The problem with metrics

Can you know something, that you haven’t measured? Of course you can. I would actually argue, that by far the largest percentage of what you know about the current state of your organization was not something you measured – it was knowledge that came to you via some other process than objective metrics. A few recents posts in different weblogs have been talking around this topic.

On Intellectual Capital Punishment Sam Marshall (via Smart Meeting Design) wrote about an article in Financial Times:

What did disappoint me though, was the quote from HP’s CKO, Craig Samuel: ‘If you can’t measure it, you can’t manage it’. Shame on him for using such an outdated cliche. It reinforces the view that management is something you do with spreadsheets. He should be pushing an agenda that changes expectations about what information you need to manage, relying much more on trusting perceptions and qualitative evidence.

On Reforming Project Management Hal Macomber wrote that:

When a supervisor, manager, or organization declares measurements people will quickly adjust their behavior to correspond to their understanding of the measurements… But most organizations have too many measurements… the practice of establishing these measurements keeps management detached from the exactly the operations that they are interested in performing well. Try something else: forego the measurements. Get engaged instead.

Chris Corrigan took a more political perspective and wrote that:

How do I know I have four apples? I count them. This is notable because the subjective truths, the good and the true (in Wilber’s terms) are truths that only exist if you participate in them… To simply sit back and accept the measured approach (pun intended) is to give up responsibility for the truth, and to become complicit in the system that generates that truth from outside of its subjects.

I was thinking about this when a thought struck me that may be painfully obvious to everyone else, but seemed kinda interesting to me. I thought that there are two reasons why we measure anything:
1: To know
2: To become aware

Measuring something will ideally give me concrete, specific knowledge, but it will also affect whatever it is that I’m measuring. Remember the experiments they performed in the car industry (in the 50’s I think) where they modified working conditions to increase productivity? For instance, they turned up the lighting in an area, and that made the workers more efficient. They tried dimming the lights in another area and, strangely, this also increased productivity! What affected the workers’ productivity in these cases was not more or less light, it was a couple of guys with clipboards in the background constantly taking notes. (On a side note, the notion that you can’t measure anything in a system without affecting the system is also a consequence of the uncertainty principle in quantum physics.)

So metrics aren’t bad. Not at all. The problem comes mostly when metrics are seen as the only way to increase knowledge and awareness – eg. when HP’s CKO, Craig Samuel says ‘If you can’t measure it, you can’t manage it’.

The question then becomes whether you will allow yourself to trust knowledge obtained without objective metrics and, frankly, I believe that not to do so is absurd. I would even take it one step further, as I did in a previous post and say that most of the important stuff that goes on in an organization is
a) Not measured
b) Not even measureable

Metrics are used to generate both awareness and knowledge, but to treat metrics as the only trustworthy source is absurd!

18 thoughts on “The problem with metrics”

  1. Alexander,

    I benefit from your discussion of metrics. I think it is a good way to get at the prevailing archetypes in most businesses. Do you think that our desire to measure often comes from a absolute value placed upon control, consistency and predictability? If so, risk taking and creativity will not be welcome. Since I am not against measuring and metrics, how would you suggest we keep our eyes open to the misuse of metrics? These are questions provoked by your blog. People like Peter Block write on this a good bit, but wondered how you might wrestle with this.

    Hope all is going well. Mike

  2. Mike,

    Those are really good questions. I think that in many cases the desire to rely on metrics is really a desire to make management simple. I’ve worked with managers on balanced scorecard implementations where the underlying idea was that, "Well, once we have our scorecard up and running, all we need to do is look for red lights and handle those". In other words, the metrics will make it simple for us. But management, by definition, cannot be simple. If something is simple, why manage it?

    And if the overall goal (whether stated or unstated) is to keep the job of managing or leading simple and easy, then, yes, risk taking and creativity are dangerous, since they introduce even more uncertainty. In fact any change effort is suspect for that very reason.

    Metrics are risky, when you can rely only on them. When there is a sense that only knowledge obtained through objective measurement is trustworthy. That everything else is undependable hunches and intuition.

    Another thing to look out for, is that metrics are often used to find fault and mistakes and to either mete out punishment or (ideally) correct the mistakes. This leads to a focus on what’s wrong and what’s not working.

    This is no fault of the metrics themselves, since they could just as weel be used to find the places where things are going find, and serve as the basis for giving out recognition and rewards. But too often, metrics are used to find fault rather than to identify and spread best practices.

  3. Alexander, thanks for the comments. Your observation that metrics are ways of finding fault and mistakes has me thinking a bit more. Always a dangerous thing…
    Metrics are introduced as neutral but I wonder if the typical line worker ever really sees them that way? Maybe if the metrics are agreed upon and brought about by a process that included the people being measured; but that is not the common practice. Both my brothers are in medicine and they talk a lot about health care "being on the wrong numbers". Well, just half baked ideas right now. But thanks for the conversation!!! Be free, Mike

  4. Mike,

    Yeah, gotta watch that thinking :o)

    What’s your viewpoint on all this? Where do you think the drive to measure stuff comes from, and what are the positive effects we can get?

    Also, you mentioned that Peter Block’s been writing about this. Could you say a little about that?

