AI Inventory Optimization & Merchandise Planning Software

Omnichannel Inventory Isn’t Shared — It’s Contested

Most teams still talk about omnichannel like it’s one clean pool of inventory. It’s not. Inventory gets pulled in different directions all day.

A unit in a store stockroom can be:

Sold to a walk-in
Picked for an online order
Reserved for BOPIS

That tension shows up fast.

You walk a store — it looks full. Then you check sizes. Core sizes are gone. The floor looks fine, but conversion drops because customers can’t find what they need.

Meanwhile, ecommerce says “available” because the unit technically exists somewhere. Orders come in. Fulfillment slows down because the stock is sitting in a store not built for picking.

And the allocation behind all this? Made weeks ago, based on demand that’s already shifted.

This is the part that gets missed: channel conflict happens at the SKU level.

Stores need depth across sizes to avoid broken runs
Ecommerce just needs one unit to convert
Fulfillment wants the fastest, cheapest node

Those priorities don’t line up.

You’re not eliminating stockouts. You’re deciding where they’re acceptable.

Stock out online to protect store sell-through? Sometimes. Pull from stores to save an ecommerce order? Also sometimes.

This isn’t “unified inventory.” It’s managed trade-offs.

Visibility Is Table Stakes — Not the Solution

Everyone starts with visibility. Fair enough. You need it.

But even “good” visibility is hard:

SKU-level accuracy across nodes
Real-time POS (not yesterday’s data)
In-transit clarity
Returns separated from sellable stock

Let’s say you actually have all that.

You can still end up with:

Size breaks in stores
Overstock in the wrong locations
Ecommerce stockouts while total inventory looks healthy

Visibility shows the problem. It doesn’t fix it.

Most teams stop at dashboards — WOS, sell-through, stock levels — and then someone decides what to do. That lag is where margin leaks.

What you need is action tied to signals.

If a store drops below a size threshold, something happens.
If ecommerce demand spikes regionally, inventory shifts.
If one cluster sits at 12+ WOS and another at 3, transfers trigger.

Without that, visibility just makes problems easier to see — not easier to solve.

Forecasting by Channel and Node — Not Just Totals

Forecasting is still too blunt in most orgs.

Plans sit at:

Category level
Total channel level

That breaks down fast in omnichannel.

A SKU can be:

Overstocked in stores
Understocked online
“Fine” in total

And you won’t catch it early.

Forecasting needs to go deeper:

Channel (store vs ecommerce)
Location (region, cluster, node)
SKU variant (size, color)

This is where things get messy. Demand isn’t clean.

Medium sells in one region. Large in another. Online behaves differently altogether.

If you don’t plan at that level, allocation becomes guesswork.

Yes, models are better now — factoring in promos, local signals, weather, events. That helps.

But forecasting doesn’t fix execution lag.
It doesn’t rebalance mid-season.
It doesn’t stop size breaks two weeks in.

Forecasting sets the starting point. Allocation decides what actually happens.

Allocation and Reallocation — Where It’s Won or Lost

Most problems don’t come from buying the wrong quantity.

They come from putting inventory in the wrong place.

Initial Allocation — Where Mistakes Get Locked In

Common issues:

Outdated store clustering
Ignoring regional ecommerce demand
Even size curves when demand isn’t even

You see it early.

A new denim fit drops. Every store gets the same size curve. Two weeks in:

Urban stores blow through smaller sizes
Suburban stores sit on larger sizes
Ecommerce pulls from both, creating gaps everywhere

Now you’ve got:

Early size breaks where demand is strongest
Slow stock where it isn’t
Missed full-price sales you won’t recover

The instinct is to blame forecasting. But often the total buy was fine.

Allocation broke it.

Dynamic Reallocation — The Missing Muscle

Strong teams don’t treat allocation as one-and-done.

They set triggers:

WOS imbalances
Sell-through thresholds
Ecommerce vs store divergence

Then they act:

Store-to-store transfers
Pulling inventory back to DC
Redirecting inbound stock

And they prioritize:

Core sizes
High-velocity SKUs
High-margin or low-cover items

Take a simple case.

A sneaker is trending online. Stores still have stock. Ecommerce is running dry.

You can:

Let ecommerce stock out
Pull from stores

Pulling drives online revenue. But it creates store gaps, which hurts walk-in conversion.

There’s no clean answer. You’re trading one lost sale for another.

The difference is whether you chose it — or reacted too late.

Allocation isn’t a plan. It’s a loop.

Fulfillment Strategy Is Inventory Strategy

Fulfillment gets treated like ops. It’s not. It reshapes inventory in real time.

The main models:

Ship-from-DC
Ship-from-store
BOPIS

Each changes inventory behavior.

Ship-from-store improves speed, but adds volatility. A few orders can wipe out key sizes before peak traffic.

BOPIS depends on accuracy. If stock is off, you cancel or disappoint at pickup.

Central fulfillment protects stores, but slows delivery and increases cost.

So teams layer in rules:

Minimum store stock thresholds
SKU restrictions for ship-from-store
Channel prioritization

These rules matter.

Unrestricted ship-from-store boosts ecommerce — but degrades store experience.
Locking everything to DC protects stores — but hurts online availability.

You’re always balancing:

Speed vs availability
Cost vs conversion
Store experience vs ecommerce growth

Every fulfillment promise is an inventory decision.

The Real Problem: Stockouts, Cash, and Margin

This is a financial equation.

You’re balancing:

Stockouts (lost revenue)
Inventory levels (cash tied up)
Markdown risk (margin loss)

You can’t optimize all three.

More inventory reduces stockouts — but increases markdown risk.
Less inventory frees cash — but increases lost sales.
Misallocation gives you both problems at once.

Inventory is frozen cash.

If it’s sitting in the wrong store, wrong size, at 10+ WOS — that’s not just operational. It’s capital inefficiency.

Then returns complicate everything.

In some categories, ecommerce returns hit 20–25%. That stock comes back:

Late
In mixed condition
Often to the wrong place

It’s technically inventory. But not sellable right away.

That creates false signals.

The system thinks you’re covered. You’re not. Replenishment slows. Allocation decisions drift.

You see it in basics.

A core tee shows healthy stock. Stores are still breaking on medium. Why? Because some of that inventory is stuck in returns or low-demand locations.

By the time it’s fixed, full-price demand is gone.

What a Practical Model Looks Like

You accept controlled stockouts.

You protect availability where it matters:

Core SKUs
High-margin items
High-traffic channels

You allow gaps elsewhere.

You rebalance continuously based on real demand — not static plans.

And you stay honest about trade-offs.

Omnichannel optimization isn’t about getting everything right.

It’s about choosing where you’re willing to be wrong.