Executive Summary
Your Supply and Demand teams are using different numbers for what should be the same metric. This leads to long debates in meetings and slows down decisions.
Time spent arguing about whose number is right is time not spent on improving the business. It wears down trust in the data and can lead to poor, reactive choices about things like inventory or marketing spend.
This usually isn't a problem with your dashboards or your people. It's a sign that your metric definitions haven't kept up as you've grown. The fix is to agree on a single, clear Metric Definition and build it centrally into your data setup.
When teams report different numbers for the same thing
It's a familiar scene. In the weekly operations meeting, the Head of Supply shares a dashboard showing 10,000 active listings. A few minutes later, the Head of Demand's dashboard shows only 8,500 'bookable' listings. The conversation gets stuck. People start questioning the data, and analysts are asked to go away and figure out why the numbers don't match. Another week is lost.
This sort of thing happens a lot. For many operations leaders in businesses with a high volume of transactions, the debate about the data becomes the main event, and the actual strategy gets pushed to next week. Over time, trust in the data, and in the team that produces it, starts to fade.
Why this happens as a company grows
It's tempting to look for a faulty SQL query, but the cause is usually a bit deeper. What looks like a data problem is often just a symptom of teams not having a shared definition for a key business idea.
I see this pattern in many online marketplaces, typically after their Series B. In the early days, a small team can get by on shared understanding. But as the company grows, teams specialise. The Supply team develops its own logic for an 'active listing', suited to how they see the world. The Demand team does the same. Both definitions make sense in isolation, but they don't match up.
You may have invested in good tools like Snowflake, dbt, and Looker, but the problem can persist. If the underlying definitions are confused, the new tools just help you generate confusing numbers more quickly. The issue isn't the technology, it's that the business logic is tucked away in different places rather than being managed centrally. It's a gap in Data Governance.
How to create a single, reliable metric
The way forward isn't usually about hiring more analysts to check reports. It's about treating your key metrics with the same care you'd give to a product feature. The goal is to stop the debate by building a Single Source of Truth.
The approach I take is quite straightforward.
#### Step 1: Get everyone to agree on a definition
Before any code is written, the first step is to get the heads of Supply, Demand, Finance, and Product in a room together. My job is to help them have the sometimes tricky conversation to agree on one, and only one, definition for each core metric. What exactly is a 'listing'? What makes it 'active'? What is the precise logic for 'bookable'?
It's more about getting people to a shared understanding than it is about the technical details. The output we're looking for is a documented consensus.
#### Step 2: Write the definition into your code
Once the definition is agreed, it shouldn't live in an analyst's local files anymore. Instead, a good Analytics Engineering team will build it into a central part of your data setup, using a tool like dbt or LookML. The definition of 'active_listing' is now written down once, tested, and kept under version control.
This definition becomes the single source. Any changes have to follow a proper process, which stops the definitions from quietly drifting apart again.
#### Step 3: Build one dashboard that uses the new metric
This is the point where we can build a dashboard that everyone trusts. We can replace the separate, conflicting reports with a single one for the management team. The Head of Supply and Head of Demand now see the same number because the dashboard is built on those same, centrally-defined models. There is no other number for it to show.
The result: Moving from arguing to deciding
In my experience, the effect is felt very quickly. Those weekly operations meetings become much more productive. The first ten minutes are no longer spent debating the data, because everyone trusts the number.
The conversation shifts from 'Whose data is right?' to 'What should we do about this number?'. The data team can stop spending its time checking reports and can instead help the business figure out what to do next. You end up with a solid, reliable foundation on which to scale your operations.