by Sarah Duffy

USE CASE
Assortment & Pricing Intelligence

Every Chief Merchandising Officer (CMO) has a version of the same story. A category is trending strong in week four of the season. Buying decisions build on that early signal. Six weeks later, the trend shifts, inventory is heavier than planned, and the markdown conversation begins.

This isn’t a reflection of poor judgement. It's the natural consequence of making high-stakes, high-velocity decisions with analytical tools built for a slower era. When the feedback loop between what's selling and what's being bought runs on weekly batch reports, even the best merchants are working with yesterday’s picture.

What is Markdown Optimization?

Retail markdown optimization is the practice of strategically reducing prices on slow-moving or end-of-life inventory to maximize gross margin while clearing stock by a target date. Rather than blanket discounts, optimization uses demand forecasts, sell-through rates, weeks of supply (WOS), and price elasticity models to recommend the right markdown depth on the right SKUs at the right time. Done well, it can lift margin rates versus reactive end-of-season markdowns.

Where Retail Markdown Optimization Breaks Down

Merchandising decisions sit at the intersection of trend data, inventory position, sell-through velocity, supplier lead times, and competitive pricing signals. Synthesizing all of that simultaneously — for a category with hundreds of SKUs, across dozens of locations — is exactly the kind of challenge where better data access creates outsized impact.

The Four Markdown Decisions

• Which SKUs: Not every slow mover warrants a markdown. Merchants must weigh sell-through velocity, weeks of supply, and trend trajectory to decide which products to act on.
• When to start: Timing is everything. Marking down too early sacrifices margin, too late forces deeper cuts and leaves less selling time.
• How deep: The discount has to be large enough to actually shift demand, but calibrated against remaining inventory, price elasticity, and margin targets.
• Where: The same SKU can be overstocked in one region and selling well in another, so markdown decisions often need to be made at the store or cluster level.

The real opportunity isn’t avoiding every markdown. The opportunity is closing the gap between when the data shows a shift and when the merchandising team can act on it.

Genie for Markdown and Merchandise Intelligence

Databricks Genie enables merchandising leaders to interrogate their entire data environment in natural language. A CMO can ask: 'Which categories are showing week-over-week sell-through deceleration greater than 10%, and what's our current inventory cover at current sell-through rates?' That question surfaces in seconds.

Why Earlier Markdown Decisions Protect More Margin

Retail competitive advantage has always had a timing dimension. The CMO who can redirect open-to-buy six weeks earlier — because they spotted the trend deceleration sooner — takes a better position on markdowns, holds more margin, and reallocates that capital to the categories that are winning. Genie doesn't make the buying decision. It gives your merchandising leaders the real-time clarity to make those decisions with confidence.

DATABRICKS GENIE · KEY DIFFERENTIATORS
Built for your data, governed by your rules, answerable to any business leader.

• Unified commerce data: Genie queries across e-commerce, stores, and wholesale channels in a single conversation — no channel-switching.
• Supplier data integration: Lead times and fill rates live in the same analytical environment as sell-through and margin data.
• Margin-aware answers: Questions about inventory automatically include margin context — decisions are grounded in financial impact, not just units.
• Historical pattern recognition: Genie can compare current sell-through patterns to comparable seasonal periods without requiring custom data pulls.

See What Genie Can Do for Your Team

Databricks Genie is available today. See how your industry peers are using it to reimagine how they access and act on their data.