We Made a Million Extra Cookies by Accident. Here's Exactly How That Happens.

Episode 5 of "How Are We Even Still in Business?!" — F&W Cookie's weekly series on the real, expensive, occasionally jaw-dropping business lessons from inside a growing food brand. Previous episodes covered a $10K newsletter ad that returned $800, paying cash for the building when a 4% loan was smarter, buying $2M in equipment before proving demand, and the $75,000 bagging machine that fell off the back of a truck. This episode involves a million cookies and a hard lesson in what demand actually means.

Here's a sentence that sounds like it should be followed by celebration: our wholesale orders were strong, our velocity was climbing, and everything our projections were telling us pointed in one direction.

Make more cookies. Make a lot more cookies. Make — and we cannot stress this enough — one million more cookies.

So we did.

And then the demand that supposedly justified all of those cookies turned out to not be real.

Which left us sitting on a mountain of cookies that don't age like fine wine, trying to figure out what you do with a million units of perishable product that nobody is currently buying.

The answer, it turns out, is liquidators. The follow-up answer is: if you have ever picked up an F&W Cookie product at a random store for a dollar, now you know why.

How the Projections Got Built — And Why They Looked So Good

To understand how this happened, you have to understand a little bit about how distribution works in the consumer packaged goods world, because the mechanism behind this mistake is not unique to F&W Cookie. This is a trap that has caught brands significantly larger and better-resourced than we were at the time.

In CPG distribution — the world of wholesale, retail shelf placement, and the supply chains that connect food manufacturers to grocery stores — velocity is the metric that matters most. Velocity, in this context, means how quickly your product moves off the shelf or out of the distributor's warehouse. High velocity means you're selling fast. High velocity means retailers want to reorder. High velocity is the signal that tells everyone in the supply chain: this product is working.

Distributors track velocity. Their systems model it. And those systems, when they see velocity climbing, do what any reasonable model does when presented with an upward trend: they project it forward. They assume the trajectory is real and sustained. They tell you — and they tell themselves — that demand is growing, and that production should grow to meet it.

Our projections were telling us to produce significantly more than we had been producing. The velocity numbers backed it up. The distribution system was essentially saying: you're about to explode. Make the cookies.

So we made the cookies.

The Promo That Broke the Algorithm

Here is what was actually happening underneath those velocity numbers.

We ran a promotion at a sales show. Discounted pricing, available to buyers at the event — the kind of deal that incentivizes stocking up because the savings are real and the product is good and why wouldn't you buy more when the price is lower than it's going to be next week?

Buyers stocked up. Really stocked up.

And from the outside, from the perspective of a distributor's demand forecasting algorithm that doesn't have the context of why those orders came in — that spike looked like organic, sustained, earned velocity. It looked like a product that was genuinely taking off in the market. It looked like the new normal.

It was not the new normal.

It was a single, time-limited, discount-driven event that pulled forward demand that would have otherwise been spread across multiple future periods — or might not have existed at all at that volume without the discount. Buyers stocked up because they got a deal. The moment the deal ended, so did the behavior that created the spike.

But the algorithm didn't know that. The algorithm saw the numbers go up and told the model: this is what this product does now. And the model told us: produce accordingly.

We produced accordingly.

A Mountain of Cookies

The inventory piled up in a way that has a specific kind of dread attached to it if you've ever experienced it in a food business.

Shelf-stable products have some forgiveness built in. Cookies — even well-made, properly packaged artisan cookies — have a shelf life. They are not getting better with time. Every day they sit is a day closer to a point where they can't be sold at full price, or at any reasonable price, or at all. The clock on perishable inventory doesn't stop because the demand projections were wrong. It just keeps moving, indifferent to your feelings about the situation.

We were looking at a very large amount of product, a very finite window to move it, and a market that had already taken on as much F&W Cookie inventory as the promotional spike had temporarily justified.

The full-price wholesale channel wasn't going to absorb it. The DTC channel wasn't going to absorb it. Retail wasn't going to absorb it. And the clock was running.

