Hello @Sinkas.
While involved in LTIPP and STIP, I am convinced that most of the analysis, at least as initial overwiew, should be something like this.
https://twitter.com/ElBarto_Crypto/status/1757036165520036299.
Will attach some screenshots for clarity (not putting the whole thread but only half).
Let me explain. In this topic, Artemis (a firm i don’t know nor i am affiliated to) has conducted an initial study of STIP effect normalized for market conditions. Their measurement was statistic based, we can likely get to similar measures by using points of control.
What does that mean? In a normale environment, you measure normal activity, you measure activity with incentives → you get some sort of quantitative answer.
However, in crypto we have such a volatility, even intra year and intra cycle, that sometimes could be difficult to understand if any meaningful change in metrics depends from the incentive program, the overall crypto market conditions, or both. If the latter (likely), what is the weight of each component.
In other way: we should use other chains and ecosystem as point of control. Quick and dirty example, for which I will use mock numbers and mock proxies:
- if arbitrum, as a chain, growth in tvl by 150%, users by 220%
- if optimism as a chain, in the same time, growth in tvl by 50%, users by 75%
- if, in positive market conditions, so far optimism and arbitrum have shown a 1:2 ratio in growth of tvl, 1:2 ratio in growth of tvl
This means that in the period of december-march, since op growth by 50% in tvl and 75% in users, for arbitrum, of the 150% growth in tvl, only 50% was due to incentives, and for the users, only 70% was due to incentives.
Now, I understand that there are a ton of nuances and caveat above. Is not trivial as a measurement, depends a lot on the control set you use (in this case is OP, and might make sense to use more than 1), there is likely to calculate this beta of metric in different market environment in relation to the total crypto market value and/or l2 total value, etcetera.
But I think this is the starting point: we need to be able to achieve a measurement of how much of the metrics changes are due to incentive programs, and how much by market. And then, we iterate this measure for single protocols.
I think there is a merit in this approach, to have clearer numbers. Let me know if you think it makes sense, also you know where to find me to have clarifications.