5 Ways to Stop Stockouts with Smarter Demand Planning
Stop losing sales to stockouts. 5 proven demand planning tactics: SKU-level forecasting, safety stock formulas, and daily early warning signals.
Stockouts aren't just frustrating—they're expensive. Every time a customer sees "out of stock," you lose revenue, erode trust, and create downstream chaos for sales and operations teams. Research by IHL Group estimates that retailers worldwide lose over $1 trillion annually to inventory distortion—stockouts and overstocks combined. For a mid-size product business, even a 5% stockout rate can translate into hundreds of thousands of dollars in missed sales each year.
The good news? Most stockouts are preventable with smarter demand planning and tighter execution. They're not random acts of fate; they're the predictable result of flawed forecasting, inadequate safety stock, and poor supplier visibility. Fix those systems and the stockouts largely go away.
Below are five proven ways to reduce stockouts, with specific actions you can take today. These strategies combine process discipline with data-driven forecasting—exactly what VNDLY is built to support.
1) Forecast at the SKU + Location Level
Global averages are convenient but misleading. Demand varies by region, channel, season, and even customer segment. If you only forecast at the global level, you'll overstock some locations and understock others—and you won't even know which until you start getting stockout complaints.
Why granularity matters
Imagine you sell a line of outdoor furniture. Your national average demand might look smooth and predictable. But when you break it down by warehouse region, you'll find that Pacific Northwest locations spike in spring while Southeast locations peak in autumn. A forecast that averages those two together will leave your Pacific Northwest warehouse empty in March while your Southeast warehouse is still overstocked.
Granular forecasting also matters for multichannel sellers. Wholesale demand from independent retailers often has a completely different pattern from direct-to-consumer demand on Shopify. Mixing those signals produces a blended forecast that serves neither channel well.
What to do:
- Build forecasts at the SKU + location level, not just the product family or national level.
- Account for regional seasonality (weather, holidays, trade shows, regional promotional calendars).
- Segment demand by channel if you sell online and wholesale simultaneously.
- Review each location's demand data independently before rolling up to aggregate reports.
VNDLY helps you set location-level reorder points and safety stock so replenishment decisions reflect actual local demand. The result: fewer "surprise" stockouts and less emergency freight. When a location dips below its reorder point, a purchase order suggestion fires automatically—you're never relying on someone to notice a problem on a spreadsheet.
Practical example
A skincare brand with three warehouse locations used national averages to reorder stock. After switching to location-level forecasting, they reduced stockouts at their East Coast distribution center by 34% in the first quarter, while simultaneously reducing total inventory by 8% because they stopped over-ordering at their Midwest location to compensate for uncertainty.
2) Use the Right Forecast Model for Each SKU
No single forecasting model fits every product. Fast movers with steady demand benefit from simple moving averages, while volatile or seasonal products require more adaptive models. Applying the wrong model is like using a weather forecast for London to plan an outdoor event in Phoenix—the structure is right but the assumptions are completely wrong.
Understanding the main forecast approaches
Simple Moving Average (SMA): Takes the mean of the last N periods. Works well for stable, non-seasonal products with low variability. Easy to understand and explain to non-technical teams. The downside is that it responds slowly to real shifts in demand.
Weighted Moving Average (WMA): Similar to SMA but assigns higher weight to recent periods. Better for products with gradual trends. If your product is slowly growing in popularity, WMA will respond to that trend faster than SMA.
Exponential Smoothing (ES): Uses a decay factor to weight recent periods more heavily while still incorporating historical data. Particularly useful for products with irregular or volatile demand. More complex to configure but often more accurate.
Seasonal Decomposition: For products with strong seasonal patterns—holiday goods, summer items, back-to-school supplies—you need to explicitly model the seasonal component. This means identifying the seasonal index for each month of the year and applying it as a multiplier on top of the base trend.
A practical approach:
- Stable demand (coefficient of variation below 0.3) → Simple Moving Average or Exponential Smoothing with low alpha
- Trending demand → Weighted Moving Average or Double Exponential Smoothing
- Volatile demand (coefficient of variation above 0.7) → Exponential Smoothing with higher alpha, or causal models
- Seasonal products → Seasonal decomposition with any of the above as the base trend model
By selecting the right model for each SKU category, you improve forecast accuracy and avoid the "forecast whiplash" that leads to alternating overstock and stockout cycles. VNDLY makes it easy to set forecast model defaults at the product category level and override them for specific high-value SKUs.
