Four analytics products on four different metrics, and the fragmentation cost 15 to 30 percent in duplicated or idle spend
Nobody buys an analytics estate. They buy four products across four budget cycles on four metrics, and the bill is the sum of decisions nobody made together.
Prepared by Redress Compliance · August 17, 2026 · SAP advisory. 20 to 30 SAP data and analytics reviews benchmarked, 2024 to 2025.
Executive summary
Fragmented buying created 15 to 30 percent of duplicated or idle spend across the analytics estate. Four products bought separately, on four metrics, through four budget conversations that never met.
Datasphere and BW carried overlapping capability paid for twice in 15 to 30 percent of estates. Modelling and warehousing capability overlap, and both carry cost for as long as both are running.
Analytics Cloud planning seats sat unused while viewers held the expensive licence. SAC bills by user type, so an assignment error is a pricing error, and it is invisible on a seat count.
BTP credits funding analytics services expired idle because nobody forecast the burn. Consumption metrics require a forecast. A credit balance with no forecast behind it is a deposit, not a budget.
Four products, four metrics
The estate is not one purchase and it does not behave like one. Each product bills on a different basis, which is why nothing reconciles across them.
| Product | Metric | Where the waste appears |
|---|---|---|
| SAP Analytics Cloud | User type | Planning seats idle while viewers hold expensive licences |
| SAP Datasphere | Consumption | Overlaps the BW estate, both paid for |
| BTP credits | Consumption | Credits expire idle with no burn forecast |
| BW | Deployment and engine metrics | Components retained after Datasphere covers them |
Read the metric column rather than the product column, because that is where the governance problem sits. A user type metric is managed by looking at people. A consumption metric is managed by forecasting burn. A deployment metric is managed by inventorying what is installed. Those are three unrelated disciplines, usually owned by three different groups, and no single review naturally covers all of them. The 15 to 30 percent is not the result of any one bad decision. It is the space between four decisions that were each defensible on their own.
The overlap between Datasphere and BW is the expensive one
Across roughly 20 to 30 SAP data and analytics reviews benchmarked between 2024 and 2025, fragmented buying created 15 to 30 percent of duplicated or idle spend across the analytics estate. SAP data and analytics spans Analytics Cloud, Datasphere, BW, and BTP credits, and each bills on a different metric: SAC by user type, Datasphere and BTP services by consumption, and BW tied to deployment and engine metrics. Four products bought at four different times, through four budget conversations, measured four different ways.
The most expensive single pattern sits between SAP Datasphere and the legacy BW estate, where modelling and warehousing capability overlap and both carry cost. That duplication was present in 15 to 30 percent of estates reviewed. It is rarely a mistake at the point of purchase, because Datasphere is usually bought as the destination while BW is still the system of record, and paying for both during a transition is a reasonable interim position. What turns a reasonable interim into a permanent overlap is the absence of a migration plan with dates. Parallel adoption, where both run indefinitely because no one committed to retiring anything, is what produces the double bill.
The other two patterns are smaller individually and equally avoidable. Analytics Cloud bills by user type, so planning seats sitting unused while viewers hold the expensive licence is a pricing error dressed as a seat count, and it is invisible unless someone compares assigned type against actual activity. BTP credits funding analytics services expired idle because nobody forecast the burn, which is the standard failure mode of a consumption metric bought by a team accustomed to seats: a credit balance with no forecast behind it is a deposit rather than a budget.
The remedy is a single inventory rather than four. Inventory which workloads run on BW versus Datasphere, then retire the BW components that Datasphere now covers, on a plan with dates attached. Compare SAC assigned user types against actual activity. Forecast BTP burn before the credits are bought rather than after they lapse. And govern the four as one estate at renewal, because the 15 to 30 percent lives in the space between them and no product level review will find it. The wider RISE bundle economics sit in the RISE pillar, the process suite in the Signavio brief, and the library in the SAP practice.
