at your target margin, and only then checked against the market. It is never a percentage of what the client spends with third parties.
Then the promise. Response times are committed, because you control them, and resolution times are published as targets. All-inclusive never means unlimited: fair use is measured against a line both sides can see, and the answer to a heavy client starts with the data and a conversation, never a surprise invoice. Every included hour is visible to the client: what was used, what is left and what it went on.
Then running it: onboarding charged as the first delivery, with a service start date; the service terms and a client service guide as the authority on scope; every agreement's margin read monthly and every price reviewed yearly, on a calendar; and a priced recurring option in every project proposal.
It is written from how I ran DevOpsGroup, from my notes and my recollection. The recurring share of revenue there more than doubled between FY18 and FY21, and the last chapter is about what moves that number: how you sell the service, and how you pay people to sell it. Every default in the model is invented and yours to change.
What you will get
Who it is for
Questions
- Should an MSP price per user or per device?
- Price on the unit that moves when your effort moves and that the client can count. For most small and mid-market clients that is the user, covering their devices up to a stated number, with servers and sites priced as add-ons. Per device suits estates where devices rather than people drive the work, such as shared terminals or many sites. The kit sets out both, and the other common models, with where each fits.
- How do you set a managed services price that holds its margin?
- Build it from the cost to serve: labour at your real cost per productive hour, the tool stack line by line, an overhead share and a small buffer. Divide by one minus your target margin to get the floor, round up, then check it against the market. The market check can move a price up; it never moves it below the floor. The model then reads every signed agreement's margin and effective hourly rate each month.
- Does all-inclusive support mean unlimited support?
- No. All-inclusive means in-scope support is included within reason, and the kit makes "within reason" a line both sides can see: the monthly hours at which the effective hourly rate falls to your rate card. A heavy month is noted, not billed. A client running above the line over several months gets the data and a conversation first, and any change in price comes with notice.
- Is the kit only for MSPs?
- No. The same method prices a retainer for a consultancy, an agency or an IT services firm: recurring deliverables plus an included allowance of hours, built from your cost to serve, with work beyond the allowance quoted from the rate card. The worked example includes one.