One model I find myself coming back to on an ongoing basis is the OODA loop.
OODA loops are the agentic cycle between Observation, Orientation, Decision, & Action. This below graphic explains it very well.
OODA Loops are important for navigating a design space where the fog of war is thick or the environment is complicated. OODA Loops = Agility. You can win over raw strength with high agility, so OODA loops are important in startups!
By virtue of their sheer size, big startups can compete on the effect of each action. Each feature release can be very impactful. Small startups dont have the network effects or resources to make each action impactful. But small startups can compete on agility by spinning the OODA Loop faster to see/execute opportunities that slower rivals don’t.
Back when I was working on Gitcoin Grants, we were in an OODA Loop in which we were testing the best way to solve problems with quadratic funding.
The problems we faced on Gitcoin Grants varied:
- The site is going down .
- We are getting sybil attacked.
- The conversion flow doesnt work in some wallet/browser/os combo.
- We are testing a new UI sorting mechanism and whether it affects user behaviour by persuading their grant contributions (answer: it does.
- This thread has details of 12+ OODA loops we went through…
Some of the meta-OODA-fu we would employ on these problems:
Faster OODA Loops = faster/more action = faster exploration of a design space
One of the techniques that I like to employ is to try and tighten my OODA Loops. Shorter OODA Loops = Faster Iteration. For example, the sybil attack problem was one that we could only really study during 2 weeks of the quarterly grants round. The other 10/12 weeks of the quarter couldnt be spent on that. So what I did was deploy a Quadratic Funding Social Network that operated as a honey pot for the sybil attackers, but the rounds were on a weekly basis instead of quarterly. This meant each quarter had 12 OODA Loops in it instead of 1 OODA loop!
Bettter Research = better orientation
Better Data = better decisions
During this sybil resistance OODA loop, I discovered a paper called Who watches the watchmen By Santi Siri + Paula Berman, which turned out to be a holy grail telling me the state of the art techniques to study sybil attackers on the system.
Once I had studied the design space and oreinted myself, we employed data collection & analytics to collect data about the sybil attackers.
Another thing we did to better orient ourselves was bring in the experts, we hired BlockScience to build an rigorous anti-sybil process.
Bias towards action == faster decisions
One of the mantras that I hold dear from Techstars is have a bias towards action. As entrepreneurs, we learn by doing. You learn infinitely more than if you think about doing it in the abstract.
One of the hardest things to tackle when trying to have a bias towards action is making decisions with incomplete information.
The other hard thing is coordinating all the different parts of your team to rally around “act” instead of “be lasseiz-faire about it”. This is an art more than a science - one part part elbow grease, one part purpose, one part coordination, and one persuasion. Some times people will rally around you to act, other times you will have to go it alone + act unilaterally if coordination fails.
During the Quadratic Funding Social Network experiment, we used to deploy to production several times per day to fix problems + gather data faster.