Get a demo

Demand forecasting

Aggregate sell-through data across retailers, project demand by SKU and channel, and flag stockouts before they hit the shelf. The planning team owns the forecast.

Get a demo

The prompt

I want to automate demand forecasting. Build me a flow that pulls sell-through data from each retailer portal, normalizes the SKU codes, projects demand for the next 8 to 13 weeks by SKU and channel, and flags any SKU that will stock out before the next replenishment lands.

Just copy and paste the prompt into a new Parabola flow to get started.

What Parabola builds

A workflow with six steps you can edit:

1. Pull sell-through. Daily or weekly POS data from each retailer portal. Plus DTC sales from Shopify or your ecommerce platform.

2. Normalize SKUs. Map each retailer's product code to your master SKU. Once at the top, then it is clean for everything below.

3. Project demand. Trailing 4, 8, and 13 weeks of sell-through. Trend, seasonality, and any known promotions layered in.

4. Compare to on-hand and on-order. Pull current inventory and outstanding POs from the ERP or WMS.

5. Flag stockouts. Any SKU where projected demand exceeds on-hand plus on-order before the next replenishment. Bucket by severity.

6. Output the forecast. Forecast table per SKU per channel, stockout queue for planning, replenishment recommendations.

Why teams stop doing this manually

Demand forecasting lives in a planner's spreadsheet. Each retailer's portal exports sell-through differently. SKU codes vary by retailer. Some retailers report weekly, some daily, some report on a lag. The planner downloads each one, normalizes the codes, joins to internal SKUs, and lays the trailing windows side by side.

The math is not hard. The data assembly is. By the time the planner has the file, half the analysis time is already gone. Stockouts are flagged after they have already happened.

What planning teams want is the assembly automated and the time spent on judgement. Which SKUs are trending up faster than the trailing window suggests. Which retailers are gaining shelf and which are losing. Which products need a deeper safety stock because the lead time just stretched.

A flow gets the data assembled. Planning takes it from there. The forecast refreshes every day, not every Monday morning.

How it works

Step 1. Paste the prompt.

Open Parabola, paste the prompt in section 2, and let it ask follow-up questions about your retailer mix, your SKU normalization rules, and your forecast horizon.

Step 2. Connect your data.

API connections or scheduled exports from each retailer portal, plus your ecommerce platform and your ERP for on-hand and on-order.

Step 3. Run it daily.

The flow refreshes the forecast each morning. Stockout queue updates. Replenishment recommendations land in the planning queue.

Related prompts

FAQ

Does this work if retailers report sell-through on different cadences?

Yes. Daily retailers feed daily. Weekly retailers feed weekly. The flow aligns them to a common week-ending convention.

How does it handle promotional uplift?

Add a promotion calendar as a side input. The flow applies an uplift multiplier per SKU per retailer during the promotion window and removes it after.

What about lead time and replenishment timing?

The replenishment recommendation factors in lead time per SKU. A SKU with a 12-week lead time gets flagged earlier than one with a 2-week lead time.

Can the flow handle new SKUs without trailing history?

Yes. New SKUs get a comparable-SKU forecast based on the planner's mapping. As real sell-through accumulates, the forecast switches over.

How is this different from a dedicated demand planning tool?

Dedicated tools are great when the demand planning team has 20 people. The flow is great when there is one planner and the work is currently in Excel. Same logic, less overhead.
See the stockout before it hits the shelf.
Paste the prompt, point it at your retailer portals and your ERP, and let the forecast refresh on its own.
Start for free