BDC in Practice: From Raw S/4 and CX Data to Decisions

The start of this series explained the theory; this post is the practice. It describes how we plan a first Business Data Cloud implementation for a Swiss mid-market company on the integrated SAP stack: the timeline, the friction points, and the moment it starts to pay off.

TL;DR: A first BDC project should cover one domain and one question that management already asks. Our plan runs six to eight weeks: activation and a join-key workshop, one semantic model (usually pipeline-to-cash), one analytic app, then a management review that works from a single number. Activation is quick and modelling is fast when data products carry the semantics; aligning keys and owners is the real work, so budget for workshops rather than platform.

This is our recommended plan, not a report on a specific client. The durations assume clean enough join keys; where they are not, we say so in week one and adjust.

The starting point

The typical company we have in mind has 150 to 400 employees, runs S/4HANA Public Cloud or is on the way there, has Sales Cloud V2 live, and holds its analytics together with exports. On Monday morning, someone in sales operations downloads opportunities to a spreadsheet; someone in controlling downloads order intake; the two files meet in a pivot table with hand-maintained lookup columns.

It works, which is why it survives. It also costs people-days every week and a recurring argument in every management meeting about whose number is right.

The plan at a glance

Phase	Weeks	Output	Who is involved
Activation and ownership	1 and 2	Tenant live, data products subscribed, join keys and owners agreed	IT, sales operations, controlling
Semantic model	3 and 4	One governed model for one domain	Controlling, Spadoom
App and review	5 and 6	One analytic app, first management review	Sponsor, management
Buffer	7 and 8	Key clean-up at the source, if week one found history	Data owners

Weeks 1 and 2: activation and the ownership workshop

Activating the tenant and subscribing to data products is the quick part. SAP describes a data product as a refined data set, the metadata describing it, and an access port or API (SAP News). SAP’s launch announcement named finance, spend and supply chain data from S/4HANA among the first SAP-managed data products (SAP News). For Sales Cloud V2, check the catalogue for your release; where the entity you need is not yet shipped as a data product, replicate it via Datasphere and govern it as your own data product (SAP Help).

The real work of week one is a workshop no tool replaces: agreeing the join keys and who owns them. Account to business partner, opportunity to sales order, material to product. Where the mapping has history (duplicate accounts, opportunities never linked to orders), we either fix it at the source or map it explicitly. This is where the foundation sequence from earlier in this series pays off, or where it has to be caught up.

Weeks 3 and 4: the semantic model

One domain, modelled end to end. For most companies that is pipeline-to-cash: won opportunities joined to order intake and invoiced revenue, by sales organisation, product group and period. The integration patterns post covers the mechanics.

Modelling moves quickly where data products carry the semantics. The effort goes into the company’s last mile: its KPI definitions, fiscal calendar quirks and extension fields. Keep it deliberately narrow. The temptation to “also quickly add” service data or marketing attribution in week three is how a six-week project becomes a six-month one. That is domain two. For a finance-led alternative first domain, starting with service margin follows the same discipline.

Weeks 5 and 6: the app and the first review

The output is one analytic app (an SAP Analytics Cloud story or a small custom app) answering the questions the sponsor named in week one: where is the pipeline against forecast, how accurate was last quarter’s forecast against invoiced revenue, which deals moved. It goes to the people who asked, with a short walkthrough rather than a training programme.

SAP also offers packaged intelligent applications on BDC. As of October 2025, Finance Intelligence, People Intelligence and Cloud ERP Intelligence Private were generally available, and Revenue Intelligence was in restricted preview (SAP News). Check what is available for your release before building something SAP already ships.

The goal of the first management review is simple: sales and finance look at the same pipeline-to-cash view, and the time normally spent reconciling numbers goes to the decisions behind them.

What changes afterwards

Three effects are what we design for:

• Forecast discipline improves without a new policy. When forecast accuracy is measured against invoiced revenue and shown monthly, sales reps keep opportunities current because the number is visible.
• Domain two is cheaper than domain one. Keys are aligned, ownership is decided and the platform stands. Service profitability (Service Cloud V2 cases against contract margins) is a natural next step.
• The AI conversation becomes concrete. With a governed model underneath, a Joule agent answering “how is Q4 tracking against forecast” reads the same numbers the CFO does. The Joule hands-on guide shows what that looks like in Sales Cloud V2. The final post in this series covers the AI dependency in detail.

Why one team for ERP and CRM helps

Most of the friction in weeks one and two sits at the seam between CRM and ERP: which system owns the customer, how an opportunity becomes an order, where prices live. When the same team configured both sides, those answers are already documented. That is the argument we make in the ERP and CRM partner guide, and it applies to analytics as much as to processes. If you are still planning the ERP move, the S/4HANA readiness assessment is the place to bring the data questions in early.

The pattern to hold onto: activation is quick, modelling is fast, alignment is the work. Budget the workshops, not the platform.

FAQ

How long does a first BDC implementation take?

We plan six to eight weeks for a focused first scope: one business domain, the available standard data products from S/4HANA plus replicated CRM data where needed, and one analytic app. That assumes the join keys between CRM and ERP are largely clean. Unresolved master-data ownership and keys with a long history are what stretch it; the integrated SAP cloud stack, where the keys align by design, is what shortens it.

What team do I need on my side?

Smaller than most expect: one person who owns the business question (usually sales operations or controlling), one who knows the history of the data, and access to whoever administers the SAP tenants. At mid-market scale, BDC removes the need for a dedicated data engineering team for this first domain.

What does it cost to run after go-live?

BDC is licensed on a consumption basis in capacity units, and the details depend on your contract. On your side, model stewardship typically takes a fraction of one full-time role once the first domain stands. Compare that against the extractors, warehouse jobs and spreadsheet reconciliation hours it replaces, not against zero.