How to audit your current forecast accuracy
If you're not sure whether your current forecast model is working, calculate Mean Absolute Percentage Error (MAPE) for your top 50 SKUs over the last three months. A MAPE above 30% is a red flag—it means your forecasts are regularly off by more than 30%, and you're likely experiencing unnecessary stockouts as a result. Aim for MAPE below 15% on your high-volume SKUs.
3) Set Safety Stock Based on Variability, Not Intuition
Safety stock is your buffer against uncertainty. It's the inventory you hold above your expected demand to absorb variability in both demand and supply. But many teams set it arbitrarily—"let's keep two weeks of stock as buffer"—or skip it entirely. Both approaches are costly.
Setting safety stock too high wastes capital and inflates carrying costs. Setting it too low leads to stockouts during demand spikes or supplier delays. The right approach is statistical.
The safety stock formula
The most commonly used formula for safety stock is:
Safety Stock = Z × σ_demand × √Lead Time
Where Z is the service level factor (1.28 for 90%, 1.65 for 95%, 2.05 for 98%), σ_demand is the standard deviation of daily demand, and Lead Time is the supplier lead time in days.
A more complete formula that also accounts for lead time variability:
Safety Stock = Z × √(Avg Lead Time × σ_demand² + Avg Daily Demand² × σ_lead_time²)
Practical example
You sell a product with average daily demand of 10 units, demand standard deviation of 3 units, average lead time of 14 days, and lead time standard deviation of 2 days. You want 95% service level (Z = 1.65).
Safety Stock = 1.65 × √(14 × 9 + 100 × 4) = 1.65 × √(126 + 400) = 1.65 × √526 = 1.65 × 22.9 = 38 units
Without this calculation, a team might guess "keep 2 weeks stock" = 140 units, massively overcapitalizing on buffer. Or they might guess "keep 5 days" = 50 units, leaving them frequently understocked during demand spikes.
Service level trade-offs
Higher service levels require exponentially more safety stock. Going from 90% to 99% service level can require 2-3x more safety stock. This is why ABC analysis matters—your top-margin, high-velocity SKUs warrant 98-99% service levels, while slower movers might be fine with 90-92%.
VNDLY's planning engine can automate safety stock calculations, ensuring high service levels without excessive carrying costs. It tracks your actual demand variability and lead time history and recalculates safety stock recommendations quarterly.
VNDLY automates steps 1–3 for you
Set location-level reorder points, pick the right forecast model per SKU, and get safety stock calculated automatically. Free 14-day trial.
Try VNDLY free →4) Align Purchasing with Lead Times and Supplier Reliability
Stockouts often happen when lead times are underestimated or supplier reliability slips. You can't plan effectively without accurate lead time data—and "accurate" means based on what actually happened, not what was promised.
The lead time accuracy problem
Many businesses use the supplier's quoted lead time as their planning assumption. But if your supplier quotes 10 days and actually delivers in 12-15 days 40% of the time, your reorder point is systematically too low. Every quarter, several SKUs will stockout during the gap between when you expected delivery and when it actually arrived.
The fix is straightforward: track actual vs. promised delivery dates for every purchase order, then use the 85th or 90th percentile of actual lead times in your planning calculations—not the average and certainly not the quote.
Building a supplier lead time database
- Record the promised delivery date on every PO at the time of creation
- Record the actual receipt date when inventory arrives
- Calculate the variance: actual minus promised (positive means late)
- Review supplier performance quarterly and update lead time assumptions
Flagging unreliable suppliers
A supplier who is consistently late isn't just an inconvenience—they're a stockout risk that you're embedding in your inventory model. Consider these thresholds:
- On-time delivery below 80%: Increase safety stock for all SKUs from this supplier
- On-time delivery below 60%: Escalate with supplier, identify backup source, or increase reorder lead time assumption by 50%
- Average delay above 5 days: Consider shifting to a more reliable alternative
In VNDLY, supplier lead times are captured directly on supplier records and can be overridden at the SKU level. When you receive a shipment late, the system prompts you to update the lead time record. Over time, your planning assumptions become a data-driven reflection of actual supplier performance rather than optimistic quotes.
Diversifying your supplier base
For your top 20% of SKUs by revenue, consider qualifying a secondary supplier even if you rarely use them. The cost of qualifying a backup supplier is almost always less than the cost of a week-long stockout on a high-velocity product. Having that fallback means you can place an emergency order when your primary supplier signals a delay—before the stockout actually occurs.
