Most merchants check their store health the same way: glance at revenue, skim the top sellers, maybe look at traffic. If the numbers are up, things feel fine. If they're down, it's time to worry.
The problem is that a store can look healthy on the surface while quietly bleeding money underneath. Revenue might be climbing, but margins might be shrinking. Orders might be strong, but your best customers might be churning. Inventory might be moving, but half of it might be sitting dead in a warehouse.
That's why we built the 12-layer store diagnostic in Spark by MishiPay. It's a comprehensive ecommerce health check that analyzes your store across twelve distinct dimensions in under 60 seconds. Not a dashboard you have to interpret. Not a spreadsheet you have to build. A structured audit that tells you exactly what's working, what's broken, and what to do about it.
Every layer maps to a lever in the core ecommerce equation:
Revenue = Traffic × Conversion × AOV × Margin × Repeat
The diagnostic doesn't just measure each lever — it calculates the dollar impact of moving it. Here's what each layer examines and the kind of findings it surfaces.
Layer 1: Revenue Health
The diagnostic starts where most merchants start — revenue — but goes deeper than the top line. It examines revenue trends across 30, 60, and 90-day windows, identifies acceleration or deceleration patterns, and flags anomalies.
Example finding: "Revenue grew 8% month-over-month, but the growth is concentrated in 3 SKUs. Excluding those, revenue was flat. Dependency risk is elevated."
Layer 2: Margin Analysis
Revenue means nothing without margin context. This layer applies the contribution margin formula to every order in your store:
CM$ = Revenue − COGS − Shipping − Transaction Fees − Discounts
It then aggregates by product, category, and time period to identify where margin is growing, shrinking, or hiding. When COGS data isn't available, the diagnostic uses configurable assumptions (default: 40%) and flags items that need cost data for accuracy.
Example finding: "Overall margin is 42%, but your top-selling category (accessories) runs at 18% margin. It accounts for 34% of orders but only 11% of gross profit. Your home goods category at 61% margin is undermarketed."
Layer 3: Discount Effectiveness
Discounts are the most abused lever in ecommerce. This layer evaluates whether your promotions drive incremental revenue or simply cannibalize full-price sales. It measures discount depth, frequency, and the percentage of revenue sold at a discount.
Example finding: "67% of orders in the last 30 days used a discount code. Average discount depth was 22%. Only 31% of discounted orders came from new customers — the rest were repeat buyers who likely would have purchased at full price."
Layer 4: Customer Lifetime Value
LTV analysis segments your customers by acquisition cohort, channel, and first product purchased. It uses the formula:
LTV = Average Order Value × Purchase Frequency × Average Customer Lifespan
Each variable is calculated from your actual order data, not industry benchmarks. The diagnostic then compares LTV against customer acquisition cost (where available) to calculate break-even ROAS thresholds — telling you exactly how much you can afford to spend to acquire each type of customer.
Example finding: "Customers acquired via organic search have an average LTV of $284 over 18 months. Customers acquired via paid social have an LTV of $97. Your current ad spend allocates 70% of budget to paid social."
Layer 5: Retention Cohorts
Customer acquisition gets all the attention. Retention is where the money is. This layer builds monthly cohorts and tracks repeat purchase rates over time. It identifies when customers typically churn and which cohorts are outperforming or underperforming.
Example finding: "Your January cohort had a 24% repeat purchase rate within 60 days — down from 33% for the same period last year. Customers who bought from your winter collection were 40% less likely to return than those who bought core products."
Layer 6: Inventory Efficiency
Every dollar in unsold inventory is a dollar not working for you. This layer calculates sell-through rates, days of supply, dead stock ratios, and identifies products at risk of stockout or overstock.
Example finding: "14% of your inventory (by cost value) hasn't sold a single unit in 90+ days — that's $23,400 in dead capital. Meanwhile, 6 SKUs are projected to stock out within 12 days at current velocity."
Layer 7: Product Performance
Beyond simple "top sellers" lists, this layer ranks products by a composite score that factors in revenue, margin, velocity, return rate, and customer satisfaction. It identifies hidden winners and overrated performers.
Example finding: "Your #3 product by revenue ranks #1 by composite score — high margin (54%), low return rate (2.1%), and strong repeat purchase correlation. Your #1 product by revenue ranks #11 due to a 14% return rate and 19% margin."
Layer 8: Basket Analysis
This layer examines co-purchase patterns — which products are frequently bought together, what drives larger basket sizes, and where cross-sell opportunities exist. It uses association rules to surface non-obvious product affinities.
