top of page
Post: Blog2_Post
  • Writer's pictureJohn Graham

How to Build a Learning Organization: Step 5 - Thought Leadership

Where are We

In the last post, we used the skill-will matrix (the Situational Leadership Model) and the Dreyfus Model of Skill Acquisition to help provide guideposts on different parts of a learning organization.

I hope to cover today the last bit of this effort. We're calling this Thought Leadership.

If I'm successful, you should come away from this with:

  • An idea of what Thought Leadership is

  • How to build people into influential Thought Leaders

  • How Thought Leaders discover new ideas

  • How an organization can enable and encourage this

We're defining a learning organization as basically an organization that learns. A good chunk of the early steps ensures that you have an organization that teaches.

It's a colossal waste of effort for people to learn something they could have been taught. When there are no resources or curricula, one of the few ways you can learn is through trial and error. This is also one of the most expensive ways to learn.

We need to teach so we have as many people as possible pushing the boundaries. If much of your organization's energy is wasted learning the things it already technically knew, there's not much left over to learn something you didn't know.

And when we talk about Thought Leadership, we're talking about learning things no one knows.

It's not unlike this model of how a Ph.D. works:

Still, we're at the phase of "what got you here won't get you there." We're shifting gears; and we've been steadily shifting gears throughout the entire climb of the Dreyfus Model of skill acquisition. This is the final and top gear - learning and discovering when you've got no prior art to inspire you or work to reference.

We'll mostly be talking about models here. The business advice version of this post is to ensure you have an organization that supports these kinds of models.

For the most part, this means ensuring your top-level experts - the ones you've put through the entire skill-will process - have the autonomy to push the boundaries how they see fit. They need to understand the Why (the Purpose) of where the organization moves.

Convergent and Divergent Thinking

Most of the early tools we've talked about, like training and mentoring, work through convergent thinking. Convergent thinking says that there's a known way to get from the problem to the solution, and all it takes is a little bit of effort and practice to do so.

The OODA loop is one such process. You need to take in data, make sense of it, make some decisions, then execute. The models, the decisions, or "moves" you might take from that sense-making usually draw from prior work.

The higher we get in the Dreyfus Model, the more we rely on divergent thinking skills. There is no known path from problem to solution. So it's a different toolbox and a different set of skills. It's also much slower.

We started being more divergent with our Coaching advice. Here, the coach didn't know the solution; it was assumed no one did. But we showed there are still ways to ensure the "solution machine" - the person trying to solve the problem - was in good working order. We called this process Coaching.

Then as we grew Experts, we needed to ensure they continued to have more and more experience and authority. We're mainly banking on just plain insight at this point. There may be convergent ways of thinking left or may not be. But we put someone with as much knowledge of the problem on the problem, then gave them the emotional support they needed. Now we just wait.

We have an excellent model of divergent thinking in general, though. And it gives us some strategies to use while tackling problems with no known solution. That model is evolution.

I'm not one of them fancy scientist types

So my knowledge of evolution beyond the high school level is sketchy; and I do not want to imply that what I'm about to discuss is how evolution works. Instead, I'm more familiar with genetic algorithms and their own inspiration from evolution. From genetic algorithms, there are three strategies to draw from.


Mutation is the first and most widely known evolutionary problem-solving method. This is akin to trial and error, as we mentioned above. We take a known solution, apply a few tweaks, then see what happens.

How does mutation apply to a learning organization?

Here, we'd advise a risk-taking culture - especially a meritocratic risk-taking culture. In other words, as people prove competence in concrete ways (taking training, having happy mentors, gaining experience), they get more autonomy and more direction to take calculated risks.

But this doesn't have to apply just to senior leaders. Mutation also happens when information fails to "copy" in the right way. In other words, there's going to be a rate of misunderstanding coming out of your training and mentoring programs. People aren't always going to "get" things the first time around.

And veeeery occasionally what they'll produce isn't just wrong - it's actually more right than right. These mistakes may accidentally improve the process!

Keep an eye out for these happy accidents. Even if you want people to have proven their efforts before taking risks, that only mitigates risks. There will be risks that can't be controlled daily, or people aren't even aware they're taking them. And these risks may turn up something - rarely - that's worth saving.


There are millions of microbes in your gut right now exchanging genetic information. This is called "conjugation." In genetic algorithms, I was taught it was called crossover.