- 520 vendor benchmarks, from SAP RISE to Oracle ULA to Microsoft EA
- Overlapping capability identified across products bought in separate cycles
- Every risky clause flagged with the exact quote, the page, and the replacement language
Consolidating the estate
- Inventory which workloads run on BW versus Datasphere, which is the only way to see the overlap that 15 to 30 percent of estates are paying for twice.
- Retire the BW components that Datasphere now covers, on a migration plan with dates rather than through parallel adoption.
- Compare SAC assigned user types against actual activity, since planning seats sitting idle while viewers hold expensive licences is a pricing error, not a seat count issue.
- Forecast BTP credit burn before buying, because a consumption metric without a forecast behind it produces a deposit rather than a budget.
- Govern the four products as one estate at renewal, as the waste lives between them and no product level review will surface it.
- Match the discipline to the metric, people for user types, burn forecasting for consumption, installed inventory for deployment, rather than applying one review to all four.
What the analytics reviews showed, 2024 to 2025
Across roughly 20 to 30 SAP data and analytics reviews benchmarked:
Share of analytics spend created by fragmented buying across four products purchased separately on four different metrics.
Estates carrying overlapping modelling and warehousing capability, paid for twice, usually because no migration plan had dates attached.
Analytics Cloud planning seats sat unused while viewers held the expensive licence. BTP credits funding analytics services expired idle because nobody forecast the burn.
SAC bills by user type, Datasphere and BTP services bill by consumption, and BW is tied to deployment and engine metrics. A migration plan, rather than parallel adoption, is what stops Datasphere and BW being paid for at the same time.
Watch the briefing · 3:455 Tips for Your SAP NegotiationWhy an estate bought in four conversations gets governed in none of them.
Your first five moves
- Build one inventory across all four products, because the waste lives between them rather than inside any one.
- Map every workload to BW or Datasphere and identify what both are covering.
- Put dates on the BW retirement plan, which is what separates a transition from a permanent double bill.
- Reconcile SAC user types against activity, moving viewers off the expensive licence.
- Forecast BTP burn before the next credit purchase. The SAP practice consolidates the estate with you.
Frequently asked questions
How much does analytics fragmentation cost?
15 to 30 percent of the analytics estate in duplicated or idle spend, across the 20 to 30 reviews benchmarked. It is the space between four separate decisions rather than any one bad one.
Which four products are involved?
SAP Analytics Cloud, SAP Datasphere, BW, and BTP credits. SAC bills by user type, Datasphere and BTP services bill by consumption, and BW is tied to deployment and engine metrics.
Where is the largest duplication?
Between Datasphere and the legacy BW estate, where modelling and warehousing capability overlap and both carry cost. It was present in 15 to 30 percent of estates reviewed.
Is paying for both always a mistake?
Not at the point of purchase. Datasphere is usually bought as the destination while BW is still the system of record, and paying for both during a transition is reasonable. It becomes waste when no retirement plan has dates.
What stops the double bill?
A migration plan rather than parallel adoption. Parallel adoption, where both run indefinitely because nobody committed to retiring anything, is precisely what produces the permanent overlap.
What goes wrong in Analytics Cloud?
Planning seats sit unused while viewers hold the expensive licence. Because SAC bills by user type, that is a pricing error rather than a seat count issue, and it is invisible without comparing assigned type to activity.
Why do BTP credits expire idle?
Because nobody forecast the burn. A consumption metric bought by a team accustomed to seats produces a credit balance with no forecast behind it, which is a deposit rather than a budget.
Why does no single review find this?
Because the four metrics require three unrelated disciplines: people for user types, burn forecasting for consumption, installed inventory for deployment. They are usually owned by different groups, and no standard review covers all of them.
What should we inventory first?
Which workloads run on BW versus Datasphere. That single comparison surfaces the largest duplication and it is the prerequisite for a retirement plan with real dates.
How should the estate be governed at renewal?
As one estate, not four products. The waste sits between them, so a product level review will confirm that each purchase was defensible while missing the 15 to 30 percent entirely.
