A boutique owner in St. Augustine sells handmade and curated home goods. Candles, ceramics, textiles, small furniture. She has one location and an online store that accounts for about 30% of revenue.
Her inventory problem was not complicated. She had 340 SKUs. She knew her best sellers. She knew her seasonal patterns. She even had a spreadsheet that tracked sales velocity and estimated reorder dates.
The spreadsheet was always a day late.
Not because the spreadsheet was wrong. Because updating it was a manual process that happened on Mondays, and her best sellers would stock out on Thursdays or Fridays. By the time she placed the reorder on Monday, she had lost a weekend of sales on products she knew people wanted to buy.
She estimated the stockout losses at $4,000 per month based on the gap between when products stocked out and when they were available again. For a boutique doing $25,000 to $35,000 per month, that is 12 to 16% of revenue lost to a timing problem.
What we built
One tool with three functions: monitor, alert, and reorder.
Monitor: The tool connects to her point-of-sale system and her online store. It tracks daily sales by SKU and calculates a rolling 14-day sales velocity for each product. It knows that the lavender soy candle sells 2.3 units per day and the ceramic planter sells 0.8 units per day.
Alert: When current inventory for any SKU will run out within the lead time for that supplier plus a 2-day safety buffer, the tool sends an alert. If the candle supplier delivers in 5 days and she has 14 candles left at 2.3 per day, that is 6 days of stock. The tool alerts when she crosses the 7-day threshold (5 days lead time plus 2 days buffer).
Reorder: The alert includes a suggested reorder quantity based on the sales velocity and her preferred stock level (she likes to keep 3 weeks of inventory for best sellers, 6 weeks for slow movers). She reviews the suggestion, adjusts if she wants to, and the tool generates the purchase order in the format each supplier expects.
The seasonal adjustment
The basic velocity calculation works for 10 months of the year. It breaks in November and December, when her sales volume doubles and the product mix shifts toward gift items.
We added a seasonal modifier. The tool looks at the same period from the previous year (she had 14 months of POS data) and applies a multiplier to the velocity calculation. If lavender candles sold at 2.3 per day in October but 4.8 per day in November last year, the tool applies approximately a 2x multiplier starting in the last week of October.
This is not machine learning. This is a lookup table with a ratio. The owner could have done this math herself. The difference is the tool does it for all 340 SKUs simultaneously and adjusts daily as actual November sales data comes in.
What happened in the first month
She placed 23 reorders in the first month. Before the tool, she was placing about 15 reorders per month on Mondays. The 8 additional reorders were not over-ordering — they were reorders that would have been placed a week late under the old system.
Zero stockouts on her top 20 products for the entire month. The previous three months averaged 4 to 6 stockout events per month on top-20 products.
Her estimate of $4,000 in monthly stockout losses turned out to be conservative. Actual sales increase in the first full month was $5,200, which includes both recovered stockout sales and the natural sales increase from having popular products consistently available. Customers who find what they want are more likely to buy additional items.
The supplier relationship effect
An unexpected benefit: her suppliers started treating her differently. When you order consistently and predictably, suppliers prioritize your orders. Two of her five main suppliers moved her to preferred status with shorter lead times, which further reduced her stockout risk.
One supplier told her directly: "You are the only retail customer I have who orders before they run out instead of after." That is not a technology compliment. That is a process compliment. The tool just made the process executable.
What this cost
Three days of build time. The tool connects to her Shopify store and Square POS system through their standard APIs. No custom hardware. No monthly SaaS subscription. She owns the code.
The ongoing cost is the compute for running the daily calculations, which is negligible. The ROI was positive in the first week.
The inventory insight she did not expect
After two months of data, the tool surfaced something she had not seen in her spreadsheet: 47 of her 340 SKUs had not sold a single unit in 60 days. She knew she had slow movers, but she did not realize 14% of her inventory was dead stock.
She ran a clearance event on the dead stock, recovered $3,800 in capital, and reallocated that shelf space and cash to her top performers. The tool did not recommend this — it just made the data visible in a way the weekly spreadsheet never did.
What this means for your retail business
If you have a POS system and an online store, you have the data. If you have supplier lead times, you have the parameters. The tool we built is not sophisticated. It is basic math — velocity times lead time equals reorder point — applied consistently across your entire inventory every single day instead of once a week.
The value is not in the math. The value is in the timing. A reorder placed 3 days earlier catches the weekend sales. A reorder placed 3 days late loses them. Over 12 months, that timing difference is worth more than most retailers spend on marketing.
Get posts like this in your inbox
No spam. New articles on AI strategy, governance, and building with AI for small business.