Crossover is when you take two known reasonable solutions and attempt to combine them. This can be faster than mutation because good solutions often work differently and synergize when combined.

We talked about one kind of crossover in our coaching blog. We pointed out that certain common mental models appear all over. These models are akin to bits of genetic code that are very useful to "crossover" into the solution at hand.

How do you build an organization that enables crossover?

First, ensure that people can break down power hierarchies and silos. People with very different skillsets should be able to talk to each other easily. Your most essential innovators will be a team of experts, all with expertise in different things.

They're not just using this expertise to fill in each other's gaps to build a system from soup to nuts but also inspire new ideas within each domain.

You can continue to emphasize crossover by encouraging your experts to build wider and wider expertise. Learn about things slightly outside of their field and things entirely outside of their area.

This gives them more and more genetic information to draw from when they attempt to do crossover and come up with novel solutions.

This won't be the last time you'll hear philosophy pitched on this blog as one of the best uses of time; but you're basically attempting to build little Philosophers of X, where X is their job in your company in your industry. Both wide and narrow: you need people who just love to think, learn, do a particular thing, and think about how that specific little thing ought to be done.

As you get up in the Dreyfus Model, the "returns" to mastery - or that boost you get to productivity when you learn something new - start to slow down. You need people who are okay just learning things to know things because you have no idea when the next breakthrough will happen.

Philosophy has given birth to many disciplines and sciences, from physics itself (at one time called natural philosophy) to, more recently, anthropology and sociology. The broader you go into mental models that "apply nearly anywhere," the more you will get into the realm of philosophy anyway.

Do you need some new Management ideas to crossover to your company culture and process? Contact Us!

Sexual Selection

Now, we already have a crossover model. What could two models pairing up to produce offspring do more than that? These microbes just share and recombine genetic material by chance in our gut. So it's a little bit reuse, a little bit of mutation.

Sexual selection attempts to add some sort of intelligence to that.

For sexual selection to work, each member of a species needs to have a model in their head about what they think is fit and what they believe is not fit. We call that attractiveness. Then they probabilistically are more likely to mate with whom they believe is suitable (and who thinks they are ideal in return) than those who are unfit.

This can speed progress as long as the model/attractiveness is reasonably good.

We've got a simpler model trying to predict a more complicated model. The simpler model will capture the salient details and speed us up in many cases.

You've got to practice building models in a learning organization and ensure your experts are great model builders.

They need to, for instance, be very good at estimation. These Thought Leaders need to be able to take an idea and roll it out in their heads as fast as possible to see the upsides and downsides. They may even need to be superforecasters.

They need to be very good at prototyping - getting from concept to "yes, this will work" or "no, this won't work" quickly. This will allow experts to "make mistakes quicker." After all, we're still relying on trial and error/mutation; we're just trying to make the mutation in our heads and see the consequences.

Note: Woe be unto you who allows someone to run amok prototyping who hasn't demonstrated a Proficient level of testing, peer review, and quality.

An organization can ensure experts have the time and resources for the above by loosening or redefining specific goals. You don't want to hold your Thought Leaders to hard goals such as "Will solve the X problem by Date." If X is a complex problem, you don't even know if it's solvable. And if X isn't a complex problem, then it doesn't require an Expert; you should give it to a Proficient so they can get experience.

Instead, you want to look for an improving ability to put out MVPs, prototypes, estimations, and forecasts.

Finally, you need to allow for works of art.

A work of art may be overdone, in contrast to prototypes or estimates that are killed off when it's realized they won't work. A work of art may not be a great idea from a business sense but is still pursued.

You can see this in nature, where basically, sexual selection goes off the rails and appears to just build out entire traits that have little or no fitness value. I know some argue these traits are "signs of health," but we can just pretend they're works of art for our uses.

Works of art often become talking points at conferences, a matter of pride for the contributors, and can often be appropriated for very useful business purposes once they're done and better understood.

At the highest levels of expertise, there is no bottom-line value. There are only works of art that later are realized to have unpredicted and high bottom-line value. You climb the mountain because it's there. And once you've done it, you realize, boy, that place needs a hot dog stand halfway up.

Other Methods

I want to mention a few more techniques that are also loosely based on the evolutionary model. I tried to draw them out, though, as they're a bit more concrete and actionable in and of themselves.

