The 0 to 1 startup manifesto: the entropy version

Two systems at different temperatures eventually attain the same temperature when exposed to each other. In doing so, the systems approach a state of thermal equilibrium. This movement toward equilibrium is spontaneous and continuous.

Sometimes, however, we may not desire equilibrium if the groups are set up for different objectives. The contrast between them is essential to their individual success. A scenario where such a contrast is crucial is when an organization has different groups assigned with the task of taking something from 0 to 1 and from 1 to n. Nothing exists and we create something new that works — that is 0 to 1; something already exists, is known to work, and we make it work on a bigger scale — that is 1 to n. 0 to 1 could be an innovation unit, an early stage startup, or a Horizon 3 business; 1 to n could be the cash cow, the core business.

On the 0 to 1 path we may not yet know what works and are in search of the repeatable and scalable. So we want to try things out quickly until we can settle upon something (business model, product-market fit). Along 1 to n, we have a fair idea of what is repeatable and scalable and any change is a potential diversion from a carefully chosen path and hence is subjected to a set of checks and balances first, sacrificing speed for (known) correctness. 0 to 1 has to stay hot; orderliness introduced too soon cools things down, slows down the rate of iteration. 1 to n, on the other hand, has to consistently expend energy to maintain order.

What kind of setup will increase the chances of failure for a 0 to 1?

Entropy of a system is the measure of energy in it that is unavailable for doing work. Entropy is thus a measure of disorderliness, of volatility, of chaos. Isolated, a system’s entropy can only increase — the second law of thermodynamics. The argument I try to make here is that disorder can be synonymous with fluidity in a 0 to 1 setup and that a measure of it is conducive to the process of turning new ideas into tangible things.

Success in the early stage is about testing out the business hypothesis: Is this opportunity big enough and Am I measuring it right? The simplicity of work at this stage is that these are just two questions but they are all that matter. Yet, fluidity is essential to answering these questions well. If answers are jumped to too soon, the incentive to run cycles of build-test-fail and learn from them is suppressed. Factors that tend to introduce an early organizing effect at the expense of this fluidity slow down business growth. This kind of organization typically manifests itself in the way a 0 to 1 chooses to do things (solve problems, make decisions); the boundaries we choose to respect (hierarchy, process); the mindset we nurture (inner captain vs role-bound), and the incentives that drive principals and agents (ownership vs compensation).

An overarching point in considering a 0 to 1 is that it is essentially a lab for running business experiments. Running business experiments is different from running a business. So it is not just important to fail; it is inevitable. But doing things wrong is not enough; doing things wrong the right way is what differentiates. Those on the inside must therefore be incentivized enough to fail right, not simply act out of self-interest. Yet, business failure attracts attention. It tests patience. It searches for causal pegs to hang from. To run a laboratory for failure means to fight against our very nature.

How can we maintain the fundamental character of a 0 to 1 needed for its optimal functioning?

Just as two groups move toward equilibrium when there is agreement about the optimal way of working, they need to be isolated from each other when a state of equilibrium is not desirable. This can be done by:

1. Reducing their exposure to each other: Introduce a wall. A wall limits exchange, which helps maintain important (ideological) differences between 0 to 1 and 1 to n.
2. Reducing the need for exposure to each other: Have a different set of incentives for the 0 to 1 group. A startup setting demands more accountability than a mature business does. People should not feel like they’re simply renting out their time, but are investing in something for better (> 1X) returns.

Like it is with all non-spontaneous processes, maintaining a separation to keep systems separate comes at the cost of energy. A wall is hard to maintain. It requires dedicated personnel. It needs loyalty and incentives to stop detours. It requires a balance between long-term thinking and response to short-lived opportunities. And sometimes holes need to be poked in it to make it, well, less of a wall. When Alexa began to take off, the team realized that it needed more hands on deck. The team could have spent all their time recruiting, but instead they focused on building Alexa. This was possible because Amazon stepped in. It offered anyone taking up a position anywhere at Amazon the option of joining Alexa. Alexa’s recruitment engine suddenly was on steroids.

From startup to scaleup

Thus far I have spoken about the value of entropy, listed a set of factors that were conducive to maintaining a measure of it, and outlined how we may go about harnessing it for business success. So while I have championed staying loose and nimble, it is worth reiterating that fluidity in itself is not a virtue; it is what it allows a 0 to 1 to do that is coveted in business.

Sensitivity is responsiveness to stimuli. A young startup should score high on sensitivity. It ought to design itself to pick up signals and act on them iteratively. New experiences help add to or fine-tune a set of first principles. As systems mature and codify a set of principles based on well-evaluated experiences, they appear to move closer to a steady state. It is this journey that is the rite of passage from 0 to 1 to 1 to n.

The reason to make this journey is simple: predictability. That is predictability in annualized growth, customer acquisition, brand value, and profitability. Arriving at your own playbook for predictability is what takes time. Yet the epilogue in that playbook is never written. There will always be unforeseen stressors threatening to shake well-set things up. [Insert the over-wrought example of the coronavirus.]

That is when resilience comes into play. Resilience is bounce-backability. Successfully moving from 0 to 1 to 1 to n and continuing to grow to n is about learning to stay sensitive while becoming resilient. Startups that don’t take shortcuts during the build-test-fail phase are more likely to sign up for early access to resilience. This testing is not just of the product, but of business model, of culture. Future scaling is built on the hard lessons learnt, and a good way to learn is by allowing some entropy early on.

Thank you Atul Sinha for reading drafts of this and for sharing your thoughts.