Some additional considerations, upon reading comments & thinking:
Buyback or not?
It appears that Fluid is spending roughly the same amount of incentives as it generates revenue. As such, the protocol is not yet profitable, and the buyback program cannot be considered as a sustainable way to return value to token holders. Instead, it should be viewed as a growth strategy. Here are some scenarios to be evaluated:
- (Positive) Buyback ~> FDV ↑ ~> Incentives ↑ ~> TVL ↑ ~> Revenue ↑
- (Positive) Buyback ~> FDV ↑ ~> # of holders ↑ ~> Narrative & Distribution ↑ ~> TVL ↑ ~> Revenue ↑
- (Negative) Buyback ~> FDV ↓ (for some reason) ~> Treasury ↓ & Incentives ↓
It is not clear if benefits of 1 & 2 outweigh the risks of 3. It may make sense to try out a 6 month buyback program, and see how it works. To minimize the risk of 3, it is important to correctly choose the signal of Fluid’s valuation (see below), in order to maximize the chances of positive effect on the FDV after buybacks.
Implementation
Agree with @DMH that for a comparative analysis with other protocols it is probably better to use FDV/annualized-fees metric, since fees is the most comparable metric. So one could try to determine undervalued levels of FDV/annualized-fees, and based on that allocate % of revenue to buybacks.
However, as @bmpalatiello pointed out, the market of DeFi protocols is (1) heterogeneous, and also (2) nascent. What this means is that choosing the right levels will be (1) complicated and (2) mandatory to review. In view of this, it may make sense to adopt a “self-correcting” strategy that @bmpalatiello suggested, i.e. take the long-term moving average as the fair valuation, and short-term moving average as the current valuation. And based on these two signals allocate a % of revenue to buybacks.
The choice of signal still remains, however. One may choose FDV/annualized-fees, FDV/annualized-revenue, or choose to account for incentives & use FDV/annualized-earnings. Note that even though earnings are close to 0 in case of Fluid, the difference between long-term moving average and short-term moving average can still indicate that token is undervalued.