For mid-market manufacturers & retailers

Your demand plan is costing you twice.

If you plan demand in spreadsheets, forecast error shows up on both ends of the balance sheet — lost margin when you stock out on the movers, and cash frozen in overstock on everything else. Most operators never put a number on it. This does.

I ran planning at a $250M furniture manufacturer. Spreadsheet S&OP quietly cost us every quarter — stockouts on the sellers, dead inventory on the rest. Moving to real planning plus AI forecasting cut both. Here's the math on your operation.

1. Estimate what forecast error costs you

Annual revenue total company revenue

Gross margin % — product margin

Average inventory value $ held on hand

Forecast error (MAPE) % — spreadsheet plans run 30–50%+

Transparent, conservative assumptions: stockout-driven lost sales ≈ 10% of your forecast-error rate applied to revenue (a 35% MAPE → ~3.5% of sales lost to stockouts); excess inventory ≈ 30% of forecast-error rate applied to on-hand stock; carrying cost 22%/yr; obsolescence/markdown 12% of excess. Modern planning + AI forecasting typically recovers ~30–40% of the total. These are deliberately conservative ballpark figures to size the problem — a real teardown uses your actual SKU and lead-time data, and usually finds more.

2. Score your planning maturity

0 = spreadsheet / gut · 3 = systematized. Below 8 means you're leaving the money above on the table.

Forecasting method

0: last-year-plus-x in Excel · 3: statistical + AI models, accuracy tracked by SKU

Data integration

0: manual exports · 3: live pipelines from ERP/POS/orders into the plan

S&OP cadence

0: ad-hoc · 3: a real monthly S&OP with one shared number

Scenario / what-if

0: none · 3: model promos, lead-time shifts, demand shocks in minutes

Inventory optimization

0: blanket safety stock · 3: service-level-driven, per-SKU, dynamic

Tap the dots to score each axis.

3. What modern planning + AI forecasting recovers

The fix isn't a bigger spreadsheet. It's a real planning model (one shared number, live data, scenario planning) with an AI forecasting layer on top. Typical results in mid-market manufacturing:

~20% forecast-accuracy gain — fewer stockouts on the sellers, less dead stock on the rest.
~30% inventory reduction — cash off the shelf and back into the business.
S&OP in days, not weeks — the team plans instead of wrangling spreadsheets.

The build path:

Map the planning model + wire live data from your ERP.
Stand up statistical + AI forecasts, benchmarked against your current accuracy.
Add scenario planning + per-SKU inventory logic.
Run it in parallel for one cycle, prove the accuracy lift, then cut over.

Want your real number?

Send me read-only access to your current plan (or a sample) and I'll do a free teardown — your actual forecast accuracy, where the cost is hiding, and what's recoverable. You keep the analysis either way.

Book a free teardown →

Jason · ex-VP of AI, $250M furniture manufacturer · planning + AI forecasting for mid-market makers