So we did what food brands in this situation do: we called liquidators.

Liquidators buy excess and near-date inventory in bulk — at prices that are, to put it diplomatically, significantly below what you would like to receive for product that cost real money to make. They move it through discount retail channels, closeout stores, and off-price outlets. It clears your warehouse. It gets the product to consumers. And it returns pennies compared to what the inventory cost to produce.

If you have ever wandered through a discount store and seen a brand you recognize on the shelf for a dollar — that's often what happened. A projection went wrong somewhere, inventory built up, and the liquidator is how it gets out the door.

If you've ever seen F&W Cookie at a random store for a dollar: that was us. That was this.

What the Algorithm Was Actually Measuring

The mistake here wasn't naïvety or carelessness. The mistake was treating a distributor's algorithmic demand signal as ground truth without understanding what was actually driving it.

Distribution algorithms are powerful tools. They're built on real data and they do exactly what they're designed to do. But they measure what happened — not why it happened. They see the spike in orders. They do not see the discount that caused the spike. They do not know whether that spike represents durable consumer demand or a one-time inventory load-in driven by a price incentive. They just see the number go up and model accordingly.

The burden of adding that context — of understanding the why behind your velocity numbers — sits with the brand. And in this case, we didn't add it. We saw projections that matched what we wanted to believe about our trajectory and we didn't ask hard enough questions about what was underneath them.

Real demand and promotional demand are completely different things, and confusing them is one of the most expensive mistakes a CPG brand can make.

Real demand is what happens when consumers buy your product because they want it at the regular price. It's repeatable, sustainable, and tells you something true about whether your product has a market.

Promotional demand is what happens when consumers — or more often, in the wholesale context, buyers — stock up because you gave them a reason to that isn't going to be there next time. It can look exactly like organic growth in the short term. It will not look like anything once the promotion ends.

The only way to tell the difference, before it shows up as a million extra cookies in your warehouse, is to track your velocity in context — to know what promotions ran when, what events influenced purchase behavior, and what the baseline order pattern looks like without any external artificial lift.

We know that now. We paid a million cookies for the lesson.

What We'd Tell a CPG Brand Watching Their Velocity Numbers Climb

This is the version of the lesson we wish someone had handed us before we ramped production:

Contextualize every spike. Before you treat a velocity increase as proof of real demand growth, ask what happened in that period that might have caused it. Promotions, shows, price changes, competitor stockouts, seasonal timing — any of these can create short-term volume that has nothing to do with your product's sustainable demand trajectory.

Separate sell-through from sell-in. In wholesale distribution, "sell-in" is how much product moved from you to the distributor. "Sell-through" is how much product moved from the distributor or retailer to actual end consumers. A promo spike often shows up as a massive sell-in event followed by a very slow sell-through period, because buyers stocked up on more than they can actually sell. Watching both numbers tells a much more honest story than watching either one alone.

Build a demand baseline without promotions. Your real demand signal is what velocity looks like in periods with no promotional activity. That's your floor. Everything above it should be understood as incremental, not permanent, until it sustains itself over multiple non-promotional periods.

Model conservatively when signals are ambiguous. When velocity is spiking but you're not sure why, the responsible production decision is to produce toward the conservative end of your range until you understand what's driving it. The upside of slightly undersupplying a real demand surge — a lost sale here and there — is much cheaper than the downside of massively overproducing against fake demand.

Know your shelf life before you scale. The more perishable your product, the lower your tolerance for inventory miscalculation. Cookies are not wine. Know exactly how long you have to move product at full value and make sure your production and inventory decisions account for that constraint explicitly.

The Part Where We Admit the Rest of It

There's a version of this story where we just share the lesson and move on. But that's not quite how this series works.

So here's the rest of it: the liquidation process is demoralizing in a specific way that the financial loss doesn't fully capture. You made those cookies. You made them with real ingredients, with real labor, with real care for quality. Watching them leave the warehouse at a price that doesn't cover your cost of goods, headed to an outlet store shelf where someone will pick them up without knowing anything about the brand behind them — it's not nothing.