Example finding: "Customers who buy Product A are 4.2x more likely to also buy Product D (not Product B, which you currently recommend). Average basket size is $67 but jumps to $94 when the order includes items from two or more categories."
Layer 9: Refund Patterns
Returns eat margins quietly. This layer tracks refund rates by product, category, customer segment, and time period. It identifies products with abnormal return rates and estimates the true margin impact after returns.
Example finding: "Your overall refund rate is 6.8%, but SKU #4471 has a 28% refund rate — the most common reason is 'not as described.' This single product has cost $4,200 in refund-related losses over the past quarter, including shipping costs."
Layer 10: Channel Performance
If you sell across multiple channels — your main site, marketplaces, social commerce, in-store — this layer compares performance across all of them. It normalizes for fees, shipping costs, and return rates to show true profitability by channel.
Example finding: "Your marketplace channel generates 22% of revenue but only 9% of profit after marketplace fees (15%) and a higher return rate (11% vs. 5% on your direct site). Direct site traffic converts at 3.4% vs. 1.9% on marketplace."
Layer 11: Seasonal Trends
This layer compares current performance against the same periods in prior years, identifies seasonal patterns, and flags when current trends deviate from expected seasonality. It helps you plan inventory, staffing, and promotions around predictable cycles.
Example finding: "Based on two years of data, your store typically sees a 35% revenue increase in weeks 10-14. This year, week 10 was only 12% above baseline. If the pattern holds, you may be overstocked for an expected Q2 surge that isn't materializing."
Layer 12: Vendor and Supplier Concentration
The final layer assesses supply chain risk. It measures how dependent your revenue and inventory are on individual vendors, flags single-source products, and identifies concentration risks.
Example finding: "73% of your COGS flows to two suppliers. Your top supplier provides 41 of your 120 active SKUs, including 7 of your top 10 sellers. A disruption from this single vendor would impact an estimated 48% of monthly revenue."
Why twelve layers — and why they run together
Any one of these layers, checked in isolation, tells an incomplete story. A product might look great by revenue but terrible by return rate. A channel might look profitable until you factor in its customer LTV. A discount might seem effective until you see it's training your best customers to wait for sales.
The value of the store diagnostic isn't just breadth — it's the cross-referencing. When all twelve layers run against the same dataset in the same analysis, the AI can connect findings across dimensions:
- "Your margin is dropping because your highest-velocity products are your lowest-margin ones, and your discount strategy is accelerating that trend."
- "Your retention is declining because your recent acquisition campaigns are bringing in one-time bargain hunters, not because your product quality dropped."
These are the kinds of insights that take hours to piece together manually. The store diagnostic surfaces them in under 60 seconds.
It works across all connected platforms
Whether your store runs on Shopify, WooCommerce, Magento, Odoo, or Square, the diagnostic pulls from the same normalized data pipeline. You don't need to export anything, configure any reports, or map any fields. Connect your store, run the diagnostic, and get a structured ecommerce audit that covers every angle.
If you sell across multiple platforms, the diagnostic runs across all of them simultaneously — giving you a unified view that no single platform's native analytics can provide.
A real diagnostic in action
Here's what a diagnostic looks like for a specialty coffee store doing ~$48K/month on Shopify. The merchant typed: "Run a full store diagnostic."
Within 60 seconds, Spark returned findings across all 12 layers. The three most actionable:
- Margin alert: The store's bestseller by revenue (Ethiopian Single Origin) had a 19% margin after discounts — ranking it 11th out of 15 products by profitability. Meanwhile, the Cold Brew Concentrate at 54% margin was barely promoted.
- Discount dependency: 67% of orders used a discount code in the last 30 days. Only 31% of discounted orders came from new customers. The store was training repeat buyers to wait for sales.
- Inventory risk: 14% of inventory by cost value ($23,400) hadn't sold in 90+ days, while 6 SKUs were projected to stock out within 12 days.
None of these findings were visible in Shopify's built-in reports. Together, they pointed to roughly $6,000/month in recoverable margin — more than 100x the cost of the tool.
Try it on your store
The 12-layer store diagnostic is available to all Spark by MishiPay users. Connect your store, type "run a store diagnostic," and see what your data reveals when you look at all twelve layers at once. Most merchants find at least three or four findings they had no idea about — and at least one that pays for itself immediately.
Run your free 12-layer diagnostic
Connect your store in 60 seconds and get a comprehensive audit across margins, retention, inventory, and 9 more dimensions.