Most of what support costs isn't in the support budget.

You already suspect some of this

• Ticket volume rises roughly in step with headcount, and nobody can say which portion of it was avoidable.
• The same twenty questions are answered every week, by people who each hold a slightly different version of the answer.
• A refund, an escalation, and a bug report take three different paths through the company. One of them is documented.
• The support tooling grew a piece at a time — a helpdesk, a chat widget, a knowledge base nobody edits, and two integrations that partly overlap.
• Engineering time disappears into escalations and is never counted as a cost of support.
• When someone senior leaves the support team, a set of judgments leaves with them.

None of this is urgent in any given week. That is precisely why it survives every budget review.

Why the real cost stays hidden

Very little of what support costs actually sits on the support line. It is spread out: escalation hours sitting inside engineering, software licences charged to IT, ramp-up time counted as headcount, credits and refunds taken off revenue. Each piece is defensible on its own. Nobody owns the total.

That has a specific consequence for whoever signs off the budget. When support asks for three more people, there is no way to test the request — because the alternative to hiring has never been priced. The question "would fixing the process cost less than three salaries?" cannot be answered, so it stops being asked.

You cannot make a decision about a number you do not have.

Get a first estimate in ten minutes →

How we work, and what it commits us to

Evidence before opinion. Findings come with the working shown — ticket data, time estimates, the arithmetic. If a claim cannot survive your own finance review, it does not go in the report.

Support operations only. A narrow scope is what makes depth possible. We would rather be precise about one function than approximate about six.

Implementation stays with you. We diagnose and sequence; your team builds. It means our findings have nothing to sell you afterwards, and it means every recommendation has to be executable by your own people, because they are the ones who will execute it.

We will tell you what not to fix. Some waste is cheaper to live with than to remove. A recommendation you can trust has to include the things we think you should leave alone.

What gets examined

Six sources, read against each other rather than in isolation. Most of what matters shows up in the disagreements between them.

• Ticket and conversation data — volume, drivers, repeat contacts, reopen rates, handling time by category
• Workflow and handoff paths — how a request actually travels, including the routes nobody drew
• Tooling — what is licensed, what is used, what overlaps, what is paid for twice
• Documentation — what is written down, what is current, what lives only in someone's head
• Escalation into engineering — the hours that leave support and are never booked against it
• Decision points — where work waits on a judgment only one or two people can make

Each finding is traced to a cost with the method shown, and rated for how confident we are in it. Estimates are labelled as estimates.

What you receive

1. A priced account of the waste — each pattern, what it costs annually, and how that figure was derived.
2. A prioritised shortlist — what to fix first, sequenced by return against effort and risk, with a suggested owner for each.
3. An explicit do-not-fix list — what we found and think you should leave alone, and why.
4. A decision briefing — a working session with you and your support leadership to walk the findings and pressure-test the sequence.

Written for a finance audience. Specific enough for your support leadership to act on.

Request a fit conversation →    Read the worked example first →

Why us

Mumford Advisory is led by Victoria Mumford — roughly twelve years in customer-facing operations, ten of them in technical support and problem-solving, including nearly five years at Shopify and, currently, complex technical escalations at Thinkific.

Escalation work is where support waste becomes visible. The job is to determine whether a problem is a platform defect, a configuration error, a broken integration or a third party — by reconciling log data, system records and human accounts that rarely agree, and by being honest about which parts the evidence cannot yet settle. Do that for a decade and the same shape keeps appearing: teams quietly absorb friction with recurring manual work, and the cost of that work never reaches the people who set budgets.

That gap is what this practice exists to close. The skill being sold is the same one the day job requires — enter an unclear situation, reconcile evidence across systems, separate what is confirmed from what is assumed, and produce a number someone can act on.

Who this is for

A good fit is a software company where support has grown faster than the process around it, where a CFO or equivalent can act on the findings, and where someone in support leadership can give us straight answers about how the work really runs.

We would turn the work down if you need someone to carry out the fixes, because we hand them to your team; if the trouble is clearly product quality, because a different diagnosis applies; if support is under about eight people, because the pattern is usually visible without us and we will say so; or if the data cannot be pulled, because the engagement runs on evidence rather than anecdote.

We would rather tell you this on a first call than three weeks into an engagement.

Start where you are

Three doors, in ascending order of commitment. None of the first two costs you anything, and none of them puts you on a list.

Want to judge the method? Read the worked example — a full assessment on invented data with every calculation exposed. Twenty minutes, no email.

Want a number for your own operation? Run the Workaround Tax Calculator. Ten minutes, no email, arithmetic printed on screen.

Ready to talk? Request a fit conversation. Thirty minutes. We will tell you honestly whether an assessment is worth your money — including when it isn't, which is a real outcome and happens.