The practical guide to KPI’s: measure what matters

Key performance indicators (KPI’s) are much easier to be misused and abused than people generally realize. This guide attempts to clarify what KPI’s do, how they should be implemented into organizations and common pitfalls to avoid.

When a measure becomes a target, it ceases to be a good measure

– Goodhart’s law

What is a KPI (Key Performance Indicator)?

Key performance indicator (KPI)‘s are statistics used to evaluate performance for various parts of an organization or individuals. KPI’s measure past data, and are used to inform people so they can make future decisions.

Why do KPI’s matter? (They’re less boring than you think)

The word KPI is easy to do get you rolling your eyes up into your skull, but they’re actually quite interesting.

They’re powerful when used properly

KPI’s can greatly help drive business outcomes because they are objective assessments of performance. They can make everyone move in lock-step towards a common goal, and show evidence of exceptional performance or progress.

When implemented properly, KPI’s help you identify problems because it allows you to objectively assess whether goals are being achieved or not. Which you can in turn use to diagnose problems to their root cause and design solutions around them.

They’re influenced by human nature

The act of measurement for an individual greatly impacts it’s results. This is generally referred to as Goodhart’s law. Whatever people feel like they are being measured on a KPI, they will typically just optimize for that metric. They influence how people think and what actions they take. More on this below.

If you’re struggling with producing good results for your team/organization, consider re-working your KPI’s. KPI’s greatly influence how everyone acts in an organization, so if you change the KPI’s, you will likely have a big impact on behaviour.

They’re important for you to progress in your career

You can use your KPI’s as a tool to show the work you’ve done, since it objectively shows results for an area you are responsible for. Read more here on how to manage KPI’s with your manager.

What are some examples of KPI’s?

Some common examples of KPI’s:

• Cost variance (project management)
• Percent of overdue invoices (accounting)
• Cost to acquire a new customer (sales and marketing)
• Employee turnover (Human resource management)

How KPI’s actually work when implemented into an organization

What gets measured gets improved

Meaning people optimize to improve KPI’s that they think are being measured on and are visible to others. Including reputation, future job prospects, how you are perceived to others, etc.

Some percentage of individuals will try to game KPI’s

Because KPI’s are used to measure an individual’s performance, some individuals will develop systems, processes, and actions that are selfish. Meaning they choose actions that benefit just themselves over the organization as a whole. Usually when this happens, it is the result of poor KPI design or a bad hire. However, it’s important to honestly ask yourself “If I were in this person’s shoes, would I do the same thing? Why or why not?” And if the answer is that you honestly would, it’s probably time to change your KPI’s.

Optimizing for KPI’s often creates tradeoffs that need to be considered well in advance

KPI’s often influence each other, because resource optimization towards one KPI can be at the expense of another.

Some examples of KPI trade-off’s:

• Optimizing sales for the next quarter can effect how long you retain customers, because you may sign less profitable customers that seem good at first pass
• Shipping features quicker can increase the number of bugs
• Ensuring a product is completed on time may decrease the quality of output

KPI’s can create a tunnel-vision effect

If you give an individual broad responsibilities over the function of a business (e.g all lead generation), but then create a north star KPI that only impacts part of their responsibilities, they will inevitably optimize for just what they are being measured on regardless of whether this is good for the business.

Consider the lead generation example above. You’ve hired a lead managerr to get good leads for your business. Meaning you think they should be optimizing the following:

Cold Leads: Get high-quality leads for the sales team to call

Inbound Leads: Get people interested in your product first, so that they sign up for a product demo

Upsell and cross sell: Get current customers interested in expanding to other services you offer or product bundles.

However, inbound revenue is the only easily measurable statistic, since there is a lot of influence in getting the other teams to help you with the cold and upsell leads, you can’t really influence whether the leads get closed.

So you end up measuring this individuals performance based on inbound revenue. You hope that because they have some stock options in the company, they will still optimize for what’s best for the company. Low-and-behold only inbound revenue gets optimized.

In turn, this person only ends optimizing for the one KPI they’re being measured on, because they think that’s how their performance is being measured.

Individuals orient their actions around the timelines for evaluation of KPI’s

If you measure performance of a KPI over short periods of time, individuals will do their best to maximize in-month statistics. Meaning if you’re looking for an individual to maximize the long-term results for the company, then you shouldn’t be looking at monthly output statistics unless an individual is really capable of increasing revenue each month.

The same goes for any other period where KPI’s are measured. A common criticism of public equity markets is that they need to report financials every quarter. Business leaders answer to their board, and the board answers to shareholders. Shareholders want stock prices to go up, so there is a constant desire to increase financial performance every quarter. Which in turn causes many leaders to optimize all of their decisions over a few quarters instead of what’s best for the business in the long term.

Watch out for these key things when implementing KPI’s into your organization

It’s very easy for KPI’s to produce unintended consequences

Because of the human element for KPI’s, it’s difficult to predict how individuals will respond when they know they are being measured. And to further complicate things, departments in an organization can influence each other’s performance. So setting kpi’s can influence the behaviour of one department

KPI’s can create diffusion of responsibility

If you set a KPI goal such as “revenue” and assign this to all of the different teams that influence revenue, it becomes very easy people to lose accountability and therefore lose effort in achieving goals.

Some common things people think to themselves if they’re assigned a high-level KPI that they do not have full control over:

• “The other department is pretty good, they’ll take care of this revenue goal”
• “Why would I work hard towards this goal if

KPI’s can make departments go to war with each other

For departments that need to cooperate with each other, if you define KPI’s at different stages of a value cycle (e.g different stages of a revenue cycle) with no incentive to cooperate, then you will have a very difficult time getting them to cooperate.

Example: A product that charges a monthly fee is sold by a sales team, and is then handed off to a customer success team to service the product.

If you incentivize the sales team to just close deals by making that the way they get their paycheck, then they have no incentive to sign customers that will stay with the company for 2 years, because someone who stays for 3 months is just as good for them.

The customer success team ends up being measured on churn, so they will do their best to point the finger at the sales team for

Build KPI’s that optimize individuals for the long term

Consider using leading indicators for KPI’s. Leading indicators are KPI’s that do not represent the output for a desirable business goal, but are supposed to predict future output.

Beware KPI Creep

An easy trap to fall into is to measure everything when it comes to inputs and outputs. This causes a “KPI Creep” effect where people lose track of what’s important versus not. KPI’s need to be prioritized due to resource constraints.

Understand how KPI’s are perceived by employees

When any person in an organization is evaluated for something they feel is not under their control, it can create a huge demoralizing effect. And this can often happen when someone is being judged by a KPI that they feel they can’t control. So asking about people feel for KPI’s can also help you understand whether KPI’s are built properly for their role or not.

Deeply understand the cause-and effect that drives KPI’s

It’s easy to jump to conclusions when you see a negative result in a KPI. But the truth of it is that KPI’s are a high-level abstraction of what is really going in an organization. Because any KPI can be decomposed into actions taken by individuals, processes, departments, etc. So it’s important to take in all of the relevant context before making any decisions based on KPI’s.

It’s tempting to create band-aids to try and change differences in behaviours

When KPI’s drive unfavourable behavior, it’s tempting to add an incentive that corrects behaviour. There’s a whole bunch of reasons this almost never works in practice, but mainly because it’s the equivalent of putting a temporary band-aid on an issue instead of addressing the underlying causes.

What are the signs of good KPI design?

To be made soon…