5) Monitor Early Warning Signals Daily
The fastest way to reduce stockouts is to catch issues before they happen. That requires daily visibility into risk indicators—not just monthly reports that tell you what already went wrong.
The reactive planning trap
Most teams use inventory reports that show current stock levels. That's useful, but it's backward-looking. By the time a report shows you're out of stock, the damage is already done. What you need instead are forward-looking signals that tell you when a stockout is likely to happen, and how many days you have to prevent it.
Key early warning signals to monitor daily:
- SKUs below safety stock level: These are already in the danger zone. A stockout may be days away depending on demand pace.
- Days of cover below threshold: Calculate days of cover as current stock divided by average daily demand. Any SKU below 14 days needs attention.
- Open purchase orders past due: These represent expected inventory that hasn't arrived. Each overdue PO is a ticking stockout clock.
- Demand spikes above 2x average over the last 7 days: A sudden increase in orders can burn through your buffer faster than your forecast expected.
- Backorder volume increasing: Rising backorders on a specific SKU are a clear signal that current stock isn't meeting demand.
Setting up an actionable daily review
The goal is a morning review that takes 10-15 minutes and surfaces the 5-10 SKUs that need action today. Not a 200-row spreadsheet that takes an hour to parse—a focused list of what to act on right now.
Structure your daily review around:
- Critical: Items already out of stock or backordered
- High risk: Items below safety stock or less than 7 days of cover
- Watch list: Items with overdue POs or unusual demand spikes
VNDLY provides real-time low-stock alerts, projected stockout dates, and dashboards that highlight the highest-risk items. You can configure alert thresholds by product category and service level target, so your team gets the right alerts for the right SKUs—not a firehose of noise.
Bonus: Align Sales, Marketing, and Planning Calendars
A common cause of stockouts is a sudden demand spike driven by promotions or new channel launches that planning never saw coming. If marketing schedules a campaign and ops doesn't know about it, forecasts will miss the true demand signal. The fix is simple: create a shared calendar that includes promotions, product launches, seasonal pushes, and wholesale events.
How promotional misalignment causes stockouts
Marketing schedules a 30% off email blast for next Tuesday. Forecasting doesn't know about it. A normal Tuesday might generate 50 orders—but this Tuesday generates 340. By Wednesday morning, your top three SKUs are backordered for two weeks.
This isn't a forecasting failure—it's a communication failure. The forecasting model did exactly what it was supposed to do with the information it had. The problem is that marketing and operations weren't sharing a planning calendar.
Creating a demand visibility process:
- Hold a monthly S&OP (Sales and Operations Planning) meeting with reps from sales, marketing, and ops
- Require marketing to submit promotional plans to ops at least 3 weeks in advance
- Build a shared promotion calendar that feeds into demand forecasting
- When a promotion is confirmed, manually adjust the forecast for affected SKUs for the promotion period
In VNDLY, you can annotate demand planning cycles with known events and adjust forecast weights accordingly. Even a quick adjustment—like increasing the forecast by 25% during a promotion window—can prevent a surprise stockout. The goal is to reduce "unknown" demand by turning it into planned demand.
Bonus: Prioritize with ABC/XYZ Segmentation
When everything feels urgent, nothing is. ABC/XYZ analysis helps you focus on the SKUs that matter most so your team doesn't spend equal time on a high-margin bestseller and a slow-moving clearance item.
ABC segmentation by revenue contribution:
- A items: Top 20% of SKUs by revenue (typically 80% of total revenue)
- B items: Next 30% of SKUs (roughly 15% of revenue)
- C items: Bottom 50% of SKUs (roughly 5% of revenue)
XYZ segmentation by demand variability:
- X items: Consistent, predictable demand (coefficient of variation below 0.3)
- Y items: Moderate variability, some seasonality (coefficient of variation 0.3-0.6)
- Z items: High variability, difficult to predict (coefficient of variation above 0.6)
How to apply it:
- AX items: Highest priority. Set aggressive service levels (98% or higher), review weekly, maintain high safety stock.
- AY items: High priority. Regular review, solid safety stock, watch for seasonal patterns.
- AZ items: High revenue but volatile. Maintain buffer stock, consider flexible sourcing, review frequently.
- BX/BY items: Moderate attention. Automated reorder points with quarterly review.
- BZ/CZ items: Low priority. Lean inventory, longer reorder cycles, may tolerate occasional stockouts.
