SAP Data Accelerators
SAP is often the hardest source to extract cleanly. Pre-built extraction patterns and modelled marts remove the groundwork every SAP analytics project repeats.
The Difficulty Is Never the Dashboard
A finance director asks for a margin view by plant and product line. On clean data that is a short build. On SAP it often becomes a long project, because:
- The data lives across BKPF/BSEG, ACDOCA, VBAK/VBAP, MARA/MARC and configuration tables, with meaning that depends on company code and ledger.
- Much of the descriptive text sits in language-keyed text tables, and much of the reporting structure in cost-centre and profit-centre hierarchies that a relational join will not flatten for you.
- The extraction route decision (direct SQL, classic extractors, ODP/OData, CDS views or SLT) carries licensing, performance and supportability consequences.
- Deltas are not simple: reversals, parked documents and back-dated postings all mean “changed” in ways a naive timestamp watermark can miss.
- The basis team has other priorities, and every open question costs calendar time.
SAP ECC or S/4HANA
Finance and operational data, with BW extractors where present.
Extraction route
The route that fits your release, licensing position and basis-team constraints.
Modelled marts
Target data models for FI, SD, MM, PP and PM, populated and reconciled.
Your platform
Snowflake, Databricks, Microsoft Fabric or a Qlik-native model.
Extraction Patterns and Modelled Marts
SQL / Direct
Qlik SAP SQL connector against transparent tables, with pooled and cluster-table handling where it applies.
Classic Extractors
Reuse of existing BW extractors where they are already set up and trusted.
ODP / OData
Operational data provisioning for delta-enabled extraction on modern releases.
CDS Views
S/4HANA CDS-based extraction, including delta-enabled views where exposed.
SLT / CDC
Real-time or near-real-time replication via SLT, or Qlik Replicate where the pattern fits.
Templates, Not One Route for All
Parameterised extraction templates per route, with delta strategy and restart behaviour decided and documented. Which route fits depends on your estate.
Five Marts, Plus the Work Every SAP Mart Needs
Each mart is a target data model, the transformation logic to populate it, and a Qlik Sense app template, deployable to Snowflake, Databricks, Microsoft Fabric or a Qlik-native model.
FI: Finance
- General ledger
- AR and AP ageing
- Document flow
- Ledger and currency handling
- Cost-centre and profit-centre hierarchies
SD: Sales
- Order-to-invoice flow
- Order book and backlog
- Pricing conditions
- Rejection and credit-block analysis
- Customer hierarchy
MM: Materials
- Inventory position and valuation
- Movement history
- Purchase order and requisition cycle
- Vendor performance
- Slow-moving stock
PP: Production
- Production orders
- Confirmations
- Yield and scrap
- Routing and work-centre load
- Order-to-completion cycle time
PM: Plant Maintenance
- Notifications
- Work orders
- Downtime
- MTBF and MTTR
- Maintenance cost by functional location
Included in Every Mart
- Language-keyed text resolution
- Unit and currency conversion against SAP configuration
- Fiscal-year variant handling
- Hierarchy flattening: cost centre, profit centre, product and customer
- A delta strategy that handles reversals and back-dated postings
From Source Assessment to Handover
Source Assessment
Release, modules in use, licensing position, existing BW landscape, available extraction routes and basis-team constraints. Output: a route recommendation with the trade-offs written down.
Mart Selection and Gap Analysis
Pick the marts that map to the business questions, then list where your configuration differs: custom fields, Z tables, unusual fiscal variants. That gap list is the actual project scope.
Extraction Build
Templates deployed and parameterised, initial loads sized and run, and deltas proven against the document-change behaviour in your system.
Model and Validate
Marts populated and reconciled against SAP standard reports. Differences are investigated and explained, not smoothed over.
Consume
Qlik Sense app templates deployed, or the semantic layer published in Snowflake, Databricks or Microsoft Fabric for whatever sits on top.
Hand Over
Under Qlik Replicate + Talend CI/CD pipelines, with documented models and a training session for the team that inherits it.
What It Covers
SAP Data Accelerators at a Glance
- Sources
- SAP ECC, S/4HANA; BW extractors where present
- Routes
- SQL, classic extractors, ODP/OData, CDS, SLT, Replicate CDC
- Marts
- FI, SD, MM, PP, PM
- Targets
- Snowflake, Databricks, Microsoft Fabric, Qlik-native
- Best fit
- ECC or S/4HANA estates feeding Qlik, Snowflake or Databricks
- Prerequisites
- A supported SAP extraction route and basis-team cooperation
- Indicative scope
- A first governed mart is typically approached within 4–8 weeks, depending on the estate (indicative, not guaranteed)
Questions We Get Asked
Both, including mixed landscapes during an ECC-to-S/4HANA conversion. The extraction layer differs (ACDOCA and CDS views change the finance picture considerably on S/4HANA), but the mart models are designed so the reporting layer can carry across your own migration rather than being rebuilt with it.
Custom fields extend the model rather than break it; that is the normal case and it is handled in the gap analysis. What genuinely adds effort is custom process logic implemented in Z tables and user exits, because the business meaning has to be re-derived, so that is scoped explicitly rather than absorbed into an estimate.
No. The extraction and the modelled marts are the substance, and the target can be Snowflake, Databricks or Microsoft Fabric. The Qlik Sense app templates are useful if you run Qlik; the marts can equally be landed for another BI tool to consume.
It is a real question, and it depends on your own licence position. We set out which extraction routes carry which considerations; the contractual interpretation is between you and SAP.
Initial builds usually run against a non-production system, and high-volume initial loads are planned with the basis team. Where read impact is the concern, SLT or Replicate-based CDC can move the load off the application layer where the pattern fits.
What Governed SAP Data Makes Possible
Once the marts exist and reconcile, further work becomes possible. These are potential next stages built on the marts, not part of the accelerator itself.
Forecasting
Demand and revenue forecasting on SD order history, with production capacity from PP as a constraint.
Anomaly Detection
Unusual posting patterns, duplicate vendor invoices and margin outliers.
Retrieval & Agents
Natural-language questions answered against the governed marts, with company-code security respected at query time.
Delivered Under the Same Release Control
The extraction and mart pipelines are handed over under version control and pipeline promotion, so the SAP layer is rebuildable rather than held in one environment.
Qlik Replicate + Talend CI/CD →Stay Ahead with AI Insights
Insights on AI, data analytics, and modern data platforms.
Ask for the Model Walkthrough
We show you the FI mart structure and the hierarchy handling on screen.
Don't Miss Out on Our Monthly Newsletter
AI, BI and data platform insights. No spam.