Month: November 2022

I discussed the nature of conditions that encourage dandelion or elephant strategies, but I only skimmed a little bit the topic of why one would choose these strategies. Here’s a sketch of digging a bit more in depth of this particular “why”.

A rubric that we want to use here stems from the observation of environments leading to the thriving of the organizations that chose either strategy. The environment must be complementary to the strategy.

Elephants and other K-selected species choose their strategy because the environment makes this strategy viable: there is a well-defined, stable niche into which the overall shape of the species’ activities fits. This niche is usually crowded: there are other species within that niche, and the game of life is mostly about optimizing the shape to vie for bettering other species. In this way, the environment of an elephant strategy is shape-focused.

This helps us define a rule of thumb for choosing to pursue the elephant strategy: if the shape of our technology or product is well-known and well-understood (and thus strongly expected) by our market, the elephant strategy is likely warranted.

The word “shape” here is meant in a broader sense: it’s not necessarily the physical dimensions of some object. Rather, it’s the whole set of expectations and constraints that define the niche in which we’re playing.

Using this definition, we can see that the basic shape of a mobile phone is becoming more and more entrenched in the mindshare of the world’s population. It’s the iconic iPhone shape that Steve Jobs revealed in 2007. There’s an expectation of a touch screen and a camera, of GPS, and various other sensors that are considered table stakes by the users. Should a company decide to ship a phone that does not fit into that shape, it will be playing at a different, less populated table.

Everyone who wants to live in this niche must be engaged in a strongly elephant-leaning strategy: year over year, continuous, incremental improvement of what can fit into the shape defined by the niche. We think of elephants as being slow, and that’s where the metaphor falls a bit short. Pursuing the elephant strategy is all about steadily accumulating mass and gaining velocity through momentum. Maybe thinking about how a capital ship reaches their tremendous speed would help here. It’s the game of consistency and compounding momentum, not leaps of faith or moonshots.

Prioritizing while pursuing elephant strategy, we must ask ourselves these questions: does this work help us accumulate momentum over the long term? Does it contribute to optimizing for the shape in the niche we’re playing? Does it help us better understand the nuanced nooks and crannies of the shape that might be critical for the relentless striving of shape-fitting? 

Having impact in pursuit of elephant strategy means contributing to gaining momentum. If I did something impressive, but not aligned with this larger goal, I probably wasted my energy – or worse yet, slowed down the ship. As I described before, top-down cultures tend to be effective here: it is easier to understand what to align with.

As a thought experiment, compare two teams: in one team, planning happens as a top-down process, and in another – bottom-up. In the top-down team, the direction is clearly stated at the beginning of the planning, and all team members must shape their work to align with this direction. Conversely, in a bottom-up planning process, the members supply what they are planning to do and then the sum vector of this work determines where the team will go. Which team will have a greater consistency of momentum accumulation over time?

On the other hand, dandelions and the r-selected species they represent choose their strategy because the space in front of them is wide open. This typically happens when a technological breakthrough suddenly provides the means to build something new, but the actual product-market fit – the shape of the product or technology – is entirely unclear. There are no right or wrong answers – yet. There aren’t clearly defined constraints and the perceived limits seem to give when pushed on. Everyone else is doing roughly the same thing: stumbling around to find out what the heck this new space is all about.

To turn it into a rule of thumb for choosing the dandelion strategy: if the technology or product space is ill-defined and little-understood, pick the dandelion strategy.

I already used the AI-generated media space as the currently-unfolding example of this scenario, and it’s still very relevant. It is not knowable where this thing will go. Early attempts to discern constraints or define a shape will likely look foolish in the long run.

It is fairly easy to see how trying to build a new dreadnought or turn the existing ship toward this momentum are both activities fraught with peril. When we don’t know where dandelion seeds will land, we are better off letting go of our habit of predicting the outcomes.

Instead of investing into enterprise-strength, scalable software, we can be better off with duct tape and popsicle sticks when adopting the dandelion strategy. Throw something together, make it go, ship it, and see what happens. Do it again. And again. Watch carefully for surprising developments. If our tiny app suddenly gains a small following – that’s something. Avoid presuming that we know what this “something” is. Very often, in a new space, even when given an opportunity to tell us directly, our users might not be able to articulate it, either. Revealed preferences rule over the stated ones. The trick is to keep trying and learning and making sure that the learning propagates to some common pool of wisdom.

In such an environment, bottom-up cultures work amazingly well. Returning to the earlier thought experiment, can you see how the fortunes of these teams will be reversed? The top-down team will form a tight fist, punch hard in one direction … to never be heard from again. In contrast, the bottom-up team will diverge in a cloud of divergent direction vectors, thus maximizing their chances of stumbling onto a fertile niche.

