Meltable Season is coming! Get the full meltable products list.

Free Download
Hero Section Background

How to Set Restock Levels to Avoid Lost Sales: 2026 Guide

how to set restock levels to avoid lost sales

TL;DR

Your restock level (reorder point) is the inventory quantity that triggers a new purchase order so replenishment arrives before you stock out. Calculate it as average daily demand multiplied by lead time, plus a safety stock buffer based on your demand and lead-time variability. On Amazon specifically, keep your historical days of supply at or above 28 days to dodge the low-inventory-level fee, and remember that inbound shipments don’t count toward that metric until Amazon finishes receiving them.

What “Restock Level” Actually Means

A restock level, more formally called a reorder point (ROP), is the on-hand inventory quantity at which you place your next purchase order. The goal is simple: the new stock should arrive before you sell through what’s left. Get this number wrong and you either run out (lost sales, damaged rank, penalty fees) or carry too much (storage fees, tied-up cash).

The canonical formula is:

ROP = (average daily demand × lead time) + safety stock

Or in shorthand: ROP = d × LT + SS

Before you can set a restock level to avoid lost sales, you need to understand three supporting concepts.

Safety Stock

Safety stock is the extra inventory buffer above what you expect to sell during lead time. It exists because demand fluctuates and shipments arrive late. The size of this buffer depends on how much variability exists in your sales and your supply chain, plus how aggressively you want to avoid stockouts.

When both demand and lead time vary (which is nearly always the case), the standard safety stock formula is:

SS = Z × √(LT × σd² + d² × σLT²)

Where σd is the standard deviation of daily demand and σLT is the standard deviation of lead time.

Service Level and Z-Scores

The Z value in the safety stock formula represents your target service level, which is the probability of not stocking out during any given lead-time cycle. Higher service levels require disproportionately more safety stock.

Service Level Z-Score
90% 1.28
95% 1.65
97.5% 1.96
99% 2.33

A 95% service level means you expect to avoid a stockout in 19 out of 20 replenishment cycles. Jumping from 95% to 99% roughly doubles the Z-score, which dramatically increases your safety stock requirement. Choose deliberately based on what a stockout actually costs you for each SKU.

Days of Supply

Days of supply (DoS) measures how long your current inventory will last at the recent sell-through rate. It’s a simple but powerful metric: stock on hand divided by average daily units shipped. Amazon tracks its own version of this, called historical days of supply (HDOS), which factors into fee calculations covered below.

The Formulas You’ll Actually Use

Knowing how to set restock levels to avoid lost sales starts with gathering the right inputs for each SKU.

Step 1: Gather Your Data

For every SKU, pull:

  • Average daily demand (d): Use a 30 to 90-day window. Bias toward 90 days if your sales are noisy or seasonal.
  • Demand variability (σd): The standard deviation of daily demand over that same window.
  • Average lead time (LT): Total days from purchase order to inventory being available for sale. This includes supplier production, freight, customs, prep, and (for FBA sellers) Amazon receiving and transfer time.
  • Lead-time variability (σLT): The standard deviation of your lead time across recent orders.
  • Target service level: Pick a Z-score from the table above.

Step 2: Calculate Safety Stock

SS = Z × √(LT × σd² + d² × σLT²)

This combined formula accounts for uncertainty on both sides, demand that could spike and lead times that could stretch. Using only average demand multiplied by average lead time, without this buffer, is the single most common reason sellers stock out.

Step 3: Calculate Your Reorder Point

ROP = d × LT + SS

When your available inventory hits this number, place the order.

Worked Example: A Fast-Moving SKU

Suppose you sell a kitchen gadget with these numbers from the last 90 days:

  • d = 20 units/day
  • σd = 7 units/day
  • LT = 45 days (supplier through FBA available)
  • σLT = 8 days
  • Target service level: 95% (Z = 1.65)

Safety stock:
SS = 1.65 × √(45 × 49 + 400 × 64)
SS = 1.65 × √(2,205 + 25,600)
SS = 1.65 × √27,805
SS = 1.65 × 166.7
SS ≈ 275 units

Reorder point:
ROP = (20 × 45) + 275
ROP = 1,175 units

When your available inventory drops to around 1,175 units, it’s time to place the next purchase order. That 275-unit buffer gives you a 95% probability of not running out before the new shipment lands.

