Building block thinking: How good metrics multiply

When your base metrics are well-defined and consistently built, calculated metrics are almost automatic. The work is upfront — and it compounds. Every well-named dimension, every consistent label, every clean definition you write today quietly enables a dozen future metrics you haven't even thought of yet.

That's what building block thinking is all about. If you ever hear your data colleagues use the word composability, this is exactly what they mean. It's a technical term for a simple idea: build your metrics in small, reliable pieces, and snap them together to make something bigger. Like Lego bricks. The power isn't in any single brick — it's in the fact that every brick is designed to connect.

Two types of metrics

Before we get into how metrics combine, it helps to understand the two fundamental types.

Base metrics are direct measurements of something that happened. They're the raw counts and sums we introduced in Article 2 — the source of truth for your business.

Base metrics answer the question: How much?

Calculated metrics are built from base metrics using arithmetic — division, multiplication, addition, or subtraction — your BEDMAS stuff from high school. They take the raw counts and sums you already have and transform them into something more revealing.

Calculated metrics answer the question: How well?

For effective analysis, you need both. Base metrics give you scale — they tell you whether your business is large or small, growing or shrinking. Calculated metrics give you proportion — they tell you whether your business is healthy, efficient, and improving. Together, they present a full picture of your business.

🛒 Grocery Store Example

Revenue tells you how much money came in. Revenue per Customer tells you which customers are spending more or less.

Units Sold tells you how much fruit moved. Sell-Through Rate tells you whether you're ordering the right amount.

The base metrics establish the facts (how much). The calculated metrics reveal what's working (how well).

Composability: The professional term for building block thinking

Your data team might describe a well-built metric library as composable. It's worth understanding what that means — because it captures something important about why this approach is worth the upfront effort.

A composable system is one where complex things are built from simpler, verified pieces — and where changing a piece at the bottom automatically updates everything above it.

In the context of metrics, composability delivers four specific advantages:

1. Consistency. If you update the definition of your base Revenue metric — say, you decide to exclude refunds — every calculated metric built from it updates automatically. Because you have one source of truth (your base metric), Gross Margin, Revenue per Customer, and Average Unit Price, all reflect the new definition without anyone touching them individually.
2. Auditability. When a stakeholder asks "Where does this number come from?", a composable metric has a clear, traceable answer: Sell-Through Rate = Units Sold ÷ Units in Stock. You can follow the chain all the way back to the raw data. That kind of transparency builds trust — and it makes errors much easier to find and fix.
3. Efficiency. Without shared base metrics, different teams often end up writing their own versions of the same formula — slightly differently each time. The marketing team calculates Revenue one way, finance calculates it another, and nobody agrees on the number. Composability encourages an agreed-upon starting point. No more starting from scratch or second-guessing.
4. Accuracy. A composable “building block” approach guarantees the math is always right — because the calculation is always built from the summed base metrics, never from pre-aggregated rates. This advantage is subtle but important. Rates and ratios should never be stored as base metrics, because you can't safely combine them later. Suppose Toronto has a Sell-Through Rate of 80% and Vancouver has a rate of 60%. If you average those two percentages, you get 70% — but that's only correct if both stores stocked the same number of Units. If Toronto stocked 1,000 units and Vancouver stocked 100, the real blended rate is much closer to 80%. The only way to get the right answer is to go back to the base numbers: sum all Units Sold, sum all Units in Stock, then divide.

In a professional data conversation, you might hear: "We should build composable metrics so our KPIs are always based on the same verified base measures." Now you know exactly what that means — and why it matters.

Why ratios and rates are so valuable

We touched on this in Article 1. It's worth going deeper here.

The core limitation of a base metric is that it carries no context about scale. A $1 million profit sounds impressive — until you learn that one company earned it on $10 million in revenue (a 10% margin) and another earned it on $100 million in revenue (a 1% margin). The absolute number is the same. The story is completely different.

Ratios and rates solve this by expressing one number in relation to another. That relationship is what creates comparability — across time, across teams, across locations, even across industries.

