Making visual language less subjective

I once heard about an argument breaking out in an ISO committee meeting about whether or not pictures should be included in the specification of a new management system standard.  The pro-picture lobby argued that the wording was too hard to understand without a diagram.  The anti-picture lobby argued that including pictures would introduce an element of subjectivity, whereas the words could all be rigorously defined in the prologue.

I think a lot of this stems from a misunderstanding of what visual language should and shouldn’t be used for.  We’ve written before about this – try explaining a 2×2 matrix, or the shape described by a quadratic equation, or the outline of North America – in words!  But when you start trying to visualise more conceptual content – business models, social systems, customer journeys and the like – the question remains, to what degree will the visual meaning be subjective?

This is a problem we’ve struggled with a lot over the years, and we’ve discovered that while you can clearly never completely eradicate subjectivity entirely, you can go an awful long way.  Once more, my favourite everyday example of this is the way we use maps.  When you pull up Google maps, you don’t spend your time thinking how subjective the visual language is, because you’re too busy using the visual language to explore, locate yourself, or find a route.  The reason is that wherever possible the visual language of the map corresponds with your experience of the world – trees are green, rivers look blue, and so on.

Now what’s true of maps is also true of conceptual pictures.  The more your visual language corresponds to the way people experience the world, the clearer the meaning will be.  Just to be clear, I’m not saying that pictures aren’t subjective.  The point is that there are ways of depicting things that 99% of people will read the same way.

There will always be 1% who see the world in a beautifully, stimulatingly non-conventional way.  But the point of all types of language – whether visual or verbal – is to convey meaning.  If you want to create visuals that 99% of people read the same way, then there are a number of principles you need to understand about the mind and the visual intelligence.  One of the main purposes of this site is to share what these principles are … stay tuned!

How meaningful is your management model diagram?

One of our ambitions is to find a way of evaluating the “meaning quotient” of diagrams, that is, how much information the diagram conveys that you wouldn’t get from just reading the text.  You would get it by measuring the % of non-verbal information conveyed proportional to the volume of visual language employed.  Even without the rigorous measurement system though it’s still a useful concept to play with.  Let’s try it out …

Here’s the famous McKinsey 7S model of organisational design.  It’s a brilliant example of how little a diagram can add to the concepts it’s intended to illustrate.

The first time I saw it I remember thinking “that’s not a model, that’s a list!”  And of course, if you look at the text of the many management books in which it’s cited, it is indeed accompanied by a list that describes each element in turn:

  • Hard elements:
    • Strategy
    • Structure
    • Systems
  • Soft elements:
    • Staff
    • Skills
    • Style
  • Central element:
    • Shared values

Now, this list of words is clearly meaningful (and presumably helpful, given how frequently it crops up).  But what meaning does the visual language of the diagram add?

The colours merely represent the categorical hierarchy already explicit in the hierarchy of the list.  The lines connect everything to everything, so don’t say anything beyond the rather obvious “these concepts are interconnected”.

In fact, the only meaning the diagram conveys significantly better than the list, is the fact that Shared Values are central.  You can’t show a spatial concept like “central” in a linear form like a bullet list.  But otherwise, it’s just a list.  You could rotate the outer balls around however you want and no meaning would be created or destroyed.

This isn’t to say the diagram isn’t useful.  The 7S framework is extremely useful, and having a recognisable visualisation makes it, well, recognisable.  But what the lack of specifically visual meaning says to me is that the 7S framework is more of a list than a model.

Google around and it’s easy to find even more extreme examples.  Check out this HR diagram, which I found on Pinterest:

hrms-diagram

At least the 7S framework has some implicit categorisation.  Here there’s nothing, it is literally a list.  If you think of my rough definition of a meaning quotient as the amount of conceptual information that the visual language conveys over and above what is conveyed by the words alone, then in this diagram the quotient would be approaching zero.

