Something I’ve picked up from conversations with CPG teams is that figuring out why a forecast is off track usually isn’t that complicated.

The hard part is getting to the answer quickly.

Most teams already have the data they need. They have shipments, depletions, inventory, retail sales, promotions, pricing and forecasts. But those signals often live across different systems, spreadsheets and reports.

So when a forecast starts slipping, there are dozens of places someone could look.

A better approach is to start with a simple decision tree.

1. Are we in stock where we’re supposed to be?

Before digging into demand, start with availability.

If a product is authorized in 1,000 stores but is only available in 750, that changes how you should interpret everything that comes after it.

A velocity problem might actually be a distribution problem.

A weak promotion might actually be an availability problem.

A forecast miss might have very little to do with consumer demand at all.

The first question should be simple: Was the product actually available to sell?

2. If we're in stock, is velocity changing?

Once availability looks healthy, then look at how quickly the product is selling.

Is velocity below plan? Is the decline happening everywhere, or is it isolated to a retailer, region or SKU?

This is where retail data starts adding context to what you’re seeing in shipments and depletions.

If distribution is stable but velocity is slowing, you’ve narrowed the problem considerably.

3. What changed around the product?

If velocity has moved, the next question is why.

Was there a promotion that ended? Did pricing change? Is the category slowing? Did a competitor go on promotion?

Rather than looking at every possible metric at once, each answer should determine the next question.

That’s the important part.

4. If retail looks healthy, what’s happening upstream?

Sometimes consumer demand looks perfectly healthy while shipments or depletions are falling behind.

That’s when it makes sense to look upstream.

Is distributor inventory building? Are shipments keeping pace? Is there enough inventory to support the expected demand? Is there a timing issue somewhere between the supplier, distributor and retailer?

Now you’re investigating a supply or ordering problem rather than a demand problem.

This is where AI gets interesting for CPG

Most reporting today is still reactive.

Someone sees that the forecast is off, opens a dashboard, pulls a report and starts working through the questions above.

But once shipment, depletion, inventory, retail and forecast data are structured and connected, AI doesn’t need to wait for someone to start looking.

It can continuously listen to those signals and work through the decision tree as things change.

Forecast slipping → check availability → availability healthy → check velocity → velocity down → check promo and pricing.

That starts to look less like traditional reporting and more like agentic reporting.

Instead of giving teams another dashboard to monitor, the system can surface what changed, why it matters and where someone should look next.

That’s the direction we’re building toward at Shopra: not more CPG data, but a faster path from “we’re off plan” to “here’s why.”

‍