Tinker, Taylor, Culture, Failure


-When are managers and supervisors going to sign their commitment?

The above question was raised when 30 or so workers were asked to read and sign their commitment to the organisational Safety Culture values. It happened during one of their fortnightly safety meetings. My presence was merely as an observer.

-Well, we started with this level, and then we’ll do supervisors’ and managers’ responsibilities as the next step, the safety supervisor replied.

Individuals are different. By definition. Any one organisation is made up from a range of unique and diverse physical and psychological dispositions, values, norms and expectations. Taken together, it’s quite messy. Untamed, the diversity may be uncoordinated, chaotic, random, accidental, and subject to whatever influences. Or so we seem to assume anyway.

To bring clarity and order to this unmanaged cacophony, to increase efficiency, or for example to improve safety, a prevalent organisational response is to standardise. Diversity is turned to uniformity. Plurality to unity. Dissonance to harmony. And messiness is reduced until manageable order prevails.

This has also long been the trademark of science – to take what is unique and seemingly savage and explain and control it with tested and validated theories and laws. As scientific knowledge has replaced local forms of understanding the world, power to decide what is best has ended up with those who have the intellectual capacity to formulate and explain concepts and models.

In an earlier post here, Sidney Dekker outlined how a similar ambition drove the early organisation theorist Frederick Taylor to locate the power to define how work should be done at the managerial level. This while workers should follow what managers had thought out, in exchange for remuneration.

Such a reverence to expertise, and a similar disrespect for non-experts, largely informs what now goes on the Safety Culture space. The standard setup goes something like this:

  • A Safety culture program is imported from somewhere far away.
  • The program normally started with a consultancy group that did a survey for a company in a respectable safety industry.
  • The group defined a set of norms, values and expectations that local people had suggested would help them to achieve safe operations.
  • Then, data was collected to show that where there was ‘more commitment’ to those particular values there were fewer safety incidents.
  • And if this circularly generated proof is not enough, additional weight is given by showing how many other companies have bought into the program in question.

Soon, it ends up in sloganized form on posters, in gap analysis sessions, and sometimes in struggles to get workers to sign a piece of paper as a symbol of their commitment to the ‘right’ values. In effect, subjectivity is removed; experts define the way forward; local judgement is considered irrelevant and unnecessarily complex. It is social Taylorism.

This approach is problematic for at least two reasons:

First, it is based on the presumption that we can control risks by specifying acceptable behaviours. This, however, requires knowing about every possible situation, and also having universal definition of the terms used. In a world that is dynamic, complex and in which meaning is defined locally, providing fixed standards is bound to miss a lot, and create confusion.

Second, the idea about a set of standard behaviours that should happen across every situation, denies culture as an emergent, complex and ambiguous phenomenon. When responsibilities are organised into functions, and people assigned to roles, mess is simultaneously organised to a controlled order. Which is what was wanted. But more importantly, such solidification of expectations leaves little room for discussing and learning about reasons behind deviations from such ideals. By providing a standard, by establishing what’s right and wrong, we create an organisation that is less interested in learning about difficulties in creating safety, and more interested in enforcing an imagined preferred state. Consequently we fail to capture discussion about issues like conflicting goals, production pressures, difficult trade-offs, ambiguity, and other safety-critical matters.

Now, if it is not possible to create a safety culture by doing copy-paste and specifying the ‘right way’, then what is the option? Is it possible to create a safety culture that is flexible and adaptive? And how can we create a culture that enables rather than constrain?

I cannot see any other option than to turn to each other. And to embrace the things we want to avoid. By engaging with social mess and diversity, by including ever more complexity, by listening to what’s difficult and challenging, we can start taking responsibility for what is. But also, we need to hear about more potential ways than just one. What do people believe about safety? And how do they want to relate to each other to achieve safety?

This way, diversity stops being a threat. Instead it becomes a source for creativity and making things better. By seeking out the disturbing, what does not fit, and engage with it, we stand a much better chance to adapt our way to a solution that is home-grown, firmly rooted in the local workforce, more resilient, that relies on collaborative intelligence, and one which is easy to sign.

Note: This post was collaboratively thought-out with Eder Henriqson and Roel van Winsen, from the Safety Innovation Lab, Griffith University, Brisbane. Also, kudos to Sidney Dekker for suggesting the title for this post.


  1. John Culvenor Reply

    Hi Daniel,

    Yes, good points. But no one really should expect that these programs do anything. The safety culture programs are just a product to sell. Consumers follow this purchasing pattern all the time. A bit like selling cosmetics. Get a well known model or actor who already looks good, have them use and endorse the face cream, put them on the ad, then consumers buy the face cream.

    So in safety, find an organization with first class engineering who uses XXX program and get everyone else to buy it too.

