QA Manager to Head of QA: The Escape Bill That Wins Budget

TL;DR A QA manager who reports counts nobody acts on, or asks for headcount with "we're overloaded", loses the room: bugs-per-tester turns collaboration into hoarding within a week, and "we're overloaded" loses to a competing roadmap every time. Report outcome metrics that lead with a judgment, and make the budget case as arithmetic: FieldFlow's escapes cost about £23k a quarter against £15k for the SDET who would prevent most of them.

Do this: Pull last quarter's escapes, price each one as support hours plus engineering firefighting, and write one sentence comparing that bill with what prevention would cost.

Part 5 ended with you as QA Manager, signing FieldFlow's releases, and two asks from above: a director who wants every tester to log 5 bugs a day, and your own case for one more SDET with last quarter's 11 escapes in the records. What do you put in front of them?

Not a tester leaderboard, and not a plea. A number with a pound sign on it.

QA Career Path, in seven parts:

1. The QA Career Path: Map Yourself by Scope, Not by Title
2. New QA Engineer? Test the Risk, Not the Ticket
3. Manual to Automation Testing in Six Months: Skip the Tutorial Loop
4. Senior QA: Depth Everywhere Is Depth Nowhere
5. After Senior QA: Lead the Release or Own the Platform?
6. QA Manager to Head of QA: The Escape Bill That Wins Budget (you are here)
7. Will AI Replace Testers? Make Your Judgment the Evidence

Your output is decisions now

A QA Manager's week has very little hands-on testing in it: triage, 1:1s, a release readiness review, an interview, the weekly status report. The output is decisions, plans, unblocked people and clear communication.

Refuse the anti-metric, answer its question

Every metric should answer a question someone is actually asking. When leadership asks for an activity count, offer the outcome metric behind the real question.

The rule underneath is Goodhart's law: when a measure becomes a target, it stops being a good measure. So the director's 5 bugs a day gets a reply, not a yes: "you want to know if we're testing enough; here's escape rate and risk coverage."

Then lead every report with the judgment: "Release is at risk: invoicing regression is 60% done with 2 Sev-2s open" beats a table with no conclusion. And never bury a Sev-1. Credibility is your whole currency.

Price the escapes

The argument that wins budget is arithmetic. Here it is on FieldFlow's fictional numbers from the repo: 11 escapes last quarter, 3 in invoicing.

Average escape    ~12 support hours × £25  +  ~2 engineer-days × £350   ≈ £1,000
Invoicing escape  the same, plus customer recovery work ≈ £4,000        ≈ £5,000
Failure cost      8 × £1,000  +  3 × £5,000                             ≈ £23,000
Prevention        one SDET on invoicing and API regression              ≈ £15,000
Projected effect  no invoicing escapes, and half the rest

Presented as one sentence: each quarter we choose £23k of failure over £15k of prevention, and here is last quarter's evidence. Substitute your own rates; the shape is what carries.

At Head of QA, the organisation is the product

A Head of QA has effectively zero hands-on time. The work is org design, budget, growing managers and leads, and the career framework itself, including the IC track from part 5. The repo's default shapes:

1. Up to 3 squads, one product: embedded testers and one QA Lead.
2. 4 to 8 squads on a shared platform: embedded testers plus a QA chapter that owns standards, tooling and the hiring bar.
3. Many products or heavy compliance: embedded testers plus a small central team for standards and audits.

FieldFlow, at 6 squads, lands on the second row. Its 8 testers to 40 developers is 1:5, inside the 1:4 to 1:8 band the repo calls common. Benchmarks open that conversation; your escape data closes it.

Try it

Here are last quarter's production escapes, with rough support hours,
engineer-days and any customer recovery work for each: <list>.
Using my rates (<support £/h>, <engineer £/day>), price each escape,
total the quarter, and write one sentence comparing it with <the prevention I want>.
Flag any number I gave you that looks like a guess.

The SDET gets funded. Then the tools move under everyone's feet: AI drafts test cases and test code in seconds, and a vendor demos an agent that "tested all of FieldFlow in 20 minutes, 0 failures". If tests are that cheap to produce, what is a tester's work worth, and how does anyone, at any level, show it to the next employer? Part 7, Will AI Replace Testers? Make Your Judgment the Evidence, closes the series there.

Read this with your AI

Read the post "QA Manager to Head of QA: The Escape Bill That Wins Budget".
1. Summarise it in 5 bullets for a new QA manager.
2. Ask me 3 questions to check I understood; wait, then correct me.
3. Draft my reply to a director who wants each tester to log 5 bugs a day.