SKU Profitability: Why Your Dashboards Hide Net Margin

---

Lots of data, but no clarity on margin

It's the Monday morning trading meeting. The Head of Marketing presents a deck showing a very good Return on Ad Spend (ROAS) for a new campaign. The Head of Operations shows that fulfilment costs are stable. Yet, when the finance report lands a week later, the overall net margin is down. Everyone looks at each other, and the arguments about whose data is right begin.

As COO, you’re stuck in the middle. You have access to hundreds of reports in Looker or Tableau. You have dashboards for everything from website traffic to warehouse pick-rates. But the one number you really need, the true, fully-loaded net margin of a single SKU, seems impossible to find.

This isn't a problem with your people. Your team is smart and your analysts are busy. The issue is that the business has scaled faster than its data foundations. It's a common situation, a kind of Dashboard Sprawl: Curing the Chaos of 500 Reports, where producing more reports feels like progress, but isn't.

The underlying problem: how blended costs hide the truth

The reason your dashboards can be misleading is that they are built on averages and blended costs. You have probably invested in a modern data stack, with tools like Snowflake, dbt, and Fivetran, but this has often just moved the same problems to the cloud. You are producing the same confusing numbers, only faster.

The technical issue usually comes down to a lack of detail. The way your data is structured often reflects the silos in the business:

Marketing Costs: Ad spend from Google and Facebook is rarely allocated accurately to the specific SKUs in a customer's basket.

Fulfilment Costs: Pick-and-pack fees, packaging, and shipping are often applied as a 'blended average' across all orders, hiding the fact that a large, fragile item costs far more to ship than a small, durable one.

Transaction Fees: Payment processing fees from Stripe or Shopify are often treated as a general overhead, not tied to the specific order.

When you can't connect these specific costs to a single SKU, you can't get a true picture of SKU Profitability: Architecting True Margin Visibility. Instead, your analysts often find themselves creating endless variations of reports, hoping one of them will uncover the real story. It's a search that rarely succeeds.

In my experience, this is a common pattern in fast-growing e-commerce companies. I worked with one recently where we reviewed their Looker setup and ended up removing over 200 dashboards that weren't being used or trusted. We replaced them with a much smaller set of about 30 reports, all built on a single data model that used an agreed method for allocating costs. The goal isn't more data, it's more trust in the data you have.

How to fix the problem: from averages to unit economics

You don't fix this by buying a new business intelligence tool or hiring more analysts to build more reports. You fix it by rebuilding the engine underneath. A modern COO Data Strategy: Architecting Operational Visibility really depends on getting this right.

The work is more about process and structure than it is about technology:

Pause building new reports: The first step is often to simply stop building new reports. The existing ones aren't trusted, and adding more just adds to the noise.

Map out the real-world costs: This means sitting down with Operations, Marketing, and Finance to map every single variable cost that goes into selling one unit. It's a process of working backwards from the P&L and getting agreement on how each cost should be allocated. It's the unglamorous, difficult work that many teams prefer to avoid.

Build the logic into your data model: Once everyone agrees, these rules are written into the data transformation layer, usually in a tool like dbt. We create a single, definite meaning for 'Net Margin'. This ensures that whenever someone in the business looks at that metric, it means the exact same thing. This is the foundation for a Single Source of Truth: Architecting Trust in Data.

Only after this foundational work is complete do we build the small number of visualisations required to track it. We move from manufacturing dashboards to curating signals.

The difficult part: getting agreement

To be direct, this part can be uncomfortable. It might mean explaining to the Head of Marketing that their platform-reported ROAS isn't the full picture, because it ignores contribution margin. It could mean showing the Head of Operations that a shipping strategy that looks efficient is actually losing money on certain types of product.

Creating a single source of truth is as much a political exercise as a technical one. It needs someone senior to be able to say, 'This is how we will measure profitability from now on, and all other definitions are retired'. It might feel like you are moving slower for a few weeks, but it lets you move much faster for years to come.

The result: clarity on what is actually profitable

The result of this foundational work is that the 'data debates' tend to fade away. The weekly trading meeting stops being an argument about whose numbers are right. It becomes a strategic discussion based on a shared, trusted view of the business.

You can finally see which products are contributing to growth and which are a drain on resources. You can allocate marketing spend with more confidence, adjust your fulfilment strategy, and make inventory decisions that help build a sustainable, profitable business. You can get a defensible view of Net Margin: Architecting Profitable Unit Economics and run the business with the clarity you've been looking for.