Here's a conversation that takes place quite often when starting a new project: a company comes to us wanting a dashboard. They've got a list of charts they want to see … maybe some metrics their board asked about or a report that someone at a conference mentioned. So they say, "We need to track X, Y, and Z. Can you build that?"
And the answer is almost always yes, we absolutely can. But the more interesting question is whether we should. Or more precisely, whether those are the right things to track in the first place.
The reality is that while making the decision to modernize the data stack is almost always the right call, many of us are prone to think about the process backwards. We have a natural tendency to focus on the output KPIs (revenue, churn, customer count) that we ultimately get measured against while not emphasizing the input KPIs that make those outputs possible. It's like a football coach who obsesses over the final score but never watches game film to understand why his team keeps losing in the fourth quarter (shoutout to Brian Daboll).
What We're Actually Talking About
Let me be concrete about the distinction here, because the terminology gets fuzzy.
Output KPIs measure results. They're the scoreboard: revenue, profit margin, customer acquisition cost, churn rate, NPS scores. These are the numbers your board cares about, the ones that show up in investor decks - they tell you what happened.
Input KPIs measure the activities and resources that produce those results. They're the game film: sales calls made, marketing spend by channel, support tickets resolved, features shipped, onboarding emails sent - they tell you why it did (or didn’t) happen.
The fascinating thing about this distinction is that it maps almost perfectly onto controllability. Outputs are lagging indicators that you can't directly influence. You can't just decide to have more revenue tomorrow. But inputs? Those are leading indicators, the behaviors and investments you have more control over. You can decide to make more sales calls tomorrow. You can decide to increase your ad spend in a specific channel.
Why This Matters for Your Data Stack
When we kick off a project with a new client, we don't begin by asking "what dashboards do you want?" We start by walking through their business processes and asking two deceptively simple questions:
- What are the 3-5 output KPIs that define success for each part of your business?
- What inputs do you believe are driving (or should be driving) those outputs?
This sounds basic, but it's remarkable how often the answers reveal disconnects. A company might say their top priority is reducing customer churn, but when you look at what they're actually measuring, there's nothing about customer health scores, support ticket volume, product usage patterns, or any of the leading indicators that would tell you before a customer leaves that something's wrong.
We use a data architecture template that forces this kind of thinking (more on that in a minute). It maps business processes to KPIs to source systems to data models. The structure isn't complicated, but it ensures that every metric we build traces back to an actual business question, and that we're capturing both the what and the why.
The Diagnosis Before the Prescription
Think of it like receiving medical treatment. If you show up at the doctor's office and say "my head hurts, give me something for it," a good doctor doesn't just hand you Tylenol and push you on your way. They ask questions: where does it hurt, when did it start, what were you doing, have you had this before? They're trying to identify the root cause, not just treat the symptom.
The input / output framework does the same thing for data. When a client says "we need to improve our close rate," the output KPI is obvious: close rate. But the input KPIs require investigation. Is the problem lead quality? Sales rep activity levels? Time-to-response on inbound inquiries? Pricing? Demo-to-proposal conversion?
Each of those inputs can be measured, tracked, and improved. And once you know which input is broken, you can actually do something about it. You can train reps, adjust targeting, change your pricing strategy, or fix your demo flow. The output KPI just tells you there's a problem. The input KPIs tell you where to look.
A Practical Example
One of our clients runs a network of locations, and they came to us wanting a "simple revenue dashboard." Fair enough. But when we dug into their business processes, we realized revenue was the output of a chain of inputs:
- Marketing spend and channel mix drove lead volume
- Lead response time and sales rep activity drove conversion rates
- Onboarding quality drove customer retention
- Customer health scores predicted churn risk
If we'd just built a revenue dashboard, they'd have known whether they hit their number each month. But they wouldn't have known why, which means they couldn't have done much about it when they missed.
Instead, we suggested a system that tracked the full chain. Now when revenue dips, they don't have to guess. They can trace it back: was lead volume down? Was conversion suffering? Did we have a retention problem? Each input KPI points to a specific team and a specific set of actions they can take.
The Bottom Line
Starting with input vs. output thinking isn't about being academic or overly structured. It's about building data systems that actually help you run your business, rather than just reporting on it after the fact.
The companies that win with data aren't the ones with the fanciest dashboards or the most expensive tools. They're the ones who understand the relationship between what they do (inputs) and what they get (outputs), and who measure both sides of that equation.
If you're about to embark on a data project, whether it's your first warehouse or your fifth BI tool migration, start by getting clear on this distinction. Map your business processes, identify your output KPIs, then work backwards to find the inputs that drive them. Everything else follows from there!
If this resonates, we've got a free template we use with clients to kick off exactly this kind of diagnosis: mapping business processes to KPIs, identifying source systems, and designing data models that actually answer the questions that matter. Drop a comment or shoot us a note if you'd like a copy.
Thanks for reading!
#DataStrategy #Analytics #KPIs #BusinessIntelligence #SouthShoreAnalytics