Marketing on Real Revenue: Feeding SAP Engagement Cloud (Emarsys) With Sales Data From S/4HANA Public Cloud

Marketing segments on clicks. Newsletter opened, link clicked, form filled: that’s what feeds the target groups, the scoring models, entire campaign plans. Meanwhile the actual gold sits two systems away, in the ERP. Orders, invoices, product mix, ordering rhythm: the ERP knows to the cent who buys what, when, and how often. The marketing automation knows who opened an email on Tuesday. Which of the two would you plan your budget on?

With SAP Engagement Cloud (formerly SAP Emarsys, renamed by SAP in February 2026) and S/4HANA Public Cloud, you can close that gap with tools that are already part of both products. The effort is modest; the effect on campaign quality is not. The component names below still carry the Emarsys label, because SAP kept it for the Emarsys edition.

Sales data in Smart Insight: the pattern

The Emarsys edition of Engagement Cloud ships with Smart Insight, an analytics layer built for exactly this: turning sales data into segments. It comes from the e-commerce world, but it does not care where revenue comes from. A sales order from S/4HANA is simply a transaction as far as Smart Insight is concerned.

The path there is unspectacular. S/4HANA Public Cloud exposes sales orders and billing documents through released APIs such as Sales Order (A2X). SAP Integration Suite reads them, maps them and hands them to the sales data upload of Engagement Cloud as a file. One record per line item: customer, product, quantity, value, date. That is all Smart Insight needs, and SAP asks for a daily upload anyway, which is fresh enough (more on that below).

From then on, marketing segments on actual buying behaviour:

• RFM: recency, frequency, monetary. Who bought last, how often, for how much? Smart Insight calculates the scores automatically and keeps them current.
• First-time versus repeat buyers: two completely different audiences that usually get the same email in click-based setups.
• Category buyers: who buys which product group, who has never touched a category at all.
• Churn risk: no order in X months, measured against that customer’s own usual rhythm, not against one fixed threshold for everyone.

That last point is the important one. A customer who orders every two weeks and has gone quiet for six is an alarm. A customer on a yearly rhythm is, after six weeks, simply normal. Without ERP data, marketing can’t see the difference.

Four B2B programmes that work in real projects

From our projects, anonymised and deliberately unspectacular, which is why they work:

1. Replenishment reminders based on real order cycles. Consumables, spare parts, wear items: Smart Insight knows each customer’s rhythm per product. Shortly before the next order would be due, a reminder goes out. Nobody has to guess, because the cycle is already written in the order data.
2. Cross-sell on the first category purchase. When an account buys from a product category for the first time, a short programme kicks off with accessories, consumables and application know-how for exactly that category. Not the full catalogue, just what fits.
3. Win-back on dropping order frequency. When an account’s frequency falls well below its own historical average, marketing gets active before field sales discovers the revenue drop in the quarterly report. That is the difference between a call at the right moment and a missed opportunity.
4. New-product announcements only to buyers of the predecessor. The successor generation of a product interests, above all, the people running the predecessor. One segment on the buyers of the old material numbers is enough. Higher relevance, less waste, and nobody markets products to customers who never bought them.

None of this needs a data science team. All of it needs sales data in the marketing system. Where a customer also runs Sales Cloud V2, the same account and contact model carries through, as described in our post on the integrated S/4HANA Public Cloud and Sales Cloud V2 stack.

One ID strategy, one consent system

For this to hold, two questions need clean answers before the first campaign runs.

Identity: in B2B, the account buys, but a human reads the email. The key is the contact’s email address, connected through the CRM sync: the contact in the CRM hangs on the account, the account hangs on the ERP customer. That way the account’s revenue reaches the right person in Engagement Cloud. One ID strategy, applied consistently, saves months of data cleansing. Two half ID strategies produce duplicates that never go away.

Consent: consent lives in exactly one system, the one where it’s collected, and gets distributed from there. And under both the Swiss Federal Act on Data Protection (FADP) and the GDPR, data minimisation applies: Engagement Cloud needs IDs, product categories and aggregates, not complete invoices. Conditions, payment terms and pricing details have no business being in the marketing system. As a side effect, that makes the conversation with your data protection officer considerably shorter.

Start small, measure properly

Two pieces of advice from project experience, both uncomfortable.

First: three lifecycle programmes, not twenty. A welcome journey for first-time buyers, replenishment reminders, win-back. Build those three properly, let them run for a quarter, then extend. Whoever starts with twenty programmes maintains none of them well.

Second: success is measured against a holdout group, not against open rates. Part of the audience deliberately doesn’t get the programme; the revenue difference between the two groups is the programme’s value. Open rates flatter, incremental revenue proves. And since the sales data flows in daily from S/4HANA anyway, the measurement basis is already sitting there.

Frequently asked questions

Do we need a customer data platform between the ERP and Engagement Cloud?

Not necessarily. For segments on sales data, Smart Insight already covers the commerce-style analytics: RFM, categories, purchase cycles. A customer data platform becomes interesting when many additional sources need consolidating. It is not a prerequisite for this pattern.

How fresh does the sales data need to be?

A daily batch is enough. Lifecycle programmes react to patterns over weeks and months, not minutes. Whether the replenishment reminder fires at 8 am or 2 pm changes nothing about the outcome, so real-time integration here is effort without return.

Does this even work in B2B?

It works well. Revenue hangs on the account, marketing runs at contact level: through the account-to-contact link from the CRM, every contact person gets campaigns based on what their company actually buys. Especially in B2B, where order cycles are stable and product ranges clearly structured, these segments bite better than in many a B2C shop.

What about orders that don’t come through the webshop?

They are the main reason to do this. Phone, email, EDI, field sales: in B2B, most revenue happens away from the shop. Because the data comes from the ERP and not from web tracking, Engagement Cloud sees the entire revenue, no matter how the order was placed.

---

Your sales figures live in the ERP and your marketing segments on clicks? We will walk you through the pattern with your own order data, and if you are still choosing a partner, our SAP Engagement Cloud partner guide and the Cloud ERP and CRM partner guide explain what to look for. Talk to us.