Kobai.io | Resources

Why Promotions Create Stock-Outs Even When You Have the Data

Written by Kobai | Jun 5, 2026 8:22:54 AM

Marketing launches the flash sale. The shelves empty out. Supply chain is scrambling. And somewhere in your data platform, every signal that could have predicted it was already sitting there — unconnected.

Marketing launches a promotion. Supply chain isn’t ready. Shelves go empty. Everyone has data, but nobody saw it coming.

It is one of the most common and most costly operational failures in consumer goods and retail. A flash sale drives a demand spike that the regional warehouses were not positioned to serve. By the time the out-of-stock reports surface, the lost sales have already happened, the promotion budget has been spent, and the ROI calculation looks nothing like the plan.

The frustrating part is that this is not a data problem. The promotion calendar existed. The inventory levels were tracked. The demand forecasts were modelled. The warehouse capacity was known. None of it was missing. What was missing was the connection between it all — a shared operational picture that lets marketing, supply chain, and operations see the same reality at the same time.

The problem isn’t missing data. It’s missing connections. Every signal that could have prevented the stock-out was already in the data platform, just not connected to anything that could act on it in time.

THE BUSINESS COST

What disconnected decisions actually cost

The business impact of poor coordination between marketing, supply chain, and inventory is visible across multiple lines of the P&L and it compounds quickly.

Lost sales and promotional waste

When a promotion drives demand that the supply chain cannot fulfil, two things happen simultaneously. Sales are lost — customers who intended to buy walk away or substitute. And promotion spend is wasted — the marketing budget that drove the demand surge generated no revenue because the shelves were empty. In high-velocity promotional events, this combination can erode the ROI of a campaign significantly within hours.

Excess inventory in the wrong location

The inverse of a stock-out is equally costly. When demand is concentrated in one region and inventory is sitting in warehouses elsewhere — because the promotion campaign was not connected to the inventory positioning plan — working capital is tied up in the wrong place. The cost of moving that inventory, or of emergency replenishment shipments at premium freight rates, often exceeds the additional margin the promotion was designed to generate.

Reactive expediting and its hidden costs

The response to a stock-out or an unplanned demand spike is typically expediting: emergency replenishment orders at above-market prices, premium freight to move inventory faster, labour overtime to increase warehouse throughput. These costs are rarely captured in the promotion’s post-event ROI analysis, but they are real and they are largely avoidable when the connection between demand signals and supply chain positioning is established before the event rather than discovered during it.

The compounding cost of disconnected decisions

Stock-outs, wasted promotion spend, misplaced inventory, expediting costs, and degraded customer satisfaction do not occur in isolation. Each feeds the next. A stock-out triggered by a poorly coordinated promotion leads to emergency replenishment, which drives excess inventory once demand normalizes, which leads to markdowns, which erodes margin. The root cause of the chain is the absence of shared operational context across marketing, supply chain, and inventory.

THE STORY

A flash sale that nobody saw coming and what changes when they do

Here is the scenario that plays out across CPG and retail every promotional cycle.

The situation

Marketing has just launched a flash sale for the Northeast market. It is performing well — click-through rates are strong, orders are coming in. Three hours later, the operations team gets the first stock-out alert. Two of the featured SKUs are out of stock at the Northeast regional warehouse. The replenishment order is four days away. The promotion runs for six more days.

The questions that needed answers three days ago

What was the expected demand uplift for these SKUs in this market? Did current inventory levels support the promotion run? Did the warehouse have the capacity to handle an inbound surge? Were the right purchase orders already in flight? Could the launch timing have been adjusted by 48 hours to align with a scheduled replenishment arrival?

Every one of those questions was answerable before the promotion launched. The data existed. What was missing was a connected operational picture — one that made the relationship between the promotion calendar, the inventory levels, the warehouse throughput, and the demand forecast visible to the people who needed it, at the time they needed it.

When that connection exists, the conversation changes. Instead of a post-event stock-out investigation, it becomes a pre-launch coordination review. Operations and supply chain can see the same demand picture that marketing is working from. Replenishment timing can be adjusted. High-risk SKUs can be flagged before the campaign goes live. The promotion still happens with better outcomes.

The flash sale was not the problem. The lack of shared context between marketing, supply chain, and operations was the problem. Shared context turns a reactive scramble into a coordinated response.

WHAT CHANGES

What becomes possible with connected operational context

Faster decisions — fewer meetings

Today, coordinating between marketing, supply chain, and inventory requires assembling people from multiple teams to pull data from multiple systems, reconcile it in a spreadsheet, and agree on what the picture means. That process takes time the promotional calendar does not always allow. When operational context is shared and connected, teams can see the same picture without a coordination call. Supply chain can see the promotion launch in the context of current inventory. Marketing can see the inventory position in the context of the demand plan. Decisions that take days to coordinate take hours.

