POS-Based Demand Forecasting: Why CPG Companies Must Flip from Shipments to Sell-Through
๐ Most CPG demand planning models may not be forecasting the right signal.
We build sophisticated statistical and ML models on outbound shipments – what left our factory or DC, and call it demand. It isn’t.
Demand is what the trade ordered, distorted by promotional loading, quarter-end pushes, retailer inventory strategies, and truckload economics.
The actual demand signal, the one that drives your business is happening at the shelf and on the retailer’s app.
๐ POS. Sell-through.
The flip is this: ๐ forecast POS first, then translate POS back into shipments; yes a bit more nuanced but gets you to the promised land!
To make this happen:
1๏ธโฃ ๐ ๐ผ๐ฑ๐ฒ๐น ๐ฃ๐ข๐ฆ ๐ฎ๐ ๐๐ต๐ฒ ๐ฏ๐ฎ๐๐ฒ ๐ฑ๐ฒ๐บ๐ฎ๐ป๐ฑ ๐๐ถ๐ด๐ป๐ฎ๐น – Syndicated data (Nielsen, Circana) or direct retailer feeds (Walmart Luminate, Kroger Stratum, Target POL, Amazon Vendor Central) become the primary input.
2๏ธโฃ ๐๐ฎ๐๐ฒ๐ฟ ๐ฐ๐ฎ๐๐๐ฎ๐น ๐ฑ๐ฟ๐ถ๐๐ฒ๐ฟ๐ ๐๐ต๐ฒ๐ฟ๐ฒ ๐๐ต๐ฒ๐ ๐ฎ๐ฐ๐๐๐ฎ๐น๐น๐ ๐น๐ถ๐๐ฒ – ๐ฐ Price, promo, distribution, weather, competitive activity โ these move consumer purchases, not truck departures. Modeling them against POS gives you real elasticity coefficients instead of coefficients contaminated by trade behavior.
3๏ธโฃ ๐ง๐ฟ๐ฎ๐ป๐๐น๐ฎ๐๐ฒ ๐ฃ๐ข๐ฆ ๐๐ผ ๐๐ต๐ถ๐ฝ๐บ๐ฒ๐ป๐๐ ๐๐ต๐ฟ๐ผ๐๐ด๐ต ๐ฎ ๐ฐ๐ต๐ฎ๐ป๐ป๐ฒ๐น ๐ถ๐ป๐๐ฒ๐ป๐๐ผ๐ฟ๐ ๐ฏ๐ฟ๐ถ๐ฑ๐ด๐ฒ – ๐ฆ Retailer on hand + weeks of supply targets + pipeline for new items + promo pre-builds.
Ironically this is where most POS forecasting implementations fail. They either skip the bridge and get whipsawed, or they build it in spreadsheets and lose it every cycle – rather strenuous without a system.
4๏ธโฃ ๐ ๐ฒ๐ฎ๐๐๐ฟ๐ฒ ๐๐๐ผ ๐ณ๐ผ๐ฟ๐ฒ๐ฐ๐ฎ๐๐๐, ๐ป๐ผ๐ ๐ผ๐ป๐ฒ – ๐ POS accuracy tells you if you understand the consumer. Shipment accuracy tells you if you understand the trade. Different problems, different owners, different fixes.
Forecasting shipments feels safer because that’s what the ERP measures and what Sales gets paid on. But you’re optimizing a lagging, distorted signal ๐.
For executives ๐ฏ: the payoff isn’t just accuracy. It’s lower channel inventory, fewer end-of-quarter surprises, cleaner S&OP conversations, and margin protection when promotions don’t lift the way sales promised.
It’s also the foundation for any serious conversation about CPFR, VMI, or joint business planning with your top retailers.
Shipment-based forecasting made sense when POS was expensive, late, and dirty. It’s now cheap, daily, and clean. โ
The question isn’t whether to flip the model. It’s why so many CPG companies still haven’t.
๐ฌ Curious to hear from planning leaders.
What’s actually stopping the flip in your organization?
Systems?
Monthly cadence discipline?
Sales still committing in shipment units?
Or something else?
Read our related posts:
Leave A Comment