Map out the domains and subdomains

From our training blog to our delegation blog, we implied that you need to have clear, concrete paths to navigating the domains your business needs to survive.

We'll cover this more in-depth in a later article. We call this a Badge System. It's the careful analysis of the skills people at your company need to succeed, paired with training, mentoring, and other support to teach those skills.

We've mentioned how this Badge System needs to align with your strategy. Resources will flow into Badges (or domains of knowledge) in which you think you'll be most competitive. In contrast, other Badges may just get table stakes or be left to partners.

But the Badge System also serves as a map. Employees can easily self-guide their own training by having a layout of all possible skills sets to learn and easy links to get started.

This map is helpful for thought leaders in two ways. First, they can continue to use the map to grow a wider and wider base. This helps with crossover. Second, they can play with the map itself, arranging and rearranging domains and potentially discovering new ones that are missing. This would be an example of model-driven sexual selection. They can attempt to make the map "less ugly" and, through that work, potentially discover new approaches.

Wider Rotations Across Skillsets

We've talked about rotations becoming a valuable tool as you go up the chain - again, building a wider base.

Here, you want to focus rotations on things that may not be related to your company or industry. This might rely on sending people to conferences where you'd be complete neophytes or even having people at your organization pursue certifications, masters, or Ph.D. programs in skills you don't anticipate needing.

Because you just might use it once you have it.

Presentations at Conferences

Presentations make an excellent concrete milestone for prototypes, works of art, and other artifacts experts might create while attempting to discover some new business angle.

It also builds credibility in the industry, attracts talent to your company, and provides a simmering cauldron of crossover ideas.

Avoid the No Patrol

Hidden in all of this is yet another layer of anti-gatekeeping efforts.

Suppose you have people astonishingly good at estimations and guesswork, at seeing the future, who are your go-to people when churning out a prototype. In that case, that group may begin to identify themselves as the only people allowed to prototype.

They may, ironically, make the rest of the company more risk-averse by claiming that risks are only theirs to take.

Nothing in this blog should be construed as empowering only your Thought Leaders as the people able to take risks. Instead, only folks capable of taking calculated risks should have the resources dedicated to them to be pushed into a Thought Leadership role.

You need to keep your thought leaders mentoring your earlier Experts and late Proficients. As we've said multiple times, teaching is one of the best ways to learn. And since your Thought Leaders are often the only ones who might know a thing, possibly in the entire industry, they and they alone can teach certain things.

This should, hopefully, keep them in the trenches enough with their colleagues, so they don't become the No Patrol. They should not employ their expert power to shoot down any idea that wasn't theirs.

Additionally, though, people management must ensure that Thought Leaders remain aware of their own growth. One of the only reasons they grew into Thought Leaders was that they were allowed to take risks to make mistakes. It's a Thought Leader's job - inherited from their duties to delegate - to think about future Thought Leaders. Part of that is ensuring those up-and-comers are not only allowed to make inevitable mistakes but encouraged to take calculated risks.


Thought Leadership is having new ideas - new in the industry, possibly the world. And simply having those new ideas isn't enough; it's about teaching the people in your organization these new ideas. This is yet another function of a Learning (and teaching) Organization.

Thought Leadership comes at the end of the Dreyfus Model of skill acquisition. This is so you don't waste effort reinventing the wheel or "learning" what could be taught.

New ideas for thought leaders are discovered primarily through divergent thinking. A great model for divergent thinking is genetic algorithms/evolution:

  • Mutation (take calculated risks)

  • Crossover (apply old ideas to new problems)

  • Sexual selection (have an idea of what is "beautiful" so you can prioritize your efforts)

Failing fast through forecasts and prototypes, or bypassing failure through works of art, are all techniques to grow knowledge via the genetic algorithm method.

We're Guildmaster Consulting. We're trying to be Thought Leaders in the software management space. Should we prune this branch, or is it a work of art? Let us know!


28 views0 comments

Recent Posts

See All

Will AI Phase Out Software Developement Roles?

John recently sat down with Daniel Litwin at MarketScale to talk about AI. AI won't change programming in the way you think... Check it out here. Key Takeaways: Junior devs need to treat AI as a doubl


Soapbox Logo (Transparent).png
bottom of page