A Successful Demo Does Not Always Mean a Successful Handover

A delivery team demonstrates a feature. It works. The customer tries it in their own environment and cannot complete the same workflow.
From the customer’s perspective, the delivery is not ready. From the team’s perspective, the feature is implemented, but something is preventing the customer from using it. Progress becomes difficult when those two views are discussed as if they answer the same question.
The situation below is an anonymised illustration based on delivery experience. It focuses on the acceptance process and does not identify a customer, project or individual.
Three questions need separate evidence
- Does the agreed functionality work?
- Does it work in the customer’s environment?
- Can the customer use and maintain it without the developer guiding every step?
Each question matters. A successful demonstration provides evidence for a particular build, setup and sequence of actions. Acceptance also needs evidence that the agreed workflow can be completed under the conditions the customer will actually use.
Separate application behaviour from environment dependencies
An environment or connectivity issue can prevent acceptance testing before the feature itself is reached. Network settings, virtualisation, device capabilities and external services can all affect the workflow. A hardware limitation or fault can also affect software behaviour, but it needs evidence before being treated as the cause.
At the same time, a successful build does not prove that the application behaves as agreed. An environment problem and an application defect can exist together. Resolving one does not automatically close the other.
Record the build, operating environment, steps, expected result and actual result. Compare the same scenario across environments where practical. If a problem disappears after several settings change, record the recovery while keeping the root cause open until the evidence supports a conclusion.
Broad requirements can hide different acceptance criteria
Even with a specification, a capability can leave its exact behaviour open. “Restore the previous state” sounds clear until one person means returning from another screen and another means recovering after the application has been terminated.
Both interpretations can be reasonable. They require different implementations and tests. When there is no detailed specification, these gaps can become larger, and the depth of checking can differ between the team and the customer.
A practical way forward is to turn broad requirements into observable scenarios:
“Make an unsaved change, leave the screen, return, and confirm that the change is still there.”
If recovery after application termination is also required, write that as a separate scenario. Agree which scenarios belong to the current delivery and which represent additional scope before promising a completion date.
Make the acceptance record usable by both sides
- Identify the build and environment used for the check.
- Write the starting conditions, steps and expected result.
- Record whether the scenario passed, failed or could not be tested.
- Keep application defects, environment blockers and scope questions distinguishable.
- Agree who will provide the missing evidence or decision, and when the check will be repeated.
A blocked check is not a passed check. It also does not, by itself, establish an application defect. Recording that distinction helps both sides focus on the next useful action instead of debating a single overall “done” label.
Include independent use in the handover
For a handover, add one more test: ask the receiving team to complete a representative task using the documentation, without the original developer leading the process. That can reveal missing setup steps, assumptions or documentation gaps that a developer-led demonstration will miss.
Acceptance becomes easier to discuss when both sides can point to the same scenario, the same conditions and the same expected outcome. Unresolved items can then be recorded explicitly, with an agreed owner and next step.
“What does your team use as the final acceptance test: a demonstration, a checklist, or the customer completing the workflow independently?”
Sergey Lapidus
CEO
Sergey leads Sparkler Soft and works with clients on product direction and delivery.
LinkedIn →Have a similar challenge?
Tell us about your project — we'll tell you honestly whether we can help.