Demand Planning for CPG Brands: Why Your Channel Matters More Than Your Forecast

Most demand planning advice treats your business like a single stream of orders. Build a forecast. Watch your inventory. Reorder when it gets low.

That works if you're selling one way to one customer.

CPG brands don't do that. You're selling DTC on your own site, fulfilling orders on Amazon, restocking shelves at Whole Foods, and managing quarterly resets at Target — sometimes all at once. Each of those channels has different lead time requirements, different visibility into what's actually happening at the point of sale, and different consequences when you get it wrong.

If you're running a blended demand plan, you're probably getting it wrong in all four places simultaneously.

Here's how to think about demand planning by channel — and what to watch for in each one.

DTC: The Channel With the Best Data and the Shortest Memory

Direct-to-consumer is where most CPG brands start, and it's also where your demand signal is cleanest. You can see every order in real time. You know your repeat purchase rate. You can tie a specific email to a spike in sales the next morning.

The problem isn't visibility. It's that founders learn to trust their DTC data a little too much, then apply that same confidence to channels where the data doesn't exist.

What makes DTC demand planning different:

• Replenishment is immediate. When someone runs out of your product, they can reorder in minutes. That means your demand curve is relatively smooth — there's no retailer buying in bulk and sitting on inventory before consumers ever see it.
• Promotions hit fast and decay fast. A 20%-off email sends a spike through your fulfillment center within 24 hours, then it's over. Plan for the spike, but also plan for the inventory you'll need to get back to baseline.
• Seasonality is visible. If you sell a sunscreen, you can watch DTC orders start climbing in April and plan your production run accordingly. Don't assume retail follows the same curve on the same timeline.

The DTC trap: Founders who scaled DTC often set inventory targets based on DTC velocity, then underproduce when they add retail accounts. The DTC signal is real — but it represents maybe 30-40% of the total demand once you're in a few regional chains. Plan your total demand, then allocate to channels.

Amazon: The Channel That Lies to You Until It Doesn't

Amazon sits between DTC and retail in most planning models, and it behaves like neither.

The demand signal is strong when things are going well. You can see unit velocity, review cadence, and search rank. But what you can't easily see is how much inventory Amazon is holding in their fulfillment centers, or whether a sales spike is organic demand or a search ranking artifact that won't repeat.

What makes Amazon demand planning different:

• FBA inventory is yours, but you can't see it clearly. Amazon receives your inventory, distributes it across their fulfillment network, and doesn't give you a clean view of what's sitting where. You can see aggregate on-hand, but not regional splits. If you're selling nationally, you may have 300 units in New Jersey and 12 in California — and Amazon's algorithm is starting to suppress your listing in the West.
• The velocity-rank feedback loop. Amazon ranks products based on sales velocity. Higher rank = more visibility = more sales. This works in your favor when you're stocked and in your against when you run out. A stockout on Amazon doesn't just lose you sales today — it drops your ranking, which costs you sales for weeks after you restock.
• Lightning Deals and promotions create artificial spikes. If you participate in Prime Day or run a coupon, your sell-through data for that period is noise. Strip it out of your baseline before using it to forecast the next period.

The Amazon trap: Brands that run out of FBA inventory often don't realize it's happening until the rank has already dropped. Build a 4-week restock lead time assumption into your Amazon plan, and check your FBA levels at least weekly during high-velocity periods.

Specialty Retail (Whole Foods, Sprouts, Target Food Hall): The Channel That Rewards Precision

Specialty retail is where CPG brands often land first before mass, and it's where demand planning starts to get genuinely complicated.

You're now dealing with real planograms, reset calendars, and buyers who expect you to know your velocity before they do.

What makes specialty retail demand planning different:

• Door count × velocity = your demand. Your demand plan is no longer about what consumers want in aggregate — it's about what consumers in the specific stores you're in want. A Whole Foods in Manhattan and a Whole Foods in suburban Ohio have different velocities for the same product. If you're in 200 doors, you need to plan by region, not just by total.
• Resets happen on a schedule. Specialty retailers typically reset shelves 2-4 times a year. If you miss a reset because you were out of stock, you may lose the shelf placement entirely until the next cycle. The cost of a stockout isn't just lost sales — it's lost distribution.
• You probably don't have POS data. Most specialty retailers don't share point-of-sale data with small suppliers. You're buying your product in (sell-in), but you don't know what's actually selling off the shelf (sell-through). You're flying partially blind. If you can buy syndicated data (Circana or SPINS), do it — it's worth the investment once you're in 100+ doors.

The specialty retail trap: Brands new to specialty retail often plan their inventory based on their initial buy order. The retailer places a 500-unit order, you produce 500 units, you're done. Then sell-through is better than expected, shelves go bare in week 3, and the buyer is calling you asking why half the doors are empty. Always produce more than the initial buy order for a new launch. Add 20-30% as a launch buffer.

Mass Retail (Target, Walmart, CVS): The Channel That Requires Its Own Plan

Once you're in mass, your demand planning complexity multiplies. This is a different game than specialty retail, and brands that treat it the same way get into serious trouble.

What makes mass retail demand planning different:

• Pantry loading is real. When a retailer takes on your product for the first time, they're not just filling their own shelves — they're loading inventory into their distribution centers across the country. That initial buy is much larger than what steady-state replenishment will look like. Plan for a big month 1 buy, followed by low or no orders in months 2-3 while they sell through what they have. Many brands misread month 2 silence as a demand problem. It's not.
• Chargebacks will find your gaps. Mass retailers have strict compliance requirements: label specs, case pack configurations, EDI setup, routing guides for their carriers. If any of these are wrong, you get charged back — which means you produced and shipped the inventory, but you're getting deducted from your invoice. Chargebacks are a planning problem as much as an operations problem. Build compliance verification into your launch checklist.
• Promotional calendars drive your production schedule. If you're participating in a Walmart Rollback or a Target Circle promotion, the lift can be 3-5x your baseline velocity for that window. Your demand plan needs to include every planned promotional event for the next 12 months, and your production timeline needs to account for the inventory required to support them.
• You need separate safety stock for each retailer. Don't pool your safety stock across channels. If Target goes out of stock, that's not Walmart's problem to solve. Each mass retailer has different replenishment timing, different DC locations, and different chargeback exposure. Treat them as separate demand streams.

The mass retail trap: The first year looks great. The retailer loaded inventory, you shipped everything on time, revenue hit the number. Year two looks like a disaster. Replenishment orders are low, and you're convinced the product isn't working. But the retailer is sitting on leftover inventory from year-one pantry loading — they don't need to reorder yet. Before you cut production or panic about the product, check their warehouse inventory levels. The sell-through may be fine.

The One Thing That Makes All of This Manageable

Every channel has different data quality, different timing, and different failure modes. That's the nature of CPG.

What makes it manageable isn't a better spreadsheet. It's treating each channel as a separate demand stream with its own forecast, its own safety stock calculation, and its own lead time logic — and then running a monthly process where you look at all of them together and make decisions about where to allocate inventory when you can't fully cover everything.

That monthly process is S&OP. And most CPG brands at $5M–$20M don't have a formal version of it.

If you've been running a blended plan, you already know what the symptoms look like: you're always short somewhere, always over somewhere else, and every planning cycle feels like you're reacting instead of deciding.

Not sure how to structure your demand plan by channel?

We do a free 20-minute demand planning audit — we'll look at how you're currently forecasting, where your channel gaps are, and what a working plan would look like for your business.

Book your audit at izba.co