A few properties that make calculated metrics especially powerful:

They're not size-dependent. Sell-Through Rate is just as meaningful for a small corner store as for a national chain, because it's always expressed as a proportion. You don't need to adjust for scale.

They expose efficiency. Inventory Turnover doesn't just tell you how much sold — it tells you how effectively you're using what you have. Revenue per Square Foot doesn't only tell you total sales — it tells you how hard each part of your store is working.

They're better for trend analysis. A 10% increase in Revenue could mean you got bigger. A 10% improvement in Gross Margin means you got more efficient. Trends in ratios tend to be more meaningful signals than trends in raw numbers.

That said, ratios do have a weakness. They can be misleading when the underlying numbers are very small. As we noted in Article 1, How to choose the right metrics, a 100% increase in sales means very little if it represents one extra banana sold! Always consider the base numbers behind the ratios.

The compatibility principle

Here's the insight that will change how you think about building metrics.

In PowerMetrics, every metric shares the same fundamental structure: a measure, a date/time dimension, and (optionally) categorical dimensions. This common structure is what makes metrics compatible with each other — and compatibility is what makes calculated metrics possible.

When you create a calculated metric from two base metrics, the dimensions carry through.

🛒 Grocery Store Example

Units Sold has the dimensions: City, Product, Store Type.
Units in Stock has the same dimensions: City, Product, Store Type.

When you divide one base metric by the other to create Sell-Through Rate, the resulting calculated metric automatically inherits those dimensions.

You can now filter Sell-Through Rate by City, by Product, and by Store Type — exactly the same way you could with the base metrics.

Toronto vs. Vancouver? ✓ Still works. Apples vs. Oranges? ✓ Still works. In-Store vs. Warehouse? ✓ Still works.

This is the payoff of building block thinking. You define the dimensions once on the base metrics, and every calculated metric built from them automatically inherits that capability.

Where it breaks: Dimension inconsistency

Now for the cautionary side.

Dimension compatibility only works when the dimension names match. If two metrics measure the same concept but use different names for it, the calculated metric won’t be able to connect them, and you’ll lose the ability to slice and filter.

🛒 Grocery Store Example

Suppose Units Sold uses a dimension called City and Units in Stock uses a dimension called Location_City.

When you try to build a Sell-Through Rate calculated metric using these two metrics, the system sees two different dimensions — City and Location_City. The resulting metric can't be filtered by city because differently-named dimensions can’t act as a shared lens.

The data exists, but the connection is broken. The fix — renaming dimensions retroactively across multiple metrics — is far more painful than getting it right the first time.

This is data debt in action. It accumulates quietly, one small inconsistency at a time, and it costs you later. Learn more about best-practices when working with metrics and dimensions.

Practical tips for compatibility

As we’ve learned, building block thinking only works if the blocks are compatible. Here are some practical tips to keep your metrics aligned and avoid data debt:

1. Align Dimension Names Early-on

Choose one name for each dimension and use it everywhere. Not City in one metric and Location in another. Not Product here and Product_Name there.

In PowerMetrics, when you connect a data source to a metric, you can rename the dimension fields before saving. This is the ideal time to align your naming. If your raw data calls it Location_City but your standard dimension name is City, rename it now — not after you've built twenty metrics on top of it.

The aligned dimension name should reflect business language, not database language. Your team will thank you.

2. Use Consistent Names for Dimension Members too

A dimension member is one of the values within a dimension. City is the dimension. Toronto, Vancouver, and Montreal are its members.

With consistent naming, these values will align across every metric that uses that dimension. The terms Toronto, toronto, TOR, and Toronto, ON may mean the same thing to a person but to a computer, they represent four different values. If one metric says Toronto and another says toronto, a calculated metric that combines the two won’t be able to match them up.

The fix is simple: Standardize member values in your source data, or clean them up when defining your metrics in PowerMetrics. Use consistent capitalization, abbreviation rules, and remove trailing spaces. A little bit of discipline for a big payoff.

Level up data-driven decision making

Make metric analysis easy for everyone.