  5. Alex,
    Metrics seem (in part at least) a matter of world views; mechanical vs. organic. Nature doesn’t measure. It doesn’t say "this is how tall an oak tree is in order to be an approved oak tree". Rather nature forms crucial relationships; DNA, water, sunlight, forms of resistence and says "grow!". This is part of what I like about Block’s notion of leaders as conveners who invite people into relationship with each other.

    Block covers this some in The Answer to How is Yes under his discussion of archetypes. When engineer and economist archetypes dominate our corporate cultures we can expect metrics to dominate. Metrics combine with the values of consistency and predictability. Metrics fit into a world view that is intrumental (mechanical) rather than idealist and intimate.

    I have been asked lately to address subjects of change, creativity and innovation; in May I will be speaking to nearly 600 city officals in a message entitled; Kissing Change on the Lips. I find that the cultural cores of my client’s companies/organizations are so instrumental and mechanical, that creativity and innovation cannot thrive. Most of my work these days is in off site meetings working on cultural core issues. It’s a blast!

    Alex, thank you for inviting me to this space to be part of this conversation.

    Mike

  6. Thanks for your input Mike. "Nature doesn’t measure". I love that.

    Also the link between measurements and relationships is interesting! I’ll have to chew on that a bit.

  7. Hi Alex, thanks for the reference.

    A variant on the measurement question is about Return on Investment. Its a classic killer question from the management-by-spreadsheet school. So I did a little list of expensive things that people do without worrying about measurement.

    * Global workshops where execs jet in from all over
    * Most training courses
    * Furniture beyond the bare necessities (including the ROI of carpets, plants and large desks)
    * Most downsizing programmes (they show apparent savings, but they imply one investment strategy over another, and nobody works out which ultimately yields more)
    * Attractive buildings vs. concrete monsters
    * Public parks
    * Travelling non-Economy class
    * Owning a cat
    * Calculating ROI

    Sam
    ps the experiment you refer to was the Hawthorne Effect and it was, surprisingly, way back in the 1920s at Western Electric. see http://en.wikipedia.org/wiki/Hawthorne_effect

  8. Ecellent point, Sam, metrics and ROI are closely related and subject to many of the same fallacies.

    I like your list of things we don’t require to demonstrate ROI and especially the last item: What’s the ROI of ROI-calculations. Had me laughing out loud.

    I think that this “management by spreadsheet” (and thank you for that term, also) is in many cases a defence againt unwanted ideas. If managers don’t want something to happen they can ask for ROI calculations and dispute them ’till the cows come home.

    If they do want something to happen (like corporat jets), they can simply do it without demanding ROI analysis. Thus metrics become a tool for defending the status quo and any preconceived notions management mey have.

    Thanks for the reference on the Hawthorne effect – I’ve been looking all over for the original story!

  9. I don’t think metrics have to be bad. There’s just so much bad management going around and metrics are misused for “management by spreadsheet”. I think good management is like good news; you never hear about it.

    thanks for Positive Sharing btw… I intend to scour your archives! :)
    Gav

  10. True, metrics don’t hav to be bad, Gav – they’re like any other tool and can be used for good or bad.

    I like your point about good management being like good news – and be my guest in the archives :o)

  11. Isn’t this post in contradiction with “Should salaries remain secret”, where you recommend that there should be clear measurements of success, openly available, so that you can ensure people get fair salaries?

  12. What lazar, you’re suggesting that I actually need to be consistent in my views across different posts? Damn, that makes blogging a lot harder :o)

    But good question – is there a contradiction? I don’t think so.

    In the company i co-founded we obiously used the hard, objective measurements we did have. For instance how much income did your project generate, how satisfied is your customer, etc.

    But we didn’t fall into the trap of believing that those numbers were the whole truth, there was still plenty of room for the unmeasurables like commitment to the company and how much fun it was to have you around and people were equally valued on those parameters.

  13. I see your point. However, by introducing unmeasurables you are kind of making the whole process if not completely subjective, then definitelly less objective, and therefore less fair. ‘Less fair’ is something you wanted to avoid in the article on salaries.

    Nonetheless, probably the most interesting blog on the topic. Thanks for the time.

  14. Well lazar: I don’t think you can equate subjective/unmeasurable and unfair. Our measures of performance were partly subjective but just as fair as the objective ones.

  15. I agree, but I guess its a bit more difficult to understand that one scores lower on a friendship scale than his/her colleagues, and therefore may seem unfair. Then again, maybe I’m just too dumb for this.

  16. Geek’s point of view: why not just automate management-by-spreadsheet managements?

    I mean, if the world is so simple (as it apparently is in HP), why not just add some macros to those spreadsheets and be done with it? This would sure better the company economics right away (saving those salaries) … ;o)
    (+++: buggy spreadsheets are easily fixed!)

    Seriously: metrics are part of the big picture – not all of the big picture. Maybe 33%. The talk-to-your-employees-and-this-way-learn-your-business exercise is the second 33%. I personally think that visionary thinking is the rest.

Leave a Reply

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

This site uses Akismet to reduce spam. Learn how your comment data is processed.