Set the right KPIs before you kill your business

How the wrong KPIs will lead your organisation to successfully achieve negative results.

Ben Ralph
May 28, 2020

In uncertain, complex, and ambiguous times where change is the only constant, it is easy to lose sleep to heavy-hitting questions about your business.

  • Is my organisation heading for a cliff?
  • Am I moving too slow and not taking on enough risk?
  • Are the naysayers right and I should take their advice?
  • Are the naysayers wrong and I should trust my instincts and push forward.

If you’re like most, the go-to position for fending off insomnia are your KPIs. You put our faith in data, because cold hard facts can never lead you astray.


The problem with Key Performance Indicators (KPIs)

As the saying goes you cannot manage what you do not measure.

Problematically, too many companies are measuring the wrong things.

Too often we fall into common traps:

  • We measure what is easy
  • We measure what we can control
  • We measure what we have always measured
  • We measure what makes us look good

This can lead us to measure a complex constellation of metrics that we are all too afraid to admit have no real impact on our decision making.

We see this with organisations that have a KPI measuring:

  • Staff surveys rather than staff engagement.
  • Brand Awareness over Qualified Leads.
  • The number of employees rather than the lifetime value of each employee.
  • Corporate acquisitions rather than the strategic benefits of actions taken as a result of the acquisition
  • New features delivered rather than customer problems solved.

How to avoid the common traps and pick the right metrics

Our mantra for selecting a reliable set of dependable KPIs is to obsess over outcomes, monitor signals, and eliminate noise.

Step 1: Obsess Over Outcomes

An outcome is simply a statement that reflects what you want to achieve.

Good outcomes equate directly to the success of your organisation. They are not abstract, they are clear statements that if achieved mean you have in fact succeeded.

Do not worry if you don’t know how to measure them at first. Remember, we don’t want to measure something simply because it is easy to measure.

To avoid the trap of measuring what’s easy, it’s important to not go straight to defining or selecting a measure. Define what you will achieve, then figure out how to measure it. If you’re not sure, we can help.

Example One — I want all my employees to be motivated and committed

Example Two — I want my profits to increase substantially over the next two years, without sacrificing customer satisfaction.

A checklist for writing a good outcome statement

Step 2: Monitor Signals

Most outcomes don’t have a single measure of success so a signal is a possible indicator of success.

It should offer insight and be used as a tool for diagnosis.

Step two is to select a limited amount of quantitative measures that will provide you with meaningful information about if you are going to achieve your desired outcomes (or not).

We call these signals, because they should literally guide your day to day decision making.

Think of them like road signs. They tell you to speed up, slow down, alert you to detours or unexpected impediments.

Again, don’t simply measure what is easy or readily available. More than likely you will need to change a process, implement a new technology or hiring someone with specific expertise.

When you’re thirsty you want a glass not an ocean. Data is the same.

Example OneI want all my employees to be motivated and committed.

Potential Signals:

  • Percentage of meetings in which everyone contributes and feels psychologically safe enough to speak up.
  • Sick days are low (under x on average)
  • Attrition is low (average employee stay is x)
  • Team members are successfully achieving their desired outcomes.


  • Employee surveys
  • Total hours spent in meetings or answering email
  • The person who spends the most time in the office.

Example TwoI want my profits to increase substantially over the next two years, without sacrificing customer satisfaction.

Potential Signals:

  • Increased leads through the new website
  • Increased conversion of online leads
  • Increased profit of online leads
  • Sustain or improve customer lifetime value
  • Reduced customer support requests


  • Total website views
  • Brand awareness
  • Ad impressions

Remember that a signal does not equal success; it just helps us navigate to our desired outcome.

While it’s critical not to simply measure what is easy, it’s always better to frame a signal so it is easy to monitor.

Ideally, all measures are viewable in a dashboard and stay out of the way unless they’re moving in the wrong direction or you’re making an impactful decision.

For example, if you were to visualise sick days taken to show green when trending down, yellow when static and red when it is on the way up you could safely ignore it unless it turns red.

Signal Checklist

Step 3: Eliminate Noise

Lastly, take the time to list the metrics you’ve considered, but are intentionally going to ignore.

When busy or stressed, people are inclined to seek information that makes them feel good — we look for data that reflects well on us and challenge data that makes us look bad. This is known as confirmation bias.

After a bad or unsatisfying quarter, it can be tempting to look for any data point that goes ‘up and to the right.’

Or for an underperforming employee to paint a positive narrative using whatever positive data is at hand.

Intentionally making a list of all of the data we consider to be unreliable or purely aesthetic saves us from ourselves.

Example OneI want all my employees to be motivated and committed.

Potential Noise:

  • Number of new employees
  • Positive rating on informal staff surveys

Example TwoI want my profits to increase substantially over the next two years, without sacrificing customer satisfaction.

Potential Noise:

  • Number of new site visits
  • Increased ‘brand awareness’

As with most things, there is no magic set of infinitely applicable KPIs for you to use that will predict success 100% of the time.

But if you obsess over outcomes, monitor tangible signals, and eliminate the noise, you will be well on your way to a good night’s sleep.

Special thanks to Tim Stroh for all of your advice and insight in relation to this topic.


Reach Out!

Ready to find the perfect fit for your project?

Do you need a world-class team to help you deliver your digital product? We’ve got just the people for you.

Let’s Talk!