See the method before you buy anything

Fictional company. Not client work. Every company detail, figure, finding and calculation below is invented. They exist to demonstrate the structure and analytical standard of a Support Operations Waste Assessment. Nothing here is a case study, a testimonial, a benchmark, or a claim about results achieved for anyone.

Why this exists

A real client report cannot be published. It would be redacted into uselessness, and the parts that matter most — the arithmetic, the confidence grading, the findings we argued down — are exactly the parts that would go. So we built one on a company that does not exist, which means nothing has to be hidden.

Read it the way a sceptical CFO would: check whether the arithmetic reconciles, whether the confidence grading is honest, and whether the recommendations actually follow from the evidence. That is the scrutiny a real engagement has to survive, and it is the only kind of proof worth offering before there is a track record to point at.

Illustrative company: Northstar Learning Systems — a fictional 128-person B2B software company with 16 frontline support agents and 4 support operations specialists.

The one number, and its qualification

Adding up everything found at Northstar gives $114,186 a year. Discounting it for weak evidence and for how much you would realistically recover gives $72,737.

That gap is the point. A firm selling on headline numbers would quote the first figure. Three separate numbers are kept apart here so you can argue with the analysis without first having to untangle an inflated claim:

Of that final figure, $19,200 is software you could stop paying for — a direct budget reduction the month the licences are cancelled. The remaining $53,537 is time, which only becomes value if the work is genuinely removed and the freed hours are put to use. It assumes nobody is let go, and it does not claim your payroll will fall.

The findings register

Note what finding 5 does. It carries the largest gross exposure in the register and the weakest evidence, so it is ranked last and its payback is explicitly not claimed. A firm optimising for headline numbers would lead with it.

One finding in full

Finding 02 — Routing mismatch creates daily manual reclassification

The routing configuration sends a recurring subset of tickets to the wrong queue or applies an unusable classification. Agents correct the queue and tags manually before productive work begins.

16 agents
× 3.2 affected tickets per agent per day
× 2.8 minutes per ticket
× 240 working days
÷ 60
× $46 loaded hourly rate
= $26,378/year          (573 hours of annual capacity)

Confidence: High. Ticket-event data provides the reclassification count and elapsed handling time. A workflow walkthrough confirms the changes are corrective rather than part of intended handling.

Root cause. Routing rules were added incrementally while the tag structure grew without retirement rules. The system reflects historical exceptions rather than the current support model.

Recommendation. Identify the top three misrouting patterns by volume; correct the deterministic routing rules; reduce the active tag set to decision-relevant classifications; add a weekly exception report for four weeks; assign one accountable owner for rule and taxonomy changes.

Decision-case value. $26,378 × 70% (the share we think you would realistically recover) = $18,465/year. Against roughly 40 internal hours of effort at $55, payback lands at 1.4 months.

Decision: do first. Fix the deterministic problem before considering automation or a platform change.

The 90-day sequence

Software goes first because it is direct cash, high confidence, and requires no behaviour change. Knowledge search goes last because the evidence must improve before investment expands.

The measurement rule. No labour finding is reported as realised value until the client observes a lower frequency or duration for the targeted work, stable performance across the agreed window, and a named use for the recovered capacity.

The decline list

Restraint is part of the assessment. These actions are not recommended, and the reasoning is stated:

How evidence is graded

Two different questions are being asked, and they are kept apart deliberately. Confidence asks: how well evidenced is this? Recovery asks: how much of it would you realistically get back once people actually change how they work? This specimen assumes 100% recovery on cancellable software spend, 70% on well-evidenced labour, and 60% on moderately evidenced labour after discounting. These are stated planning assumptions, not industry benchmarks — we do not have industry benchmarks and will not pretend to.

Working rules applied throughout: system data establishes what occurs and interviews explain why; a finding is downgraded when quantification cannot be independently reproduced; no interview statement alone earns High confidence; no value is counted twice across findings.

Limitations, stated plainly

1. This document is entirely invented and cannot demonstrate market proof or client results.
2. Annualised labour exposure is not equivalent to payroll savings.
3. Recovered time creates value only when the work is removed and the capacity redirected.
4. Implementation effort is illustrative and would require validation by the client's system owners.
5. No benchmark data is used.
6. No company-wide conclusions can be drawn from a support-only assessment.

What a live engagement produces

An executive findings brief; a traceable findings register; direct software-waste analysis for the support stack; quantified analysis of recurring support workarounds; confidence and evidence grading; a prioritised 90-day recovery plan; a live executive readout; and an appendix containing formulas, evidence references, assumptions and excluded items.

Delivered over fifteen business days after complete required data is received.

Next step

If the reasoning above holds up to your scrutiny, that is the best available basis for a conversation. Request a fit conversation and we will tell you whether an assessment on your actual operation is worth the fee — including when it isn't.