Get Started Now

3. Name Your Metrics for Reuse

A base metric named Rev might make sense to the person who built it. It will likely make much less sense six months later, to a different team member, who isn't sure if it means total revenue, net revenue, or something else entirely.

Name your base metrics as if they'll be used by someone who isn’t in the room when you build them — because, eventually, they will be. Revenue (Gross), Units Sold, Inventory Levels are clear. Rev_v2, Inv_Curr, US_final are not.

4. Use Tags to Organize Your Metric Library

As your metric library grows, tags help you find and group related metrics quickly. In PowerMetrics, you can tag metrics by team (Marketing, Operations, Finance), by data source (Hubspot, CSV, DataWarehouse), by topic (Inventory, Sales, Customer), or by any combination.

Think of tags less as filing and more as filtering. A well-tagged library means everyone on your team can find the base metrics they need, making building calculated metrics faster and less error-prone.

The Compounding Effect

Here's the thing about building block thinking: the value isn't linear. It compounds.

Every base metric you define well doesn't just create one useful number. It creates the potential to combine with other well-defined metrics to answer questions you haven't asked yet.

🛒 Grocery Store Example

You create Revenue, Units Sold, Customers, Cost of Goods Sold, and Inventory Levels base metrics — each with consistently-named dimensions (Product, City, Store Type).

From those five metrics, you can immediately build:

• Sell-Through Rate (Units Sold ÷ Inventory Levels)
• Revenue per Customer (Revenue ÷ Customers)
• Average Unit Price (Revenue ÷ Units Sold)
• Gross Margin (Revenue − Cost of Goods Sold ÷ Revenue)
• Inventory Turnover (Units Sold ÷ Average Inventory Levels)

And every one of these calculated metrics can be filtered by Product, by City, and by Store Type — without any additional setup — because the dimensions were defined consistently on the base metrics from the start.

Five base metrics. Five calculated metrics. All fully dimensional. Built once, usable forever.

That's the compounding effect of building block thinking. The discipline you apply at the beginning quietly multiplies everything you build afterward.

Quick recap

• Two types of metrics: base metrics answer how much? — calculated metrics answer how well?
• Base metrics are the foundation — direct counts and sums that establish the facts of your business.
• Calculated metrics are built from base metrics using division, multiplication, addition, or subtraction — revealing efficiency, proportion, and performance.
• Composability is the technical term for building block thinking — building complex metrics from verified, reusable base metrics. It delivers consistency (changes flow upward automatically), auditability (every number is traceable), efficiency (one formula, used everywhere), and accuracy (gained by always calculating rates from summed base metrics, never by averaging pre-aggregated rates).
• The compatibility principle: in PowerMetrics, metrics share a common structure. When two metrics share the same dimension names, a calculated metric built from them automatically inherits those dimensions.
• Dimension alignment is everything. Use the same name for the same concept across every metric. Rename fields during setup in PowerMetrics — not after.
• Member naming matters. The system interprets Toronto and toronto as two different things. Standardize early and always.
• Name metrics for reuse. Clear, plain-language names mean anyone can build on your work.
• Use tags to keep your metric library organized and findable as it grows.
• The work is upfront, and it compounds. Every well-defined base metric multiplies your ability to build calculated metrics quickly and reliably.

What's Next

Articles 1 through 3 covered how to choose metrics, how to define them, and how to build them into a consistent, compatible library. Article 4 takes this one step further.

Graph thinking is about adding a layer of context, relationships, and metadata that makes your metric library not just useful to your team today — but understandable to any analyst, any new team member, and any AI tool you bring in tomorrow. If composability is about making metrics reliable and reusable, graph thinking is about making them intelligent.

Continue Learning

Measures, metrics, and KPIs: Read up on definitions, practical examples, and a simple workflow to understand and apply these terms successfully.

Why metrics: Additional resources on why a metric-centric approach is the modern way to turn raw data into a strategic asset for your entire organization.

PowerMetrics Concepts Series · Article 3 of 4