The S/4HANA migration by the numbers, and what they mean for your test plan

Statistics about the S/4HANA move get quoted a lot and checked very little. This guide collects the numbers from the most credible sources available in 2026, says plainly where they conflict, and translates each one into a decision for whoever owns testing. For a daily stream of sourced numbers, see the statistics section.

How far along are SAP customers?

• DSAG (Feb 2026, 198 respondents): 54% still run ECC somewhere, down from 68% in 2024. S/4HANA deployments split 56% on-premise, 17% private cloud and 5% public cloud. 37% plan to finish switching by the end of 2027, 48% by 2030 and 4% by 2033^[1].
• ASUG and Precisely (Nov 2025): 59% of respondents are fully or partly live on S/4HANA, up from 46% in 2024^[2].
• Gartner (2025): of roughly 35,000 ECC customers, about 39% had moved by the end of 2024, with around 17,000 still expected on ECC in 2027 and about 13,000 in 2030^[3].

What it means for testing: a large share of the market will be testing a migration in the same 2026 to 2030 window, all competing for the same experienced SAP testers and partners. Lock in people and tools early, and build capability you keep rather than rent.

How long, and how well, do migrations go?

• Horváth (Q1 2025, 200 companies): S/4HANA projects ran 30% longer than planned on average, and only 8% finished on time. More than 60% deviated on budget, schedule or quality, and 65% reported severe quality deficiencies. Underestimated testing and data migration were among the causes^[4].
• Gartner: migration costs for large enterprises range from about $2 million to $1 billion, over three to seven years^[3].

What it means for testing: the quality number is the one to take to your steering committee. When two-thirds of companies report severe quality problems, the testing budget is not where you economize. Put data migration reconciliation and end-to-end process testing in the plan from the first mock, and treat schedule contingency as a planning input, not an admission of failure.

What gets in the way?

In the ASUG and Precisely research, the top barriers were business process change (49%) and customizations (44%)^[2]. Both are testing problems in disguise: every changed process needs new scenarios and acceptance criteria, and every customization is code that must be found, remediated and regression tested.

What it means for testing: start scenario design during Fit-to-Standard, not after it, and use custom code analysis to decide what goes into every regression cycle (see the migration test plan).

What SAP’s own results say

SAP does not publish the share of ECC customers that have migrated. What it does publish shows the direction of travel: current cloud backlog of €22.9 billion in Q2 2026, up 27%^[5]. More of the installed base is committing to cloud contracts, which (as the RISE guide explains) means a fixed upgrade rhythm and recurring regression testing after go-live.

A number nobody has: testing’s share of effort

You will see claims that testing consumes 30%, 40% or more than 60% of an S/4HANA program. We could not find an independent source for any of them; the ones we traced came from test tool vendors. The defensible framing is the Horváth finding that testing phases are routinely underestimated^[4]. Better still, estimate your own: the free test estimation model builds the number from your process count and complexity.