Help Desk Cost per Ticket: Check Your MSP Labor Baseline

Grab the Numbers You Already Have

You need two numbers, and you already own both. Ticket counts come from your PSA: export the closed tickets for the same period, total, and then again grouped by client. Payroll comes from your bookkeeper: the burdened cost of the desk staff whose work those counts cover, meaning salary plus the employer costs and benefits included on that payroll basis. Keep unrelated roles out of this desk-payroll measure and account for them separately in the full managed-service margin calculation.

Start with exports and a question to your bookkeeper about the payroll basis. Reconcile the period and covered team before dividing.

The Simple Formula and the One That Actually Helps

The company-wide number is your starting point. In the US model, $321,304 divided by 12 gives about $26,775.33 of monthly pooled desk payroll. Divide that by 350 tickets and you get about $76.50 per ticket across the book of business. That measure includes the pooled desk's burdened payroll, not tools or the other delivery roles.

Now take that same $76.50 and multiply it by each client's closed ticket count for the same period. That allocates pooled desk payroll according to ticket share; it does not prove what the client's tickets actually cost. A short request and a long investigation receive the same allocation, so check recorded time and the delivery work outside tickets before calling the client unprofitable.

This is where the number earns its keep. Not as a price for a single ticket, but as a ruler you can hold against every client you have.

Rank Your Clients for a Gross Margin Review

Build a spreadsheet with one row per client. Start with monthly contract revenue, the desk-payroll allocation and the difference between them. Sort to find accounts worth investigating, then use their actual delivery costs to calculate gross margin, including burdened delivery labor and tools. The model's managed-service target is 50% on that fully loaded basis; the ticket allocation alone cannot establish it.

For the model's 20-seat client paying $4,500 a month, its well-run rate of 0.5 tickets per seat per month gives 20 × 0.5 = 10 monthly tickets; a typical MSP runs 1.0 to 1.4 per seat per month. Using the rounded desk baseline, the pooled desk-payroll allocation is 10 × $76.50 = $765. The remaining amount is $4,500 - $765 = $3,735 before tools and other delivery costs. That is a screening calculation, not the client's gross profit.

One caveat before you start firing anyone: a single month of high ticket volume can brand a client unfairly. One bad month, one project that spiked the count, and a healthy client looks like a loss-maker. Inspect low-ticket-volume months too, and reconcile a representative period before acting on an apparent loss. Keep one-time project or hardware revenue out of the recurring comparison, so a separate sale doesn't hide the service result.

Before You Fire the Client, Check Your Own Workflow

A negative-margin client is a symptom, and the cause is not always the client. Before you reprice or offboard, look at whether your own desk is inflating the cost.

Pull the ticket log for that client and look at what happened to their tickets. Set review triggers from the agreed response commitments and the actual workflow, including who owns waiting tickets. Check whether a tech kept spending time re-opening and chasing them. Those are workflow costs you can investigate rather than assuming the client caused them.

My view: fix your side first. If a recurring password-reset request can be handled safely through an approved self-service process, that's work your team can reduce without changing the client relationship. Improve the documentation, make the knowledge base usable and automate repeat work where it fits. None of that changes what the client pays, but it may change the work they require; measure whether workload and actual delivery cost change.

What to Do With the Unprofitable Ones

If the client's full delivery cost still exceeds the recurring service revenue after you check the records and address internal bottlenecks, then it's time to talk about the relationship. Not offboard yet, talk.

Start with scope. What does the contract actually promise, and what is the client actually using? If their users need constant hand-holding on things a documented fix would solve, a higher-priced tier that matches the service level they consume is a fair conversation, not a betrayal.

If the ticket volume is structural, if their environment or their users genuinely require the hours they consume, then the price has to reflect it or the relationship is a donation. Renegotiate the tier, reduce the scope to what they pay for, or, if neither is possible, offboard them cleanly. A client who loses you money every month is not a client, they are a liability, and keeping them funded by your profitable accounts is a decision you are making every month without meaning to.

Monday's first step is simple: pull last month's ticket counts and burdened desk payroll, build the spreadsheet, and rank the accounts for investigation. Check the apparent outliers against recorded work and a representative history before deciding which clients are funding the rest of the book.