Amazon’s 28-Day HDOS Threshold: The Fee You Need to Dodge

Setting restock levels to avoid lost sales on Amazon involves an extra layer. Amazon’s low-inventory-level fee applies to standard-size FBA items when your historical days of supply falls below 28 days. This fee is a per-unit charge that eats into margins, and it compounds the damage you’re already taking from near-stockout conditions.

Here’s what you need to know about how Amazon calculates HDOS:

  • Both windows must be low. The fee only kicks in when your 30-day and 90-day HDOS are both below 28 days. Amazon forum moderators have confirmed this repeatedly.
  • It’s calculated at the parent-ASIN level. If you sell a product in multiple colors or sizes, Amazon rolls all child ASINs up to the parent for this calculation.
  • Reserved inventory counts. Units reserved for pending orders are included in the HDOS calculation.
  • Inbound inventory does not count. This is the critical gotcha. Shipments that are in transit or sitting at an Amazon fulfillment center waiting to be checked in do not lift your HDOS. Only units fully received and made available count.
  • Amazon introduced a 7-day/20-unit exemption in 2024 for new-to-FBA products, giving sellers a brief grace period.

The HDOS formula itself is straightforward: average daily inventory divided by average daily units shipped. You can monitor it in Amazon’s FBA Inventory dashboard.

Mapping Your ROP to the HDOS Threshold

If your calculated reorder point would regularly leave your HDOS below 28 days for standard-size items, you have two options. Either increase your safety stock (or order earlier) so your typical HDOS stays at 28 or above, or accept the per-unit fee when the carrying-cost math supports it. For most sellers, the fee costs more than a modest bump in safety stock.

A practical target: maintain an operating band of 35 to 45 days of supply. This keeps you well above the 28-day threshold even during minor demand spikes or receiving delays, while preventing excessive overstock. Reconcile this band with your ROP so purchase orders land before DoS dips toward 28.

For brands selling across Amazon and their own Shopify store, this kind of cross-channel inventory coordination is where Amazon account management and inventory planning becomes essential.

What to Do When Lead Time Slips or Demand Spikes

Even perfectly calculated restock levels sometimes aren’t enough. Suppliers miss deadlines. Amazon’s receiving process takes longer than expected. A TikTok video sends your sales through the roof. When your inventory position gets tight, you need tactical levers to prevent a full stockout.

Throttle Your Ads

When your days of supply drops below your lead time plus half your safety stock, pull back on advertising spend. This slows your sell-through rate and stretches your remaining inventory. Resume full spend once DoS recovers. Practitioners on Amazon seller forums report this as one of the most effective ways to avoid a hard stockout without completely losing momentum. For a deeper look at connecting ad spend to inventory position, see this guide on Amazon advertising profit strategies.

Nudge Your Price Up

A small price increase, even 5 to 10%, can meaningfully slow velocity during a pinch. Sellers on Reddit report using this tactic specifically to keep their HDOS above the 28-day fee threshold while waiting for inbound shipments to be received. It’s not a long-term play, but it buys time.

Activate FBM/3PL Backup

Keep a small reserve of your top SKUs at a third-party logistics provider. If your FBA inventory runs critically low or receiving lags stall your shipment, switch the listing to Merchant Fulfilled (MFN) so you stay live. Vendor guides and practitioner posts consistently recommend this as a pragmatic hedge against receiving delays. For brands running both Amazon and D2C channels, a unified D2C growth system that coordinates inventory across Shopify and Amazon can prevent these gaps from becoming emergencies.

Pull Forward Your Purchase Orders

If you see demand trending above forecast, don’t wait for inventory to hit the ROP. Order early. The cost of slightly higher carrying costs is almost always less than the cost of a stockout, especially on Amazon where going out of stock can damage your product ranking and take weeks to recover.

Gate Your Promotions

Before running a Lightning Deal, coupon, or any major promotion, simulate the demand uplift and check whether your projected post-promotion days of supply will stay above your lead time plus safety stock. If the answer is no, defer the promotion. This is where building a coordinated paid and organic search strategy pays off, because it forces you to plan promotional velocity against inventory reality.

