Cross-Selling in CPG: Connecting Marketing, Supply, and Sales Data

Cross-selling is one of the most consistently under-executed commercial strategies in CPG. Not because the opportunity is not there, it almost always is. But because acting on it requires connecting data that sits in separate systems, owned by separate teams, organized around separate objectives.

The sales team has transaction history: which retail accounts buy which product lines, at what volume, and through which channels. Marketing has campaign data: upcoming activations, promotional calendars, and brand priorities. Supply chain has inventory and fulfillment data: which SKUs are constrained, which have surplus, and which markets have headroom. Put those three together and cross-sell opportunities become visible. Keep them separate and the commercial team is making decisions from one-third of the picture.

The problem is not missing data. Every large CPG organization has invested in CRM, ERP, demand planning, and marketing platforms. The problem is missing connections — a shared view of accounts, products, channels, regions, and campaigns that lets commercial teams ask the questions that actually drive revenue.

THE COMMERCIAL COST

What disconnected commercial data actually costs

Whitespace you can’t see

Whitespace — the share of wallet within an account that you don’t currently have — is one of the highest-ROI targets in CPG commercial strategy. But identifying it requires connecting purchase history against the full product range, segmented by account, region, and channel. When sales data is in the CRM, product data is in the ERP, and the analysis requires a manual extract from both, whitespace analysis becomes a quarterly exercise rather than a daily sales tool. By the time the commercial team acts on the analysis, the competitive window may have closed.

Promotions that don’t land

Marketing invests in a trade promotion for a product line that has strong brand equity but low penetration in the Southeast. Supply chain has not been told. The promotion runs. Demand spikes. The product is out of stock in three of the six targeted retail partners by day four. The sales team is fielding calls from angry buyers. The promotion ROI is half of plan, because half the incremental demand was lost to out-of-stock rather than converted to revenue.

This is not a marketing failure or a supply chain failure. It is a coordination failure, one that stems directly from the absence of a connected view across promotion, inventory, and account data.

Sales conversations without commercial context

A key account manager walking into a retailer review with last quarter’s sell-in data but no visibility into that retailer’s basket composition, competitive threat exposure, or upcoming promotional calendar is operating with one hand tied behind their back. The account knows more about the interaction than the manufacturer does. Connected commercial intelligence changes that dynamic: the KAM walks in knowing which product lines are underperforming at shelf, which promotional activations are planned, which SKUs have supply that needs pulling through, and what the retailer’s buying pattern suggests about upcoming orders.

THE SCENARIO

What connected commercial intelligence changes: A cross-sell example

With disconnected data, preparing for that meeting requires pulling reports from three systems, reconciling account-level sell-through data against the category plan, and making assumptions about inventory and promotional support based on what each team can share before the meeting. It takes a week of preparation and still produces an incomplete picture.

With connected commercial data, the KAM can ask the relevant questions directly:

The KAM walks into the review knowing which gaps to close, which promotional mechanics to propose, and which supply constraints to flag. The conversation shifts from presenting last quarter’s performance to planning next quarter’s growth together.

THE CONNECTIONS

What needs to be connected for CPG Cross-sell Intelligence

The entities that drive cross-sell opportunities in CPG are well understood by every commercial leader. The challenge is that they each live in a different system with a different schema and a different owner.

When these relationships are declared in a shared connected model built directly within the Databricks Lakehouse under Unity Catalog governance, the commercial team can work from a unified view of accounts, products, promotions, and supply without waiting for a data team to assemble a bespoke report for each question.

WHAT CHANGES

Three commercial capabilities that connected data enables

1. Whitespace visibility at account level

When purchase history is connected to the full product catalogue, whitespace is not a quarterly analysis exercise, it is a live view. Commercial teams can see, for any account, which product lines are absent from the purchase record, which categories are underpenetrated relative to comparable accounts, and which adjacent categories represent the most credible cross-sell entry point. That visibility, updated continuously as orders and sell-through data flows in, turns whitespace from a strategic aspiration into a daily commercial tool.

2. Promotion planning that closes the loop with supply and sales

Trade promotion ROI in CPG is notoriously difficult to measure precisely, in part because the data required — promotional spend, incremental sell-through, baseline sales, and competitive context — is distributed across systems that do not naturally connect. When promotion data is connected to account-level sell-through, inventory availability, and demand forecast, the loop closes.

Promotional planning improves because the team can assess supply availability before committing to a promotional volume. Promotional ROI improves because the analysis is available in real time during the campaign, not weeks after it ends. And coordination between marketing and supply chain improves because both are working from the same connected picture of what the promotion is expected to do.

3. AI-assisted commercial Q&A through Databricks Genie

For commercial teams where Databricks Genie is already the interface for data access, connected commercial context directly improves the quality of AI-assisted answers. As Genie deployments scale across sales, marketing, and category management teams, maintaining consistent definitions of accounts, products, promotions, and commercial performance becomes increasingly important. A shared connected model ensures that the KAM in the Midwest asking Genie about an account’s purchase history, and the category manager in HQ asking about promotional performance in the same region, are working from the same definitions.

The result is commercial intelligence that sales teams actually use without SQL, consistent across the organization, and grounded in the connected data that reflects how the business actually operates.

THE JOURNEY

From connected data to commercial intelligence: The maturity path

Databricks + Kobai: Connected commercial intelligence for CPG

Kobai extends the Databricks Lakehouse with the connected commercial context that CPG cross-sell strategies require. Graph structures are built directly within Databricks under Unity Catalog governance, connecting account, product, promotion, inventory, and sales data into a shared operational model. Commercial teams — sales, marketing, and category management — access the connected picture through Databricks Genie or Kobai’s own interfaces, without depending on the data team for every commercial question.

The pattern mirrors the inventory optimization work described in our earlier CPG post: the problem is not missing data, it is missing connections. The same Databricks Lakehouse that holds your CRM, ERP, and marketing data already contains everything a commercial team needs to identify cross-sell opportunities, plan promotions with supply confidence, and prepare for account reviews with the full commercial picture. Kobai surfaces those connections.

The path to getting started is through the Databricks Marketplace: the Semantic Graph Pilot provides a structured 2–4 week engagement to connect a defined set of commercial data domains and demonstrate the cross-sell intelligence the connected model produces on real data.