Here, being impactful means uncovering something interesting and surprising as quickly as possible – and bringing it back to the team. Trying and failing is just as useful, because it uncovers where not to go, or go about differently. The key difference between measuring the impact for the elephant strategy is in contributing to the common pool of knowledge while exploring a direction that’s different from the rest of the team. This can feel rather unintuitive: members reinforcing each other’s approaches can be a source of blindness when applying the dandelion strategy. The way to structure incentives here is to emphasize individual agency while rewarding contributions to collective knowledge of the space.

How does a bottom-up team prioritize? In an uncertain environment, prioritization is emergent. There aren’t well-defined metrics and clear lines to cut. Instead, the team’s stumbling into novelty is the source of knowing what’s important, leading to recurring waves of swarming and scattering. This may feel rather mercurial and drive some engineers and program managers nuts. The trick here is to zero in not on the ever-moving objects of prioritization, but rather on whether the information about these objects is flowing as quickly and clearly as possible.

To summarize these two rules of thumb, I will bring them together into a rubric. Ask these two questions:

1. Is the shape of the product/technology niche in which we are playing well-understood? 
2. Are we playing in (or closely adjacent to) the space that just opened up because of a recent technological breakthrough?

The answers form a 2×2.

For the environments where the shape is well-known, with no new space opening up, we are looking at a strongly elephant-leaning strategy. Get that colossal ship going and keep it rolling. Don’t get discouraged when first outcomes are unsatisfying. Elephant calves need nursing and care.

If the answer to both questions is “yes”, we’re probably seeing increasing potential for a new product category in a previously-stable space. Something curious will happen soon, and we don’t want to miss it. Deploy the “fuzzy elephant” stance: structure most of the team to adopt the elephant-leaning strategy, with a modest-sized group making dandelion moves. Given the recent rate of technological advance, this is an effective posture for any large player: there will always be an opportunity of surprise.

The full-on dandelion approach is warranted in the presence of a technological breakthrough in a brand new area with few well-defined niches. Manage divergence and get that insight flow going – who will be the first player to spot a niche?

The final quadrant is a bit puzzling for me. If the shape is not known and there aren’t any breakthroughs, why are we playing here? It feels like there might not be enough evolutionary pressure to get the selection process going – which means that if we find ourselves in this space, we are better off looking for a way out.

I’ve been talking about dandelions and elephants for a while now, and yes, it may seem like I’ve gone a bit nuts. Oh well. It’s just that it’s such a good framing and I keep finding uses for it nearly every day. When applied to ideas, r/K-selection strategies seem to be uncommonly generative.

It all begins with a question: what kind of new ideas do we want to produce? Do we want a collection of different, independent ideas or do we want each idea to improve upon some larger idea?

What I like about these questions is that they are objective-agnostic. They don’t ask “what do you want to achieve?” or “where do you want to go?” Instead, they require us to choose the means to generate ideas. And strategy is all about the means. In the field where I work, strategy is also about generating new ideas.

Here’s the thing. In software engineering (as likely in many technology fields), more often than not, we don’t know what the path to our objective will look like. Heck, most of the time we don’t even have a clear sense of what the objective will look like. This is assuredly not a “let’s plan all steps in advance” process. The fog of uncertainty is right there in front of us. 

If we are to navigate toward it, we must be prepared to shift course, to adjust, to learn on the spot about the next step, make it, learn again, and so on. And to do this well, we need new ideas. Our strategy must count on us continuously producing these new ideas – and applying them. In this way, my ramblings about dandelions and elephants aren’t fun side metaphors. They are the essence of business.

Summoning my inner Rumelt and putting things perhaps overly bluntly, an organization can only be effective at setting a strategy and actually following through when it is intentional about creating conditions for generating ideas. While it’s not the only crucial ingredient, the organization that doesn’t have it will suffer from strategic incoherence.

A team may accept as a truism that bottom-up cultures are superior to top-down cultures. And yes, if we are setting out to explore a large space of unknown untapped potential, then we probably want to create conditions for a dandelion strategy. The bottom-up culture has them: individual incentives (Interest), small teams, short-term objectives (Legibility), independent decision-making (Velocity) and non-hierarchical structure and mobility (Access).

However, when we’re endeavoring to care for one big idea, we likely want the conditions to encourage the elephant strategy: more structured and predictable organization and incentives (Stability), care and accountability in decision-making (Breadth), comprehensive processes and long-term thinking (Rigor), and concentrated points of organizational control (Power). These are a depiction of the top-down culture.

If we set out to do something that calls for an elephant strategy, yet the culture we have is a bottom-up one, we will have strategically incoherent outcomes (I called them the “pappus elephants” in the previous post). Our bottom-up culture will suddenly snag us like a trap, with coordination headwinds becoming universally felt and recognized. Things that worked really well for us before, like emphasizing individual impact in our incentive structures, will become a source of pain: why are our teammates acting in such a self-interested way?! Well… maybe because that was a good thing when we needed a dandelion strategy?

Even when the need to pursue a multi-year objective becomes existential, the dandelion conditions will keep blowing us off course: multi-year ideas will be simply swept away by the churn of the quarterly objective-setting and obsessive focus on individual impact. In a dandelion culture, when given a chance to make a dandelion move, most folks will take it. When strategy is incoherent, one can be a superstar while directly contributing to the team’s demise. 

