The model allows 120 labor minutes for a ticket resolved at Tier 3 and 30 at Tier 1. The difference is 90 minutes, under those calibrated assumptions. A shift left service desk makes suitable work solvable by the least costly qualified resource, through a tested procedure or automation. Changing the queue label doesn't save the minutes.
My test is simple: what does this do to the board on Monday morning? The technician taking the next ticket needs a working fix, clear authority, and a way to escalate. A new label won't answer the phone.
Measure Where the Labor Goes
Start with the tier that resolved each ticket and the labor used across everyone who worked it. Don't assume the final person to click Close did all the work. Keep waiting time separate from technician labor, and keep the logging rule consistent across the comparison.
Say you run the model's 700-seat US example at its well-run rate of 0.5 tickets per seat per month. That produces 700 × 0.5 = 350 tickets a month; the typical range is 1.0 to 1.4 tickets per seat per month. Its calibrated tier mix gives the following workload.
| Resolving Tier | Share of Tickets | Labor Minutes per Ticket | Monthly Ticket Labor Hours |
|---|---|---|---|
| Tier 1 | 65% | 30 | 700 × 0.5 × 65% × 30 / 60 = 113.8 |
| Tier 2 | 22% | 60 | 700 × 0.5 × 22% × 60 / 60 = 77.0 |
| Tier 3 | 13% | 120 | 700 × 0.5 × 13% × 120 / 60 = 91.0 |
The last row deserves attention. In the model, 13% of tickets consume about 32% of ticket labor hours. That is a reason to inspect recurring escalations, not proof that your senior technicians are doing work someone else could safely handle.
For your desk, choose a sampling period that covers the work you want to change and the technicians who handle it. Check whether the sample includes the relevant ticket types rather than borrowing a fixed number of weeks. If those cases haven't appeared, extend the sample or narrow the decision you are trying to make.
Give the Baseline a Named Owner
I'd have the person responsible for the desk sign off the baseline before moving work. Use the export and a ticket sample together. A report with missing time entries can't tell you that the work was quick.
- Desk lead: choose the reporting period and export completed tickets with their resolving tier, category, client, and logged labor.
- Technicians: check the sampled records for time spent before escalation or after a handover. Correct missing entries through the normal recording process.
- Desk lead: count tickets by resolving tier and divide each count by the total to get the split.
- Desk lead: total the logged labor by ticket type and compare it with the counts. Keep unusual or specialist work visible rather than hiding it in an average.
- Owner: confirm which recurring ticket type is worth testing, who can write the fix, and what work they will put aside to do it.
Look beyond the split when the hours don't make sense. More incoming tickets or longer handling times can increase workload without changing the tier percentages. The model's move from the well-run 0.5 rate to a typical 1.0 costs its US example about 7.8 percentage points of EBITDA, holding the other inputs constant.
Choose Work That Can Actually Move
I'd leave the growth-hack language out of this. Name the ticket type and what the next technician must be able to do. Keep work with a qualified person until the procedure and authority are ready.
| Possible Change | Who Prepares It | What to Check Before Routing Work |
|---|---|---|
| Document a repeat fix | The engineer who understands the resolution | A receiving technician can follow the steps and knows when to stop or escalate |
| Automate a repeat action | The person responsible for that automation | The action is appropriate for the supported environment and its result is checked |
| Standardize the client build | The technical owner of the standard | The client is on the intended build and the team is trained to support it |
Begin with a recurring category whose resolution can be explained and checked. Ask the senior technician whether the difficult part is diagnosis, permission, or execution. A written execution step won't remove a judgment the receiving technician is not equipped to make.
Keep the chosen change small enough for the people doing the work to review it. Set the pace from the cases arriving, the time available to prepare the fix, and the observed result of the previous change. Don't promise a monthly quota before you know what the work requires.
Price the Change as a Sensitivity
The model prices an alternative resolving split of 70% at Tier 1, 24% at Tier 2, and 6% at Tier 3. Its modeled US example gains $108,944 of EBITDA, or about 3.0 percentage points, relative to the 65%, 22%, and 13% base split. EBITDA is earnings before interest, taxes, depreciation, and amortization.
That is the effect of the alternative mix with the model's staffing and pay mechanics. It isn't a cash saving to claim after relabeling tickets. Keep the incoming ticket count and the work performed visible while you examine whether your own changed procedure uses fewer labor hours.
The 90-minute difference comes from 120 minus 30, the model's Tier 3 and Tier 1 averages. A particular ticket may use different labor, including preparation and follow-up. Compare its actual result with the old method before multiplying a modeled difference across the whole category.
Check Resolution Before Counting Capacity
Have the receiving technician try the documented procedure on suitable work with the specialist available to help. Record where the instructions were unclear and what authority was missing. The specialist should review the result before the desk lead changes the routing rule.
Watch for reopens and work returning to the senior queue. A higher Tier 1 closure share doesn't establish that the original problem was solved. Check the resolution and all recorded labor before calling those hours available.
On a small pooled desk, freeing hours doesn't immediately remove a salary. It gives the same people capacity for other work until it supports a real staffing decision, such as postponing a hire. Have the person assigning tickets show what can now fit on the board.
Set the Next Change From the Result
Ask the desk lead to bring the tested ticket type back with its outcome and labor record. If the receiving technician still needs specialist judgment, change the procedure or keep that part at the higher tier. If the fix works, record the routing decision and who maintains the instructions.
Use the observed results and the team's available capacity to choose when another category should move. Keep one standard technology stack and train the team on it, as the model recommends. The next improvement starts with work a named technician can now finish correctly.
mspoutpost.com/articles/shift-left-service-desk-move-work-with-a-tested-fix
Written by Megan Friesen
MSP Outpost publishes information, not legal, financial, insurance, tax or compliance advice, and reading it creates no professional relationship. For decisions about your business, talk to a qualified professional.