    It should especially have a trademarked name preferably with an innovative way of spelling or writing the word such as TH!Nk4SAFETy ™ which shows that a lot of thought and research has gone into the product 🙂

    Or indeed, for the best of the best, you can’t go past my own BEANIE ™ program based on the latest neuroscience http://goo.gl/C9kGU

    1. Daniel Hummerdal Post author Reply

      John, my criticism apart I do believe much good can be achieved by asking questions around what people believe is important and difficult around safety. Not a huge believer in providing the answers though..

      Is ‘first class engineering’ the biggest hope to improve safety?

  2. Jephte Lafour Reply

    Hi Daniel,

    Nice piece!

    It leads me to the question: why would anyone voluntarily sign a commitment to the organisational Safety Culture values?
    A handshake should do, right?

    1. Daniel Hummerdal Post author Reply

      Jephte, good to hear from you!
      Not sure how a handshake is different from a signature in relation to the above. Isn’t both based on ideas that we can perfect the set of behaviors and then we will arrive to the promised land?

  3. John Culvenor Reply

    Hi Daniel,

    First class engineering? Yes. No contest. Try these examples:

    Compare existing elevators with this idea; let’s have no double doors on elevators (i.e. just a hole) and rely on a line with safe behavior as the protection.

    Compare current microwaves with this innovation; have the switch on a microwave that turns the power off when the door is opened and just rely on a safe attitude.

    Compare non-toxic paint with this “new” way; switch back to lead paint on childrens’ toys. No worries; winning ‘hearts and minds’ of the mums and dads will ensure proper supervision and prevention licking of toys and ingestion of lead. Hooray, problem solved. TH!Nk4SAFETy ™…Child Protection Edition.

    TH!Nk4SAFETy ™** its the new way toward a safe leadership hearts and minds behavioral science culture.

    ** Conditions apply. It doesn’t work.

      1. mikebehm Reply

        Hi Daniel. Great post and discussion, and yes, that is an important question in this whole “culture” discussion that frequently gets overlooked or too easily dismissed. And it’s not just limited to engineers or even to formal designers for that matter, but extends to those designing and planning the work. In 1991, Joe Stephenson, in his book System Safety 2000: A Practical Guide for Planning, Managing, and Conducting System Safety Programs, stated “The safety of any operation is determined long before the people, procedures, and equipment come together at the work site.” I love that quote! Safety culture starts at the beginning or any project, product, process, technology, or at the “top” and is about arranging conditions for workers to utilize their knowledge, skills, and abilities. Workers are resilient. When something at-risk occurs, workers are usually innovative and creative to avoid loss but the feedback must go to design and planning for improvements. And, how about getting worker knowledge more directly into the design and planning phases…now that seems to be a great source of innovation and might be indicative of this safety culture thing. I think safety professionals and researchers should do more to connect culture to upstream thinking. It’s interesting, difficult, and I have so many questions. In my experience, it seems the US lags behind in this area of safe design…

        1. Andrew Townsend Reply

          Mike – Thanks for the Joe Stephenson quote. I shall be using it in my currently unequal struggle with the UK regulator.

        2. jephte lafour Reply


          Thanks for bringing that up.

          May I quote Martin Sibileau, a macroeconomist sharing a recent thought on his website: ” Why do we need a NASA? How many dreams have been frustrated by the government intervention in this area? The space shuttle program did not innovate in design during its whole life, since 1982. Yet, Sir Richard Branson challenged engineers to produce a space plane for tourism and numerous designs came forth, for Virgin Galactic…”
          And if one goes to VirginGalactic.com, under research: “1) One type of flight offers researchers—whether they be academic or corporate, scientists or engineers, teachers or students—the opportunity to board SpaceShipTwo and fly to space with their experiments, becoming astronauts themselves as they conduct their research.” And: “If interested in flying your researchers or payloads on board the revolutionary SpaceShipTwo platform, please send an email to research@virgingalactic.com.”

          So I associate the discussion with another website of which I took away that if a group of children(sec) comes to play at a birthday party the parents could set behaviour rules in order for the children to, f.i., stay within certain boundaries. They could even make the kids sign a commitment as to how to behave (and some good stuff about values). Or…the parents could challenge the kids to play a game with a goal or some target. According to these safety thinkers the former approach was bound to fail…

  4. Jephte Lafour Reply

    Daniel, not sure about the promised land, but if workers have a powerbase, than instead of signing papers – (in exchange for remuneration?) – they could shake hands with the managers and than the commitment would be mutual and trust based. The handshake seems culturally different than the signature, perhaps it functions as a cultural determinant (how we do things around here). Anyway, it was my takeaway from your piece, that some workers, so long after Taylor, still lack a powerbase.

  5. Jephte Lafour Reply

    Well Daniel, and I agree with you, from the heart:

    From another website, I quote Hugh-Smith’s quest post today, titled “Why suppressing feedback leads to financial crashes”.
    And quote: “If we see the economy as a system, we understand why removing or suppressing feedback inevitably leads to financial crashes. The essential feature of stable, robust systems (for example, healthy ecosystems) is their wealth of feedback loops and the low-intensity background volatility that complex feedback generates. The essential feature of unstable, crash-prone systems is monoculture, an artificial structure imposed by a central authority that eliminates or suppresses feedback in service of a simplistic goal–for example, increasing the yield on a single crop, or pushing everyone with cash into risk assets.”

