SAP Attribution: What Sales Cloud Records, and What It Never Sees

What follows is each system in turn, the migration that is happening right now underneath all of it, and the two honest ways to close the gap.

Nine out of ten attribution conversations in an SAP organisation stall here, because two people are talking about two different products. The current names, as of September 2026:

• SAP Sales Cloud

  The CRM, previously SAP Cloud for Customer and still called C4C by everybody who has run it for years. Leads, accounts, opportunities, the sales process. Now on version 2, which is a rewrite rather than a release.

• SAP Engagement Cloud

  The marketing engagement platform. It was SAP Emarsys Customer Engagement until 19 February 2026. Journeys, email, personalisation, the engagement data that feeds lead scores.

• SAP Marketing Cloud

  A third product, and the one on the clock: support ends in December 2026, with Engagement Cloud named as the successor. More than 2,000 companies are in that migration now.

• S/4HANA and SAP Analytics Cloud

  Where the invoice and the reporting usually end up. Attribution questions often get pushed here, which is how a marketing question becomes a data warehouse project.

Two fields carry the whole answer, and both are filled in by a human or by a rule somebody wrote.

• Source, in the Lead Information sectionA single value denoting where the lead originated: telephony inquiry, email campaign, trade fair, social media and whatever else your administrator added. One value per lead, chosen at creation.
• The campaign IDAssigned to the lead when it was created as part of a campaign, for identification and reporting. Assigned, not observed: a lead that arrives any other way carries no campaign.

Now look at how leads actually get into Sales Cloud. SAP documents five ways: created manually from the quick create icon or the Sales Campaign work center, imported from Microsoft Excel, uploaded through the Cloud for Customer APIs, scanned from a business card on a mobile device, or copied from an existing lead. Not one of those five is a web session. Whatever the visitor did on your website before becoming a lead is, as far as Sales Cloud is concerned, something that happened to somebody else.

Engagement Cloud is where the marketing interactions live, and it measures them properly: link clicks, email opens, form submissions and other interactions between your company and a contact, rolled up into an engagement score that can drive a lead qualification level. Cross an engagement score of seventy and marketing operations can have the lead marked hot and routed for follow up.

Read that list again and the shape is the same one every marketing automation platform has. Every interaction it counts is one it produced itself. An email it sent and somebody opened. A form it hosts and somebody submitted. A link in its own message. The LinkedIn ad that made the person search your brand three weeks before any of that is not an interaction Engagement Cloud can have a score for, because nothing told it the visit happened.

So the engagement score is a good measure of how warm a known contact is, and a poor measure of which channel bought them. Those are different questions and an enterprise reporting request usually conflates them.

Line the three systems up against a normal B2B journey and the hole is obvious. Somebody clicks a LinkedIn ad, reads two pages and leaves. Two weeks later they search your brand, read a case study and leave again. A week after that they fill in a form.

1. Sales Cloud gets a lead with Source set to whatever the form mapping says, probably Web or Social Media, and a campaign ID if somebody wired one in.
2. Engagement Cloud starts counting at the form submission, because that is its first interaction with a person it can name.
3. Neither holds the LinkedIn click, the two anonymous visits, the brand search or the case study, so the channel that opened the journey is credited to nothing and the channel that happened to be last takes the whole record.
4. And nothing goes back out. When the opportunity closes as won, no conversion travels to LinkedIn, Google, Meta or Microsoft, so the platforms keep optimising toward form fills rather than toward the deals your sales organisation actually signed.

For an enterprise the last point is usually the expensive one. Media budgets at this size are optimised by algorithms, and an algorithm optimises toward the signal it receives. Feed it form fills and it will find you more form fills, including from the segments that never buy.

The first is the route most SAP organisations reach for, because it stays inside the estate and inside the existing vendor relationship. It is also the route that turns a marketing question into a two quarter data project, and it does not solve the underlying problem: the anonymous sessions still have to be recorded by something before any model can use them.

Being direct about what exists today: LeadJourney has no native SAP Sales Cloud connector. HubSpot, Salesforce, Microsoft Dynamics 365, Pipedrive, Attio, Close and GoHighLevel connect natively; every other CRM, SAP included, connects through the webhook and the open API, or through Zapier and Make for the middle layer. That is a real integration and it is how several customers run, but it is a webhook, not a one click connector, and anybody scoping this for an enterprise landscape should know that before the architecture review, not after it.

What LeadJourney brings to that connection is the half SAP does not collect. One script on your site, live in about 21 minutes, recording every visit server-side on your own domain at 95%+ accuracy. At the first click it captures the click ID from each platform, gclid with gbraid and wbraid, fbclid, li_fat_id and msclkid, plus the UTMs and the landing page, and it holds them first-party until the visitor identifies themselves. When the lead is created, that whole anonymous history is joined to it.

From there the loop closes in both directions: spend is pulled from Meta, Google, LinkedIn and Microsoft Ads and joined to the opportunities themselves, so cost per closed opportunity is measured rather than modelled, and when the opportunity closes the amount goes back to the platform that produced the first click, matched on that click ID. Every attribution model, first click, last click, linear, time decay and position based, runs on the same tracked events.

Two limits worth naming for an enterprise reader. Tracking is first-party and respects consent decisions rather than working around them, so a visitor who declines is not measured. And we hold no SOC 2 Type 2 report today, which for some procurement processes is the end of the conversation; the data is processed and stored in the EU under a data processing agreement, and that is what we can put in writing.

Keep going with what SAP campaign management actually tracks, and the same analysis for the CRMs that connect natively: Microsoft Dynamics 365, Salesforce and HubSpot. For the wider picture, B2B marketing attribution and the multi-touch attribution models behind it.