We are currently in the process of researching tools that serve as the basis of our knowledge graphs and decision trees.
A description of the goals and the criteria can be found in the Wiki at: Development:Tools/Knowledge Graphs and Decision Trees - InvestmentWiki
@bdavid is leading the effort.
The purpose of this topic is to discuss goals and criteria and get feedback on tools or even new ideas on which tools might be the most suitable.
Related Messages
Discord: Why we want to break down an investment thesis into its sub-components + Example on Figma - by moritz
Original Discord msg here
Yes we are obviously trying to answer a very complex and hard to answer question here which consists of many different components.
Similarly to how we are breaking down the overall company thesis into subcomponents i think the right approach is breaking down this GSV subcomponent into it’s many parts and analyze them individually. (I tried to sketch what i have in mind here (1) in Figma and started to add esp. some subcomponents for GSV. Figma is the wrong tool for us though as it does not allow for elements to be collapsable which makes it impossible to illustrate all components of a company at once due to space limitations. @MeDavid is currently looking for a tool that we can use)Once we get a good understanding of components we can try to weigh them and think about strongest correlations between those subcomponents and GSV (in this case) or in other words get an understanding of the prediction power of any of those subcomponents.
The overall goal is not to predict GSV with absolut precision but get an understanding to which factors GSV is responding and get a rough estimate how GSV is likely going to develop.
E.g. is it going to slightly grow, be constant, fall 5-10%, fall 20%, fall 30% etc.
If applicable we can even couple those GSV predictions to different macro scenarios (and how probabilities for those scenarios).By getting this rough understanding it will be much easier to think about different how Upwork might likely develop in different scenarios and make increasing or reducing an investment much easier.
As an example if i know GSV and revenue is almost certainly not falling more than 20% in any circumstance i know that the downside risk at this front is limited and can increase the investment.
Description of decision trees: Acceptable and Non-Acceptable Scenarios for Companies - #4 by raff91