Metrics Finish the Job Semantic Layers Started
Graham WattsPublished 2026-02-05Summary: Businesses thrive or fail based on the decisions that their leaders make. So it could be said that the secret to running a successful business is in making the right decisions. So how can you ensure that the decision makers in your business always make the right decisions? In the end, it comes down to having the right tools, training and experience.
The common practice is to try and equip the decision makers in the business with data and business intelligence software to try and ensure they have the tools they need. Training is provided on the software and over time experience is achieved, hopefully leading to better decisions. In principle this makes sense, but the problem is that data is not a tool for decision making, at least not in the way it is typically collected, stored and arranged. Data is a resource, and only becomes a tool when it is cleaned, understood and transformed so it can be applied to a specific business problem where a decision can be made.
Many businesses struggle with data as a tool for business decisions, simply because the state their data is in means they aren’t really working with decision making tools at all.
Semantic Layers
A good first step in preparing data to be used for decision making is to ensure that it is well cleaned, organized and described. Data models, and more recently semantic layers are technologies that enable just that, by adding metadata on top of the data describing what the data is and how it should be queried. Furthermore models and semantic layers often provide simplified query languages that can make use of the extra metadata simplifying and controlling the access to the data.
In this way semantic layers provide much the same benefit as cleaning up and organizing a messy garage. Messy garages are full of treasures and interesting things, but no one really knows where anything is, or what you will find without significant work to go searching through the mess. Data without a semantic layer is much the same, you have to roll up your sleeves and go digging through the data, and when you do you won’t know what you’ll find. However, if you took the time to organize the garage, label everything, lock away the dangerous stuff, and provide notes and instructions on what things are, well then you can find things, and you have a pretty good sense of what you have.
Organizing and describing your data is a good first step to making use of it for decision making, and as such semantic layers are a key component in many modern data stacks.
Data as a Business Tool
So now that your data is organized, and you know where to find everything, how do you know you have what you need to make good business decisions. To do that you need to first understand the types of decisions you need to make, then and only then, can you identify the data needed to support those decisions. This combination of business purpose, understanding and carefully selected data is what turns data into a tool for decision making.
Going back to our analogy, when you have a job to do around the house, and need a tool to accomplish it, one often goes searching through the garage to find the right tool. If you have carefully organized your garage and described everything then if you have the right tool you’ll know where to find it. Quite often though, with household repairs, you are missing tools you never had before so you’ll have to go to the hardware store to get them. Over time you’ll collect the common tools that you need, but each job takes more time than it should, as you need to go and get the tools you don’t have and then learn how to use them.
An experienced homeowner may go one step further and setup their garage as a workshop. They will take inventory of the tools that they will probably need, buy them, learn how to use them and organize them appropriately so they have a complete workshop. When a job needs to be done, a homeowner can immediately get the right tool, use it and return it to where it belongs, turning a whole day repair job into an hour or two.
Business decisions are quite similar. If you have identified the data you need, obtained it, organized it by purpose, and made sure decision makers know how to use and to solve what problems, then business decisions can be done quickly and efficiently, rather than requiring a long time to find and prepare the data you need. In business, decision making speed and confidence nearly always win the day.
Metrics Are Data Tools
This combination of purpose, shared understanding and applicability to business problems is what separates a collection of well defined metrics from a semantic layer or a data model. The first step in setting up your metrics is identifying the business decisions you need to make regularly or may need to make. Then you identify the data you need and make sure you have it available. Clear descriptions for the metrics and sometimes training is essential to ensure everyone knows what the data means, and more importantly, what it is used for and how it is to be used. Metrics allow you to create a well thought out workshop for solving your business problems with data.
Technology Can Help
There are metric layers and metric platforms designed around the principle of creating a set of business metrics and sharing them with the decision makers in your organization. Products such as Klipfolio PowerMetrics are built around the concept of defining metrics from a business standpoint and connecting them to a wide variety of data sources. These products allow these metrics to be shared with other users as necessary or according to the data governance rules for the business, essentially acting as a toolbox for you decision making tools in your business.
A good set of metrics is essential for unlocking efficient decision making with confidence in a modern business. It’s not always obvious which metrics you’ll need though. To help with this, there are resources describing industry standard metrics and how they should be interpreted and used in most businesses. An example of such a resource is MetricHQ. These provide a great resource for getting going when setting up your metrics.
The tooling though, only helps to unlock the benefits of a good metric discipline. Thinking about how to use data as a problem solving tool is the first step every business needs to take to use data effectively. Metrics are only the resulting tools that come from well thought out business needs combined with good quality data often enhanced with a semantic layer.