    1. Daniel Hummerdal Post author Reply

      Beautiful. Using available diversity and allowing information to flow is good stuff! (unfortunately an idea that runs counter the the popular ‘predict, plan, control and constrain’ approach to management)

      Which then begs the question: What makes a good ‘power base’ for enabling a ‘safety multiculture’?

  6. Jephte Lafour Reply

    I think any power-base will do as long as it balances central power sufficiently to keep feedback flowing. Interesting in the context of Taylor is of course Semler and his wildly succesful Brazilian company Semco were this powerbalance is designed into the organization. Wikipedia says Semco is known for its radical form of industrial democracy and corporate re-engineering.
    Semler says if you give workers freedom they will be more responsible. The idea of control is an illusion, according to the CEO.

    1. Daniel Hummerdal Post author Reply

      Interesting read about Semco!

      “People care about what they create” (Margaret Wheatley). And I increasingly wonder if the opposite is also true: people don’t care about what they have not created!?

  7. Steven Shorrock Reply

    (Safety) Culture is not a programme, except maybe in the sense of the “collective programming of the mind”, as Hofstede put it. For a start, a ‘programme’ has a start and an end. It also has some kind of structure or plan. Safety culture has none of those things. It just is. A programme to assess or evaluate and try to improve safety culture intervention does, but that is different, and the two should not be confused.

    I am yet to find one organisation that has only one (safety) culture. Rather, organisations tend to have several (safety) cultures – distinguished by profession, rank, location and even shift. Safety culture can’t be imported from elsewhere – there are too many unique factors, and a unique history. There are usually some commonalities that span sub-cultures within an organisation, and national culture imports quite a few – acceptance of uncertainty and messiness may suggests a Scandinavian one, for instance :-). There is often more difference within an organisation than between organisations, in a given sector. The commonalities may well be more along professional lines, such as pilots, surgeons, etc, than organisational lines.

    Once safety culture is manualised and defined by an organisation in terms of acceptable behaviours, it risks retiring to the same shelf as performance appraisal schemes that include such ‘organisational behaviours’; also often top-down (from HR) and contextually sterile. This is behaviour as imagined, not as done. People look at the ‘behaviours’ once to fill our the form, or read the new laminated card, then that’s it – on with the job. If, however, the people who do the work discuss and find out what what they agree on, what they don’t and what works best for them, then great. That’s safety culture. But to focus on individual behaviour is to miss the system, and thus the major opportunities for improvement.

    In many ways, safety culture can be seen as an ongoing conversation – the energy which flows through the system. Hence why what excites me most is bottom-up safety culture approaches. Hence a shameless plug to http://www.skybrary.aero/index.php/Safety_Culture_Discussion_Cards

      1. Steven Shorrock Reply

        Cheers Daniel. Yes the whole idea is to raise questions and try to demystify and democratize ‘safety culture’. The cards are available in a few languages – currently EN, FR, ES, HU, FI – and increasing soon to IT & PT. They are also free, which helps. I find out (often accidentally, which is nice) that they are used in lots of ways – for brief conversations, for longer workshops, as random ‘thought for the day’ pop ups on intranets, as posters (even using the more challenging management posters in management corridors), etc. They are not perfect and now I would change some wording, but they were done as an experiment. We move onto an ‘Edition 2’ soon, and I’ll be sending some calls for input via social media. Trying to think how best to do it…

        For me the energy within systems is conversations. I’m not really interesting in measuring them though. Capturing them, yes, and especially the meanings and stories within.

        Here is a ‘hard’ stance on this, which sees organisations AS conversations: http://informalcoalitions.typepad.com/informal_coalitions/2013/05/from-talking-in-terms-of-systems.html

        I prefer to see conversations as the lifeblood, energy, or chi that exists within a system.

  8. Alan Glenn Thompson Reply

    Hi Daniel,

    I’d just like to share my experience that not all behaviours (acceptable or otherwise) espoused by cultural improvement initiatives are meant to provide a specific response to a situation or risk; often they are often around topics such as communication, openness, fairness, inclusiveness, consultation, personal involvement and accountability etc.

    Rather than trying to programme people to react in a certain way to certain situations, they attempt to establish a more open-minded approach where different views are sought, discussed, and considered on their merits thereby drawing on the strength of diversity without attempting to homogenize it.

    These are the better ones.

    I have, of course, encountered many consultants pedaling their guaranteed, clinically proven, one-size-fits-all culture-strengthening formulations in the past. Sometimes these are purchased by desperate senior executives under pressure of recent incidents to show that they’re doing ‘something’ to improve safety outcomes. The memory of one particular product that I was recently stuck with was so bad it still makes me GOL (groan out loud) when I think about it.

    BTW AFAIK GOL is new chatspeak. TTYL.

Leave a Reply

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