Perhaps even more bizarrely, by all accounts of witnesses, these efforts will look like elephants – until they disappear in a puff. It is in everyone’s interest to create a perception that they are indeed operating in an elephant factory, despite all the dandelion moves they are making. 

When caught in this condition inconsistency, the long-term projects within this organization will inevitably find themselves in a weird cycle: set out to do big things, fail to articulate them clearly, struggle to do something very ambitious, get distracted, then quietly discontinue the effort, unable to examine what happened due to the deep sense of shame that follows – only to try again soon thereafter. When underlying conditions allow only dandelion-like moves, trying to choose an elephant strategy is a tough proposition.

The variables and symptoms might vary, but the equation will remain the same. If they sound at all familiar, consider asking different questions to get to a more productive conversation about incentives, culture, structure, and practices. What are our current conditions for generating new ideas? Do they lean dandelion or elephant? How might they be inconsistent with our desired outcomes?

As a sort of thought appetizer, here’s a vignette attempting to intersect the divergent/convergent thinking frame and … you guessed it, dandelions and elephants.

While exploring idea generation strategies, I realized that I’ve confused the opposite of divergent thinking with the lack of new ideas. That does not seem to be the case. There is a really simple 2×2 that helps illustrate that. Horizontally, we have dandelions and elephants, and vertically – the rate of new ideas that we want to encourage. Some organizations, for example, don’t want any new ideas altogether (low rate), while others have an existential need for new ideas (high rate).

The upper quadrants make sense: we have our dandelion-fueled exploration and improvements to existing ideas through the elephant strategy. The bottom-right quadrant shows up as orthodoxy – when new ideas are unwanted because we believe that there’s already one idea that is understood as well as it could be. The final, bottom-left quadrant is that of the idea desert: the absence of ideas altogether. This quadrant rarely has permanent occupants, because we humans tend to have ideas – however, we might occasionally visit it when we’re lost and disoriented.

If I were to explore how an organization might move from quadrant to quadrant, there appear to be four clockwise motions, from one quadrant to another. When our process of ⛵exploration (the top-left quadrant) yields many interesting ideas, organizations tend to shift into the convergent posture: start looking for one idea that will become bigger than others. A typical call here is “more wood behind fewer arrows” and conversations about prioritization and hard decisions. Convergence leads to the top-right quadrant of 📈 improvements, where most of our efforts are invested into reinforcing the idea.

Thinking in this quadrant does not need to be boring and uncreative. For example, one of my colleagues, upon joining our team, spent a bunch of time looking through the code and spotted a bug: in handling mouse events, we would sometimes hit an O(n²) condition – and that was unnecessary. Their first commit to the codebase was a performance breakthrough. When I think of innovation in the top-right quadrant, I keep recalling that bit of code. It was clearly a novel insight, yet it also demonstrably improved the state of the bigger idea.

Unfortunately, our residence in this quadrant usually comes to an end due to the consolidation move. In an effort to better protect the value contained within the big idea, organizations usually shift to the quadrant of ☝️orthodoxy, where any new ideas are viewed as mostly distractions. There’s so much stuff to do already! The problem lists are long and the issues are well-known. Let’s just keep on fixing them, shall we?

As you may suspect from the flow of this story, the discouraging of new ideas eventually triggers the next move: obsolescence. This move puts us into the 🏜 idea desert. We know that the old big idea no longer works, but we have nothing else to hang onto.  The discomfort of this quadrant acts a powerful motivator, manifesting as the divergence move that propels us back into the  ⛵exploration quadrant. This move might coincide with the demise of the organization, when the divergence acts as a pulling-apart force of ideas, each stakeholder pursuing their own.

If we are to believe this tall tale, we can see that divergence and convergence are somewhat orthogonal moves. The divergence is mostly about moving from zero to many ideas while relying on dandelion strategies. 

On the other hand, the convergence is a qualitative shift. It does not change how much idea-generation we do, but rather whether or not these ideas are independent from each other. A productive convergence is the one where people ideate while building on each other’s ideas, improving upon one unifying big idea – the elephant strategy.

This might not be news to you at all, but this was a pretty useful insight for me. When planning the next diverge/converge exercise:

• To get divergence truly going, put people in the idea desert. Let them have the sense that all of their previous, strongly-held beliefs might not be as sound and safe as they seem. This might mean not letting people prepare for the exercise or even creating an idea parking lot that is filled at the beginning with the ideas that we already have.
• Recognize the creative part of convergent thinking. Let people continue ideate, but shift the constraints of the exercise to encourage the clumping of ideas together. Avoid prematurely collapsing the process straight toward the orthodoxy. 

Convergence doesn’t have to be a boring prioritize-and-cut procedure. It can be fun – and who knows, maybe produce new ideas that didn’t pop up during the divergence part of the exercise?