It's the kind of thing that stays with you. The kind of thing that rewires how you think about production decisions for a long time afterward.

We are more careful now. We ask harder questions about what's driving our velocity numbers. We look for the promo context behind any spike before we treat it as a signal to produce more. We watch sell-through, not just sell-in.

And we are still standing, which continues to defy several reasonable projections.

TL;DR — The Breakdown

What happened: A sales show promo created a one-time spike in wholesale orders that the distributor's algorithm misread as real sustained velocity
What we did: Ramped production based on those projections — a million extra cookies
What the demand actually was: Promotional, one-time, not real sustained consumer pull
What happened to the excess inventory: Sold to liquidators at a severe loss
Where you might have seen it: Random stores, discount outlets, dollar price points — that was the liquidation inventory
The core lesson: Real demand and promotional demand are completely different. Contextualize every velocity spike before you produce against it.

FAQ: The F&W Cookie Million Cookie Mistake and What It Teaches About CPG Demand

What is velocity in CPG distribution? In consumer packaged goods (CPG), velocity refers to how quickly a product sells through a distribution channel — how fast it moves off shelves or out of a distributor's warehouse. High velocity signals strong consumer demand and encourages retailers and distributors to reorder. It's one of the most closely watched metrics in wholesale food distribution.

What is a promo spike in demand forecasting? A promo spike is a temporary, artificial increase in sales or orders caused by a promotional event — a discount, a sales show, a special offer — rather than by organic consumer demand. Promo spikes can look identical to real demand growth in the short term but don't reflect sustainable purchasing behavior once the promotion ends.

How did F&W Cookie end up with a million extra cookies? F&W Cookie ran a discount promotion at a sales show, which caused buyers to stock up heavily. The distributor's algorithm interpreted that one-time spike as evidence of rising organic demand and projected it forward. F&W Cookie produced against those projections — making a million more cookies than actual sustained demand could support. When the promotional lift disappeared, so did the demand.

What is the difference between sell-in and sell-through in wholesale distribution? Sell-in is the volume of product that moves from a manufacturer to a distributor or retailer. Sell-through is the volume that moves from the distributor or retailer to actual end consumers. A promo spike typically drives a large sell-in event, but if consumers aren't buying at the normal rate, sell-through stays low and the retailer or distributor is left with excess inventory they won't reorder.

What happens when a food brand overproduces? When a food brand produces more than demand can absorb, they face an inventory problem that worsens over time as products approach their shelf-life limits. Options include discounting deeply to move product, selling to liquidators at a significant loss, or in worst cases, writing off inventory entirely. Each option results in a financial loss that compounds the original production cost.

What is a liquidator in the food industry? A liquidator is a company that purchases excess, near-date, or overstock food inventory in bulk at prices far below the original wholesale value. They resell the product through discount retail channels, closeout stores, and off-price outlets. Selling to liquidators clears warehouse space but typically returns only a fraction of the product's cost of goods.

What should CPG brands do to avoid overproduction mistakes? Key practices include: contextualizing velocity spikes (understanding what events drove them), tracking sell-through separately from sell-in, establishing a demand baseline from non-promotional periods, modeling production conservatively when demand signals are ambiguous, and always accounting for product shelf life when making inventory decisions.

Is this part of the F&W Cookie "How Are We Even Still in Business?!" series? Yes. This is Episode 5 of F&W Cookie's ongoing series covering real financial and operational mistakes made while building their brand. Previous episodes include a $10K newsletter ad, paying cash for the building, $2M in premature equipment investment, and a $75K bagging machine that fell off a truck.

Follow F&W Cookie on TikTok for new episodes of "How Are We Even Still in Business?!" every week — and for Mariah's weekday morning Live flavor drops, the HEB launch updates, and whatever the secret project announcement turns out to be.