Intermittent and Long-Tail SKUs: Different Math Required

Not every SKU sells steadily. Some products go days or weeks without a sale, then get a burst of orders. These “lumpy” demand patterns break standard ROP calculations because the average daily demand is misleadingly low, and the standard deviation relative to the mean is huge.

For these SKUs, classic safety stock math underestimates risk. A product that averages 0.3 units per day but occasionally sells 5 units in a single day will look like it needs very little safety stock, right up until you stock out during one of those bursts.

The solution is to use intermittent-demand forecasting methods like Croston’s method or the TSB (Teunter-Syntetos-Babai) approach. These methods separate the forecast into two components: the probability of a sale occurring and the expected size when it does occur. The result is a more realistic demand estimate that you can feed into your safety stock formula.

For long-tail SKUs where the cost of a lost sale is high (because the customer won’t wait and the margin is strong), use a higher service level, 97% or even 99%, to set a meaningful buffer despite the irregular pattern.

Setting Service Levels by SKU Class

A 99% service level across your entire catalog is a fast way to drown in carrying costs. A 90% service level across the board means your best sellers will regularly stock out. Neither extreme makes sense.

The better approach is ABC/XYZ segmentation. Practitioners on Reddit and operations blogs recommend setting service-level targets per class to balance stockout risk against working capital:

  • A-class / high-velocity SKUs: These drive the bulk of your revenue. Set Z at 1.65 to 1.96 (95% to 97.5%). A stockout here is expensive in both lost sales and rank damage.
  • B-class / moderate-velocity SKUs: Set Z around 1.28 to 1.65 (90% to 95%). Moderate buffer, moderate risk.
  • C-class / slow movers with smooth demand: Set Z at 1.28 or even lower (90%). The cost of carrying excess stock outweighs the occasional missed sale.
  • Volatile SKUs (X/Y/Z overlay): If demand variability is high regardless of velocity, bump the Z-score up. A volatile B-class SKU might warrant a higher service level than a smooth A-class item.

Review and recalculate monthly. Seasonality shifts which SKUs belong in which class, and a product that was C-class in January might be A-class by November.

On the flip side, carrying too much safety stock creates its own problems. Excess inventory ties up cash and can lead to long-term storage fees. For guidance on the other side of this balance, see this piece on managing aged inventory and reducing long-term storage fees.

Common Mistakes That Cause Lost Sales

Even sellers who know the formulas make predictable errors when setting restock levels. Avoiding these is half the battle.

Treating lead time as a single number. Your supplier says 30 days. Freight takes 14. But Amazon receiving can add anywhere from a few days to three weeks. Practitioners on Reddit’s FBA communities regularly report shipments that are “delivered” to the fulfillment center but sit in limbo for weeks before units become available. If you don’t account for this variability in σLT, your safety stock will be too thin.

Using one service level for everything. Applying a flat 95% across hundreds of SKUs wastes money on slow movers and under-protects your best sellers. Segment and set targets per class.

Counting inbound shipments as “safe.” This is the mistake Amazon sellers make most often. You see a shipment marked as delivered and assume your HDOS will improve. It won’t, not until those units are fully received and available. Plan your reorder timing as if that inbound shipment doesn’t exist until it shows up in your available inventory count.

Ignoring parent-ASIN math. Amazon calculates HDOS at the parent level. A well-stocked variant can mask the fact that another variant is nearly out. This creates a false sense of security. Set variant-level reorder points even though HDOS is aggregated to the parent, because one depleted variant still means lost sales for customers who want that specific option.

Disconnecting ads from inventory position. Running aggressive PPC campaigns while sitting on 15 days of supply is a recipe for a stockout that craters your ranking. Operators who avoid rank crashes pace their ad spend to their inventory position. Going dark on ads entirely is also a mistake, because you lose keyword momentum. The key is to throttle, not stop.

