Week of 2023-01-30
Where I learn about the tension (and biases created by this tension) between two qualities that businesses often find desirable: growth and control.
The tension between growth and control
I alluded to this tension before, but it was somewhat buried in a fairly obscure context of compounding loops. Here’s a somewhat unkempt riff on the concept.
When introducing a product into the hands of our customers, we will nearly always face the tension between growth and control. Here, growth reflects the number of people who choose to use our product, and control is the degree to which we get to decide what this product will be and how it might be used.
Basically, think of these as two extreme ends of a line that can never move close to each other. We can travel along the line with our product strategy, but never bend it to make control and growth meet. We can either lean toward growth and thus relinquish more control, or we can lean toward control and thus give up on more growth.
Interestingly, the quests for growth and control both stem from the same intention toward value. We seek growth when we are looking to acquire value, and we seek control when we seek to protect it. And almost always, we want to have both: we are intent on holding on to some value and are looking to acquire more. Truly blessed are those who don’t have attachments: they are blissfully unaware of this tension. If one has nothing to lose, they have no need to grow. For the rest of us, the growth/control tension is something we experience every day.
For example, if we’re shipping an API, we usually want its usage to go up (growth) and we want developers who use this API to only use it the way we want to (control). When we imagine a new service, we want it to be enjoyed by as many users as possible (growth), yet we also have very firm ideas on what we want to build (control).
Often, control isn’t something we have due to value we already hold, but rather due to a debt we carry. Writing software or making hardware takes time. In the gap between us starting the project and it becoming available to our customers, we crystallize our ideas into code and circuitry. Think of this crystallization as a form of control: we make decisions about what the product should be, and by making these decisions we exercise control over its shape.
This crystallization is also an unrealized value – we won’t start accruing this value until the product is released to the users. When we do, many of these decisions are somewhat irreversible: even if we want to relinquish control and adapt our product to stimulate its growth, we can’t do so easily. Especially with hardware, a single decision made somewhere in depth of the multi-year gap between ideation and shipping can severely limit a product's capacity to grow.
The outcome of growth has a similar trappy quality. If our product is blessed with exponential growth, after reaping its benefits for a while, we will eventually arrive at the weirdest situation. No matter how much we try, we can’t do anything but optimize for what is already expected by our massive user base. Any attempt at deviation triggers the allergic reaction of irate customers: “why are you moving my cheese and how soon can you put it back?” In this trap of successful growth, control is no longer ours, no matter how much we try to wrestle for it.
🔗 https://glazkov.com/2023/01/30/the-tension-between-growth-and-control/
Growth and control biases
When we don’t recognize that growth and control are in tension with each other, we tend to develop a bias that prevents us from spotting problems with the decisions we make. To better make sense of how this bias shows up – and how to account for it – I will rely on the dandelions and elephants framing. As you may recall, the whole point of dandelions is growth, and elephants are all about holding on to value in a well-defined niche.
The more dandelion-like the environment, the more likely we are to exhibit bias toward control, having higher than warranted confidence in our ability to predict the shape of the product niche that is yet to be discovered. This bias is entrenched in the usual “brilliant engineer/scientist” stereotype: the plucky hero who can see the solution way before anyone else. My experience is that these brilliant minds are narratives: they are stories of the lucky survivors whose being in the right place at the right time is reinterpreted – and entrenched in myth – as some superhuman ability.
In a dandelion environment, everyone is just stumbling around to find out. The less firmly we hold on to our conceptions about the final form of the product, the more we invite serendipity. The more optionality we embed into our own stumbling around, the more likely we are to hit the vein of something interesting. The early pivots of products that led to Flickr and Slack are great examples of doing this well. Poor Stewart Butterfield. He keeps trying to build a video game, and these successful businesses pop out instead.
One typical pattern emerging from the control bias in a dandelion environment is placing early big bets. When we see a large organization spun up for a focus area whose shape is not fully known, we are likely seeing the “control as cost” scenario: the mere size of the organization and the need to justify its existence will prevent it from embracing serendipity. The book Showstopper! is a great read for an inside story of the suffering this anti-pattern causes. I call these kinds of products “elephantine dandelions” – they are simply too big to be agile among other dandelions in the wind of change, yet they continue to exist, buoyed by their size.
Conversely, the more elephant-like our environment, the more likely we are to bias toward growth, presuming that it is something we can simply rely on. There are a couple of patterns here that I am familiar with. One shows up as a belief that continued growth is the organization’s birthright – despite the fact that most of the effort is expended eking out the last bits of value under the curve in the well-defined niche. Thankfully, Clayton Christensen’s concept of disruptive innovation provides an antidote to this bias – though given that the pattern keeps repeating across all industries, I am not sure how effective this antidote is.
Another pattern of growth bias is what I call the “never-ending long tail.” Every so often, a new contender comes up to disrupt a well-established niche, with the tantalizing promise of breaking through the mold of the well-optimized niche. In such cases, it’s very important to watch for the long tail of consumer expectations: are they still there? When the newcomer shows promise and inspires a compelling new vision of what’s possible, are they playing in an entirely new space, or are they still on the hook to deliver on all existing, mostly implicit contracts with the customer?
A good example here is the emergence of new Web rendering engines. I have seen a couple of announcements crop up here and there, and I am totally rooting for them as a fellow engineer, but as a strategist, I have very little hope that they’ll survive under the crushing weight of all the quirks and barnacles that a modern Web rendering engine must support. After initial success and admiration of colleagues (“look! it’s so much faster than others!”), a never-ending long tail of conforming to the same shape within the same niche awaits them.