In fact, the visual language here actually gets in the way of meaning.  Is there a reason why the shapes are circles?  No.  Do the colours mean anything?  No.  Does the particular arrangement of the circles mean anything?  No.  The only visual meaning I can pick up is that the concept of a Human Resource Management System is in some way superordinate to the other concepts.  But that’s because it’s the title of the list:

Human Resource Management System

  • HR Management
  • Payroll Management
  • Recruitment Management
  • Training Management

Why do we end up with diagrams like this?  It happens because most models in business start out as a series of bullet points on a flipchart during a brainstorm.

management-bullets

If you want to publish it internally or externally it has to look decent, so it goes to a designer.  The designer doesn’t understand the meaning of the content because they weren’t in the session, but they do know how to make it look nice.  You end up with an aesthetically pleasing list, but a list nonetheless.

Here’s another famous model – Maslow’s hierarchy of needs.  It’s usually drawn something like this:

blog-piramid-2-03

Now this is not just a list.   The verticality of the diagram represents levels of dependency in a hierarchy – the upper blocks would not be there without the lower ones to rest on.  This is not the same as a list – if you switched the order here, it would mean something different.  The closest you can get in text is an ordered list:

  1. Physiological needs
  2. Safety needs
  3. Love and belonging needs
  4. Self-esteem needs
  5. Self-actualization needs

But of course the point of the model is not that the elements are discrete, but that they form a chain of dependency – you can’t meet one level of need without satisfying all the needs below it.  Saying that they’re ordinal and putting them into a chronological list ends up visually inverting the order of the pyramid.

The point is, the visual language is instantly conveying meaning to the viewer on an instinctive level that would take a paragraph to explain verbally.

The reason the pyramid works instinctively is because the visual structure matches the concepts in a way that agrees with our experience of the world.  In the world, the downward force of gravity means support always comes from underneath, because objects fall downwards:

blog-piramid-2-02

You sometimes see progressive leadership teams inverting their organisational structure diagrams, because they see themselves (or want to be seen by others) as supporting the organisation, not the other way round:

blog-piramid-2-01

The support metaphor still works because of the implicit force of gravity working downwards.  I’ve always found these diagrams unlikely, not just because gravity makes it look like the organisation is going to fall over, but because another basic meaning of verticality is that power comes from above.

But this is a problem with the concepts, not the visual language.  Whether the pyramid is upside down or the right way up, the visual elements have meaning that’s not conveyed by the text.

So, visual language can convey things that are very hard to illustrate using text alone.  A simple example is the 2×2 matrix, which adds a huge amount of meaning over the corresponding words.  Try explaining a Prisoner’s Dilemma using text alone, and see how much longer it takes:

prisoners-dilemma

Any management consultant worth his or her salt will struggle to get through a day without spotting a 2×2 matrix somewhere.

And this I suppose is the difficulty of mastering visual language:  It’s not enough just to be well versed in design principles.  You need to understand the conceptual structure of the content to be able to match it to the appropriate elements of visual language.  And a really good test of this is to ask yourself strictly, how much meaning do these visual elements add?  If they don’t add very much, then what you’ve really got is an aesthetically pleasing list.

Now let’s finish by looking at some really effective visual language, in one of the most popular business diagrams of the last five years:

business-model-generation-canvas

This is XPlane’s visualisation of Alex Osterwalder’s business model canvas.  If you reduce it to just a list of words like this …

Business model canvas

  • Key Activities
  • Value Proposition
  • Customer Relationships

… look how much meaning is lost!  It’s clear that the visual language is conveying meaning over and above the list of words, for example:

  • The distinction between vertical and horizontal blocks has meaning: Vertical blocks represent delivery elements, whereas horizontal blocks represent financial elements
  • Size of gaps between blocks has meaning: The gap that the arrows fit in represents an organisational separation between producer and consumer
  • Interconnections have meaning: Notice the tabs on the left, which represent collaboration
  • The icons have meaning, acting as a mental shortcut to the concepts they represent.
  • And so on …

So, the Meaning Quotient.  Maybe some day we’ll be able to measure it.  In the meantime, why don’t we ask ourselves the next time we visualise a management model, how much meaning is the visual language actually adding here?  And if it’s not adding anything, was it really a model to begin with, or just a list?