By segmenting inventory this way, your team can prioritize planning time on the highest-impact SKUs. That reduces the risk of stockouts where they hurt most—high-velocity, high-margin items—while avoiding over-investment in slow movers.
The Hidden Cost of Stockouts You're Probably Underestimating
Most teams measure stockout cost as lost sales: the revenue you didn't collect because you had no inventory. That's real, but it's only part of the story.
The full cost of a stockout includes:
Lost immediate revenue: The obvious one. Customer wanted to buy, couldn't.
Lost future revenue: Studies suggest 30-40% of customers who encounter a stockout don't come back. They find a competitor and stay there. For a subscription customer or a wholesale account, that's not one lost order—it's a lost lifetime value.
Customer service costs: Fielding "where's my order" emails, processing backorder communications, issuing apologies and potentially discount codes to frustrated customers.
Emergency procurement costs: Rush orders, premium freight, expedite fees. These can be 2-5x the cost of standard procurement.
Sales team disruption: Your sales agents spend time explaining stockouts to wholesale customers instead of closing new business.
Brand damage: Social media complaints, negative reviews, and word of mouth all have long-tail impacts on customer acquisition costs.
When you add it up, the true cost of a significant stockout event is often 3-5x the face value of the lost sales revenue. That context makes the investment in better demand planning look very different—the ROI on preventing even one major stockout event can be 10x or more.
Implementation Checklist: Getting Started This Week
Reducing stockouts isn't a single fix—it's a system. When you combine location-level forecasting, tailored models, data-driven safety stock, accurate lead times, and daily monitoring, you turn inventory planning into a competitive advantage.
Week 1:
- Audit your current MAPE (forecast error) by product category
- Identify your top 50 SKUs by revenue and categorize by demand variability
- Calculate proper safety stock for your A items using the formula above
Week 2:
- Set up location-level reorder points in your inventory system
- Create a shared promotional calendar with marketing
- Establish daily monitoring thresholds for low-stock alerts
Month 1:
- Implement ABC/XYZ segmentation across your full catalog
- Review and update supplier lead time assumptions with actual delivery data
- Run your first monthly S&OP meeting
Ongoing:
- Review forecast accuracy monthly for A items
- Update safety stock calculations quarterly
- Evaluate supplier performance against on-time delivery targets
With VNDLY, these workflows are built into your daily operations. Instead of juggling spreadsheets, your team gets clear replenishment recommendations, smarter purchasing, and fewer missed sales. The planning module handles location-level forecasting, safety stock automation, supplier lead time tracking, and early warning dashboards—all in one place.
If you're ready to reduce stockouts and improve service levels, start by reviewing your top 20 SKUs in VNDLY's Planning module. Small improvements there often deliver the biggest wins.
Frequently Asked Questions
What is the main cause of stockouts? The most common causes are inaccurate demand forecasts, safety stock set too low or not set at all, supplier lead times that are longer or more variable than assumed, and lack of real-time visibility into inventory levels. Promotional demand spikes that operations didn't plan for are also a leading cause.
How do you calculate the right safety stock level? The standard formula is: Safety Stock = Z × √(Avg Lead Time × σ_demand² + Avg Daily Demand² × σ_lead_time²), where Z is the service level factor. For 95% service level, Z = 1.65. You can use a simplified version—Z × σ_demand × √Lead Time—if lead time variability is low.
What's a good stockout rate to target? For A-category items, a stockout rate below 2% (98% fill rate) is a common target. For B items, 5% or less. C items can tolerate higher rates. Track your fill rate by SKU category and set improvement targets each quarter.
How often should I update my demand forecasts? For high-velocity A items, update forecasts weekly or monthly. For B items, monthly or quarterly. Seasonal products should have forecasts reviewed before each seasonal build-up period. At minimum, review forecast accuracy for your top SKUs every month.
What is the difference between a reorder point and safety stock? Safety stock is the buffer you maintain above zero to absorb variability. The reorder point is the level at which you trigger a new purchase order. Reorder Point = (Average Daily Demand × Lead Time in Days) + Safety Stock. If your average daily demand is 10 units, lead time is 14 days, and safety stock is 30 units, your reorder point is 170 units.
Can inventory software really prevent stockouts? Yes, significantly. A system like VNDLY provides the real-time visibility, automated reorder triggers, supplier lead time tracking, and demand planning tools that make systematic stockout prevention possible. Spreadsheets can theoretically do the same calculations, but the manual overhead makes it impractical to monitor hundreds or thousands of SKUs daily.