Operator-Grade Checklist: How to Set Restock Levels to Avoid Lost Sales

Run through this monthly for every active SKU:

1. Refresh your inputs.
Pull d, σd (from 30 and 90-day windows), LT, and σLT from your sales and order history.

2. Assign a service level by SKU class.
A-class gets Z = 1.65+. C-class gets Z = 1.28. Volatile items get bumped up.

3. Recalculate safety stock and ROP.
SS = Z × √(LT × σd² + d² × σLT²). Then ROP = d × LT + SS.

4. Reconcile with Amazon’s HDOS.
Check that your ROP will keep HDOS at or above 28 days for standard-size FBA items. Adjust upward if needed, or budget for the fee.

5. Set alerts.
Trigger notifications when stock drops to ROP and when DoS falls below LT + SS.

6. Build contingency rules.
If DoS drops below LT + half your SS: throttle ads, consider a price nudge, pull forward the next PO, or switch to FBM.

7. Gate promotions.
Before any deal or coupon, simulate the demand uplift and confirm post-promo DoS stays above LT + SS.

8. Review and adjust.
Seasonality, new competitors, and supply chain changes all shift these numbers. Monthly recalculation isn’t optional.

Frequently Asked Questions

What is a restock level and how is it different from safety stock?

A restock level (reorder point) is the total inventory threshold that triggers a new purchase order. Safety stock is one component of that threshold, specifically the buffer above expected demand during lead time. The reorder point equals your expected demand during lead time plus safety stock. Safety stock alone wouldn’t tell you when to order, it just defines how much cushion to carry.

How do I set restock levels to avoid lost sales if my lead times are unpredictable?

Unpredictable lead times mean a higher σLT value, which increases your safety stock requirement. Track your actual lead times across the last several orders, calculate the standard deviation, and plug it into the safety stock formula. If receiving delays at Amazon are the main source of variability, consider keeping backup FBM inventory at a 3PL so you can stay live while FBA catches up.

Does Amazon’s low-inventory-level fee apply to all products?

No. As of 2024, the fee applies to standard-size FBA items when both the 30-day and 90-day historical days of supply are below 28 days. It’s calculated at the parent-ASIN level. New-to-FBA products get a brief exemption (7 days or 20 units shipped). Check Amazon’s FBA Inventory dashboard for your current HDOS.

Should I use the same service level for every SKU?

No. A flat service level across your catalog either over-invests in slow movers or under-protects your bestsellers. Use ABC segmentation: higher service levels (95%+) for high-velocity, high-margin products, and lower levels (90%) for slow, steady items. Adjust for demand volatility as well, not just volume.

Why does my HDOS stay low even after I ship inventory to Amazon?

Because inbound inventory that hasn’t been fully received and transferred to “available” status does not count toward your HDOS calculation. Only units that Amazon has processed and made sellable affect the metric. Receiving delays are common, so always plan as though your inbound shipment won’t improve HDOS for days or even weeks after delivery.

How often should I recalculate my restock levels?

Monthly is the minimum. If you’re in a seasonal category or experiencing rapid growth, biweekly recalculations are better. Demand patterns, lead times, and supplier reliability all shift over time. A reorder point calculated in January may be dangerously low by March if spring demand picks up.

Can I just use Amazon’s restock recommendations instead of calculating my own ROP?

Amazon’s restock recommendations are a useful starting point, but they don’t account for your specific lead-time variability, supplier constraints, or margin targets. They also won’t coordinate with your ad spend or promotional calendar. Calculating your own ROP gives you control over the service level and lets you align inventory decisions with profitability.

What’s the best way to handle setting restock levels for products that sell on both Amazon and Shopify?

Track demand and lead time separately for each channel, but plan inventory holistically. Your total safety stock should account for combined demand variability. Many sellers keep the bulk of inventory at a 3PL that can ship to both FBA and fulfill D2C orders, which gives flexibility to shift allocation as demand changes across channels.


Setting restock levels to avoid lost sales is a math problem, but it’s also a coordination problem. The formula gets you the number. Connecting that number to your ad spend, your promotions calendar, your receiving timelines, and your fee thresholds is what actually prevents stockouts in practice.

If you want a team to audit your current inventory position, reorder points, and ad-to-inventory alignment, request a free eCommerce brand audit that includes a 90-day action plan. Or if you’re ready to talk specifics, reach out to an Amazon and D2C growth specialist directly.