Lesson 5 of 5
Testing implications and choosing an option
10 min
Listen to this lesson
Each deployment option creates a different regression demand: how often the system changes underneath you, how much of it you own, and how much time you get to prove it still works. This lesson turns the earlier dimensions into a test planning view and closes with a way to recommend an option for a scenario.
Regression demand by option
In the public edition, regression is frequent and fixed: two major upgrades a year, each with a test window of about three weeks, plus continuous feature delivery. You cannot postpone it, so the answer is automation. SAP recommends its Test Automation Tool for the public edition, and the Release Assessment and Scope Dependency tool and the What's New Viewer help identify which changes affect your scope.
In the private edition and on-premise, regression is less frequent but larger, and its timing is yours: each feature pack or support pack stack, each base release upgrade, and every one of your own releases. The risk shifts from "not enough time" to "too much change at once", especially when custom code at levels C and D must be retested. On-premise adds technical changes you run yourself, such as operating system and database updates, which also need at least a technical smoke test.
Automation matters in the private edition too, just for a different reason. A regression set that runs only every year or two decays: test data expires, users change and nobody remembers the steps. Contoso plans to keep its key process tests in SAP Cloud ALM, run a small automated core set after every transport cycle so it stays alive, and expand to the full set for stacks and upgrades. Kenji's team adds ABAP Unit tests to the custom code it keeps, so the core return agreements logic is checked every time it changes rather than only at upgrade time.
What stays the same everywhere
- The customer always tests and accepts. No option moves business sign-off to SAP.
- A risk-based regression set built from your processes, not SAP's whole scope, is the core asset.
- Integrations and extensions are the most fragile parts and deserve contract tests.
- Evidence must be stored where auditors can find it, in SAP Cloud ALM or your test management tool.
Recommending an option
- Must existing custom code and history be kept largely as they are?Yes → private edition or on-premise via conversion · No → next question
- Do the processes fit SAP's standard and released extension points?Yes → public edition is a strong candidate · No → private edition or on-premise
- Does the company want to stop running infrastructure and technical operations?Yes → private edition (RISE) over on-premise · No → on-premise remains possible
- Can the organization test twice a year on SAP's schedule?Yes → public edition is workable · No → build that capability first or prefer customer-scheduled options
Contoso's main program answered the first question with "yes", so it converted to the private edition. Wave 2 raised the question again: after the conversion Contoso acquires Alternadores del Norte S.A. de C.V. in Monterrey. The team considered a two-tier setup with a public edition system for the subsidiary, then chose to load Alternadores' legacy data into the converted system as company code 1910, so that intercompany flows and reporting stay in one system. Either answer could be right; what matters is that the option was compared on the same five dimensions.
What would you do?
A US distributor with 300 users, mostly standard processes and no SAP history asks Contoso's architect for advice before a new implementation. The company has a small IT team and no wish to run servers. Which option should the architect lean toward?
- Public edition: frequent, fixed regression windows; automate with SAP's Test Automation Tool and scope-based impact analysis.
- Private edition and on-premise: less frequent, larger, customer-scheduled regression; custom code drives the size.
- The customer always tests and signs off, in every option.
- Recommend an option by comparing scope fit, extensibility, cadence, operations and testing capability.
Check your understanding · not graded
Lesson check
1. Which SAP tool is recommended for regression testing in the public edition?
2. Why does custom code at clean core levels C and D increase regression effort in the private edition?
3. What does every deployment option have in common for testing?
4. Why did Contoso load Alternadores del Norte into the converted system instead of a separate public edition system?