When you manage a large Amazon business, the smallest demand shift can create a big ripple effect.
That was the reality for Infinite Commerce (now merged with Razor Group). With 18,000+ SKUs and a $400M+ Amazon business, the team had strong internal controls. They had disciplined planning across inventory, pricing, finance, and forecasting.
The problem was not growth.
It was predictability.
Amazon marketplace conditions change daily. Even small demand shifts at this scale created outsized downstream effects. Products sold faster or slower than planned. Inventory drifted out of balance. Teams were pushed into reactive decisions to protect availability and cash flow.
To solve this, Razor Group partnered with Trellis to implement Dynamic Pricing in order to use pricing as a continuous control system to absorb changes rather than forecast it away.
Keep reading to learn how.
The Challenge: Fragile Predictability
Amazon marketplaces evolve faster than operating plans can adapt. Demand shifts unevenly. Inventory falls out of sync. Teams are forced into reactive decisions to protect cash flow and availability.
At Razor Group, this challenge scaled quickly.
Even with strong internal controls and disciplined planning, daily marketplace changes created operational noise and reduced forecast reliability. Over time, forecasts became harder to trust – not because the modeling was wrong, but because conditions changed too quickly for static plans to hold.
Key challenges included:
- Marketplace volatility creating unpredictable demand shifts across 18,000+ SKUs
- Demand swings forcing emergency action toward steep discounts or
rush replenishment - Inventory forecast reliability declining due to daily market changes
- Capital tied up in slow-moving inventory
- Manual pricing interventions that couldn’t keep up
At this scale, the team didn’t need more forecasting. They needed a way to keep execution aligned as conditions changed.
Their goal was simple: use Trellis Dynamic Pricing to bring supply and demand into balance and improve forecast accuracy across the board.
The Solution: Trellis Dynamic Pricing & Forecast Accuracy
Before Trellis, price changes were handled manually. At this scale, it was tedious, error-prone, and slow. The team could not respond quickly enough to marketplace shifts, which meant demand drifted away from plan.
With Trellis Dynamic Pricing, Razor Group implemented pricing as a continuous adjustment layer between marketplace volatility and inventory and forecasting plans.
Instead of treating pricing as a one-time lever, Trellis treated it as a daily operating system. Pricing became a way to stabilize execution, using margin from overperforming products to support selective discounting on underperforming ones.
Trellis updated prices daily using live signals, including:
- Demand elasticity
- Competitor pricing
- Sales trends
- Advertising behavior
- Inventory position
Rather than aggressive discounting, Razor Group used pricing to rebalance uneven demand and keep sell-through rates closer to plan. Discounting was intentionally capped near 5%, which helped protect margin while still improving inventory movement.
The Results
With continuous pricing intervention through Dynamic Pricing from Trellis, Razor Group stabilized execution and built the predictability needed to plan with confidence.
Results at a glance:
- 18X ROI
- +4.5% improved margin
- 12% improved wMAPE (Weighted Mean Absolute Percentage Error)
Forecast accuracy improved materially, with wMAPE declining from ~48% to ~36%. Forecasts became usable operational inputs again, not just directional estimates.
Trellis also helped Razor Group improve inventory performance by shaping demand more efficiently across the catalog. Underperforming products recovered sales velocity, while overperforming products gained margin, reducing the need for broad, reactive discounting.
Compared to manual pricing, Trellis achieved better adherence to forecast with approximately 4.5% less discount, allocating pricing pressure more strategically across thousands of SKUs.
On discount efficiency alone, this delivered an estimated 18X return on investment, excluding any operational or planning benefits.
Execution became calmer and more reliable with a much more predictable cash flow, steadier sell-through, and the elimination of constant manual pricing intervention.
Trellis created a more predictable operating environment where sales targets could be met and inventory aligned much closer to plan.
Predictability at Scale Starts with Pricing You Can Trust
For large Amazon businesses, uncertainty is not an exception. It is the default.
Trellis Dynamic Pricing helps you respond to that reality with AI Precision + Human Intuition. It gives your team a way to shape demand, protect margin, and keep inventory aligned, even when the marketplace shifts daily.
Book a demo with Trellis today to see how we can work together too!