Lesson 1 of 6
Why sign-off fails
10 min
A sign-off is a promise. When a process owner signs that order-to-cash is ready, the business is told it can stop worrying about it. Most SAP programs that struggle after go-live did not lack testing; they lacked a sign-off that meant something. This lesson looks at why sign-offs fail so you can recognize the pattern early.
Three ways sign-off goes wrong
| Failure | What it looks like | What it costs |
|---|---|---|
| Vague criteria | "Billing works." "No major issues." "Users are happy." Nobody wrote down what correct looks like, so any result can be called a pass. | Defects are argued about instead of fixed; the process owner cannot defend the signature. |
| Missing evidence | A test log full of "Pass" and screenshots of success messages, with no document numbers, statuses or journal entries. | Nobody can tell whether the business result happened; problems surface in production. |
| Timeline pressure | Exit criteria are relaxed the week before the gate because the date is fixed. Open defects are reclassified rather than fixed or accepted. | Known risk goes live without an owner or a workaround. |
The three common causes of a sign-off that does not hold.
All three share a root cause: the definition of "done" was left until the end. When criteria are written after testing, they are shaped by what testing happened to find. When they are written before, testing is shaped by them.
Timeline pressure deserves a special word, because it is the one process owners feel most directly. Go-live dates are usually fixed months in advance and tied to business events: a fiscal year start, a contract expiry, a data center exit. As the date approaches, every open item looks like a threat to it, and the easiest lever to pull is the definition of done. The defense is to agree criteria early, in writing, with the same steering committee that owns the date. Then, when pressure comes, the conversation is about fixing or accepting specific risks, not about rewriting the rules.
Missing evidence is often a habit rather than a decision. Testers capture what is easy to capture, usually the message bar, because nobody told them what else was needed. A short evidence standard, issued before execution, fixes most of this at almost no cost.
The cost of late discovery
The later a defect is found, the more it costs, because more has been built on top of it and more people are affected. In SAP, a wrong revenue account found in a unit test is a configuration fix. Found in user acceptance testing, it means re-running scenarios. Found after go-live, it means wrong postings in the books, reconciliations, perhaps restated reports and a conversation with the auditors.
- DesignA missing requirement costs a conversation
- Build and unit testA configuration fix and a retest
- Integration testFix, retest the chain, re-check dependent scenarios
- User acceptanceFix, retest, re-sign; schedule pressure grows
- ProductionWrong documents in the books, corrections, reconciliation and audit questions
Late discovery also costs trust. When business users find basic defects in acceptance testing, they stop believing the system is ready, and their confidence is hard to rebuild before go-live. Clear gates protect them from that experience.
What a good sign-off looks like
A sign-off that holds has four properties. It is defined in advance: criteria were agreed before execution. It is measurable: each criterion can be judged pass or fail by someone else. It is evidenced: the pack shows the business result, not just a save message. And it is honest about risk: open defects are listed, with an owner, a workaround and a due date, and someone with authority accepted them.
The rest of this course builds each property. Lesson 2 writes acceptance criteria; lesson 3 defines entry and exit criteria for each test phase; lesson 4 sets evidence standards; lesson 5 handles open defects; lesson 6 assembles the sign-off record.
What would you do?
Two weeks before go-live, Grace (PMO lead) proposes changing the UAT exit criterion from "no open high-severity defects" to "no open critical defects" so the gate can be passed on time. Three high-severity billing defects are open. What should Dan, as process owner, do?
- Sign-off fails through vague criteria, missing evidence and timeline pressure.
- Define "done" before testing starts, not after.
- Defects cost more the later they are found; in production they reach the books.
- A good sign-off is defined in advance, measurable, evidenced and honest about risk.
Check your understanding · not graded
Lesson check
1. Which is the common root cause of failed sign-offs?
2. Why does a defect cost more when found after go-live?
3. Which is an example of vague sign-off criteria?
4. What is the right response to pressure to relax exit criteria the week before a gate?