Better promotion ROI

Promotion ROI is not just a function of the demand the campaign drives. It is a function of how much of that demand the supply chain can actually fulfil and at what cost. When inventory positioning, replenishment timing, and warehouse throughput are coordinated with the promotion plan before launch, more of the driven demand results in completed sales. Less ends up as lost revenue or expedited replenishment at a premium. The same marketing investment delivers a better commercial outcome.

Visible stock-out risk before it becomes a stock-out

The value of connected operational context is highest before an event, not during it. When demand signals, inventory levels, promotion calendars, and replenishment schedules are connected in a shared model, teams can see which SKUs are at risk before a promotion launches and act. Adjusting replenishment timing, reallocating inventory across depots, or flagging a high-risk SKU for a temporary promotion hold are all preferable to managing a live stock-out.

Genie answers grounded in the full operational picture

For organizations running Databricks Genie, connected operational context directly improves the quality and reliability of AI-assisted answers. Genie is excellent at helping users ask questions about their data. The challenge, as Genie deployments scale across marketing, supply chain, and operations teams, is ensuring that each team’s questions are grounded in a shared understanding of products, inventory, warehouses, promotions, suppliers, and demand.

When Kobai extends the Databricks Lakehouse with a shared connected model of those entities and their relationships, Genie’s answers to promotion and inventory questions are grounded in the same operational context that supply chain is using. Marketing can ask “what is the current inventory position for the SKUs in our Northeast campaign?” and get an answer that reflects the same data that operations is working from.

THE CONNECTIONS

What needs to be connected and how

Connecting demand, supply, and promotion data is not primarily a technology problem. It is a modelling problem: making explicit the relationships between business entities that exist in the real world but are not captured in the schema of any individual system.

Entity

Connects to

Business question it unlocks

Promotion / Campaign

SKUs, markets, historical demand lift, spend

What demand is this promotion expected to drive, where, and for which products?

SKU / Product

Inventory positions, demand history, substitutes

Which products are at stock-out risk given current inventory and the expected uplift?

Inventory Position

Warehouse, safety stock level, reorder point

Is inventory positioned in the right location to serve the expected demand?

Warehouse / Depot

Capacity, inbound shipments, labour schedules

Can the warehouse handle the replenishment throughput the promotion requires?

Purchase Order

Supplier, SKU, lead time, expected arrival

Which replenishment orders arrive within the promotion window, and which do not?

Demand Forecast

SKU, market, channel, promotion adjustment

How does actual demand compare to the plan as the promotion runs?

When these relationships are declared in a shared model — built within the Databricks Lakehouse under Unity Catalog governance, authored by supply chain and demand planning teams using no-code tooling — the coordination conversation changes. Teams are not reconciling separate views of the same data. They are working from a shared operational picture.

THE JOURNEY

From coordination support to connected operations

The shift from reactive to connected operations does not happen overnight. Most organizations move through a series of stages, with each stage building on the same connected foundation.

Stage

What it enables

Shared visibility

Marketing, supply chain, and operations see the same connected picture of inventory, demand, and promotion activity. Coordination conversations start from a shared baseline rather than conflicting reports from separate systems.

Pre-launch risk review

Before a promotion launches, teams can review the inventory position for featured SKUs against the expected demand uplift. High-risk SKUs are flagged. Replenishment timing is adjusted. The stock-out risk is visible before the campaign goes live.

Real-time monitoring

As a promotion runs, the connected picture updates continuously. Stock-out risk is visible as demand builds. Operations can adjust inbound priorities and warehouse throughput in response to live demand signals rather than after-the-fact reports.

Agent-assisted coordination

AI agents help surface recommendations — replenishment adjustments, inbound prioritization, at-risk order flags — for human review. Teams retain all decisions; the agent handles information gathering and recommendation drafting.

Databricks + Kobai: connected operational context for CPG and retail

Kobai extends the Databricks Lakehouse with the shared operational context that connects demand, supply, promotion, and inventory into a unified decision-making picture. Graph structures are built directly within Databricks under Unity Catalog governance. Domain experts in supply chain and demand planning author the connections using no-code tooling. Business teams across marketing, operations, and commercial functions access the shared picture through Databricks Genie, without waiting for a data engineering team to reconcile five different reports.

A global CPG manufacturer deployed this approach on their Databricks Lakehouse to address exactly the coordination gap described in this post. Purchase orders, inventory, labour schedules, demand signals, and promotion data were connected into a shared operational model. Operations teams gained real-time visibility across inbound operations. Warehouse throughput improved. Stock-outs and excess inventory both reduced. Business users were accessing insights directly without manual reconciliation.

The shift was from fragmented, reactive coordination to a connected picture that let marketing, supply chain, and operations work from the same operational reality. That is what technology enables. But the outcome is simpler: fewer stock-outs, better promotion ROI, and faster decisions when it matters most.

The problem isn’t the data. It’s the connections. To explore how shared operational context on the Databricks Lakehouse can help your teams coordinate demand, supply, and promotion in real time, visit kobai.io or contact us at contact@kobai.io.