Measure Outcomes, not People

I recently had an exchange with Jurgen Appelo on Twitter about measuring people’s performance. During the conversation, he said it was impossible to fit the dozen or so recommendations he had into a single tweet.

It all just sounded too complex (complicated? :-) ) to me and why bother with measuring individual performance when there are more useful ways to improve outcomes? So this blog post is to present my own perspective.

You can follow Jurgen’s original tweet and the conversation here.

Jurgen said ‘Before you “measure” someone else’s performance, explain how you measure your own’. It rings true enough – many managers are quick to judge and measure their staff without any sensible way of measuring their own performance.

But why are we measuring individual performance in the first place? There’s a deeper question here.

What is measuring performance supposed to achieve?

It might sound obvious and taken-for-granted but this is worth examining.

Deming said that only 5% of the performance of an organisation (system) is down to individuals, 95% to the system

In a world where the performance of organisations is believed to be principally down to the performance of their individuals, then measuring their performance makes a lot of sense. The theory is that if only the people perform well then so will the organisation.

This is a world of performance appraisals, coaching and personal incentives (money, benefits) tied solely to individual achievements.

But do most of us live in such a world?


For the vast majority of people working in organisations they are constrained by the systems in those organisations. Rules, processes, hierarchies, rituals, structures – all of these things conspire to limit the difference that an individual can make. This makes perfect sense since most people do not deliver a product or service on their own; they are part of a bigger structure.

Deming said that only 5% of the performance of an organisation (system) is down to individuals, 95% to the system. Whatever the figures in an individual case, it is a powerful argument to look at outcomes, not individuals. More recently, John Seddon and others have been raising awareness of the same issue decades after Deming.

The purpose of most organisations is to deliver an end-to-end product or service to customers. That’s what we need to focus on: performance measures that align to end-to-end delivery.

Misplaced Competition

Comparison of the consequences of Competition between Individuals versus Competition between Outcomes

Focusing on individual performance forces team members to compete within each team. This is destructive competition, since it pits people against each other – when in fact they need to collaborate together to deliver an end-to-end service.

It encourages hoarding information and protecting personal advantage. It also is time-consuming and expensive: creating an industry of performance tracking and assessment for every staff member.

In a similar vein, when an organisation is structured into functionally-seperated units or departments, competition between departments pits groups against each other – when in fact they need to collaborate together to deliver an end-to-end service.

It encourages delays and waste due to each group making their own work “the most important thing”.

True constructive competition within businesses should be between independent units who deliver an independent slice of end-to-end demand

Both approaches undermine an organisation’s true purpose. They may be well-intentioned but they undermine it nonetheless.

True constructive competition within businesses should be between independent units who deliver an independent slice of end-to-end demand (i.e. different services or products).

Then open competition (visibility and sharing of all information) is constructive: teams borrow the best from each other and compete to make their service delivery the best it can be – but never at the expense of other groups. They share the same end-to-end measures of outcomes.

Enter A Better Way

Instead of focusing on individual performance, define a simple set of end-to-end outcome measures for each product or service, based around these three simple measures:

Cycle Time – How fast the service is delivered or how fast are product increments delivered.

Lead Time – How long service requests take to process or how long new product features take to deliver.

Right-First-Time – How well the service is provided (percentage requests processed right first time) or how well product features are implemented (problems/bugs first time).

If you focus on incrementally reducing Cycle Time, then you drive waste and delay out of your organisation; simultaneously cutting costs and improving service.

If you focus on incrementally reducing Lead Time, then you create flexibility and ability to respond rapidly to changing needs. That enables taking advantage of opportunities that are impossible with a huge tanker of projects that takes months to change direction. It also avoids huge missed-opportunity and delay costs.

If you focus on incrementally improving Right-First-Time, then you drive both improved quality with lower costs (re-work is always expensive for both products and services).

You can always bring in outcome measures alongside individual ones.

As you learn what the outcome measures do for performance you can then retire the others. Implementation first starts with making the end-to-end work visible, which you can do using an approach like Kanban.

These are all group goals: they act to bring people together to cooperatively and collaboratively drive improvements. This is far more effective than isolated hero efforts.

Apply these measures to the groups responsible for operating and delivering each service or product (assuming of course that you’ve already organised your people into cross-functional teams).

Of course this isn’t exhaustive – you may well need other measures. But fewer is better because good measures are hard to define and bad measures drive bad behaviour.

Avoid any use of targets because they act to limit ambition. If the goal is to improve service quality by 5% then no-one will look for or seriously push major improvements that might improve it 50%. Remember that the goal isn’t targets, it’s the fastest route to improved outcomes.

But what about...?

At this point some people are apoplectic about the lack of individual performance measures.

“Surely if you don’t have those, some people won’t work hard or put the effort in?”

Not true.

With a properly collaborative environment where the only outcomes that matter are end-to-end ones then there is nowhere to hide – so there’s more incentive for them to shape up. Everyone can see if someone isn’t performing. It’s obvious. You don’t need a manager’s performance appraisal to know it.

When people aren’t performing, then it’s a manager’s job to find out why and help them.

If you don’t have a properly collaborative environment then (again) create one quickly; without it you’ll struggle to get anywhere.

“But what about promotions?”

In a similar vein, people who make a bigger contribution than average will stand-out.

So promote them. It’s as simple as that.

“What if I need to get rid of someone?”

Then do it.

Individual performance justifications are always excuses in such cases. You don’t need the huge bureaucracy of individual performance management to change your team. There’s always a way.

Putting it in a single tweet

My challenge to Jurgen was to condense his answer to “How do you measure your own performance?” into a single Tweet, so it is only fair that I do that here:

You don’t. Measure e2e outcome per service/prod, not individuals. How fast (cycle time), how long (lead time) & how well (right-first-time).

Good luck defining sensible performance measures in your organisation.

Posted in Collaboration, Cycle Time, Kanban, Lead Time, Leadership, Measures, Rewards, Right First Time.


  1. Nice, but I see two logical problems:

    The “end-to-end” metaphor is a gross oversimplification. Organizations do not consist of linear value streams. They are value networks. You must measure value for ALL stakeholders in all directions. There is also value for suppliers, for owners, for employees, for communities, etc. This should be measured.

    You ignore the fractal nature of complex systems. If you cannot measure performance for one person, you also cannot measure performance for one team. And you also cannot measure performance for one organization, etc… Because at every level the system does not exist on its own and performance depends on what the other systems are doing.

    In short: You have some good points, but I don’t consider it a proper answer. :-)

    My upcoming workout article addresses both of these issues. :-)

    • On the first point: yes it is a simplification (or perhaps a starting point) but all value for suppliers, owners, employees, communities should be subordinate to value for customers. They are not equal-value value networks :-) Otherwise the organisation’s purpose is confused and the real customers are not the stated ones. As I said, there are other things worth measuring but they are not normally as important. A few strong clear driving measures closely connected to purpose will go a very long way to driving improvements.

      The second point I don’t agree with at all. Of course you can measure performance of team without measuring individuals – it is trivial to measure whether an item of work delivered by several people is done. It is either done or it isn’t. The individual isn’t the lowest granularity anyway – should one infinitely break down tasks into tiny minute-by-minute fractal sub-tasks? No, that is neither helpful nor scalable, so you measure at a level that makes meaningful sense given the purpose and desired outcomes.

      And of course you can rollup these measures at a higher level but yes you may reach a level at which this isn’t all that helpful (overall corporate performance), although the measures I state are in practice quite useful for comparing quite different services and products so long as systemic differences are borne in mind.

      I’ll read your article with interest when it comes out. :-)

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>