New Delegate Compensation scheme

General
1

We’re ready to announce the v2 of the delegate compensation scheme.

Due to budget constraints the total amount distributed each month has to be halved unfortunately.

Last month we had a long and deep discussion about metrics, cutoffs and what makes delegates valuable.

What stuck with me, was that the previous scheme didn’t do a great job of creating a true community of governors. @ChronoTrigger rightfully pointed out that a hard cutoff at place 7 meant that no one could be sure they’d receive compensation for their valuable work. That is not ideal.

Hence we have developed a new scheme, that distributes the available funds along each qualifying participant, with a soft tapering to reward effort but not enough (we hope) to justify gaming.

We also have some long-term plans to tie delegate compensation to overall ecosystem success. This needs deeper simulation and thinking through, and will be announced separately. The idea there being that delegate compensation would rise with quantifiable ecosystem success.

New Delegate Compensation scheme

As of now each delegate that has:

  • Participated in 90% or more of all polls
  • Written rationales for at least 90% of their votes cast.
  • Has more than 50,000 stRIF delegated.
  • Has more than 3 hours of read time.

will receive delegate compensation.

Delegates will be ranked at the end of each month with the same formula as last month (April), consisting of read time and likes, with likes being weighted at 1/6th of the month before.

Time spent writing posts also contributes to read time, so writing answers and proposals als counts towards the ranking.

The total amount of $3,750 USDRIF will be distributed among all qualifying delegates in such a way that the bottom ranked delegate receives 30% less than the top ranked delegate.

Here is a sample distribution with 6 qualifying delegates:

1. $735
2. $691
3. $647
4. $603
5. $559
6. $515

And a sample distribution with 8 qualifying delegates:

1. $551
2. $528
3. $504
4. $481
5. $457
6. $433
7. $410
8. $386

How can delegates be added / removed?

Since adding delegates reduces compensation for everyone, we have introduced the 50k stRIF delegation threshold for new delegates. This makes it harder for new delegates to receive compensation, without either putting up significant skin in the game, or doing stellar work to attract delegation.

We want to point out the focus on SEA and APAC whales and the possibility to attract delegation there in this context.

Delegates will be removed from the compensation tracker for:

  • Two months of inactivity measured as less than 40% votes or less than 40% rationales
  • Cheating, gaming and otherwise subverting the spirit of the community
  • A peer vote by other delegates to remove one delegate. This proposal should have strong arguments and evidence, not just vibes.

Polls to finalize this:

What should the minimum threshold of delegation be for delegates to qualify?

  • 30,000 stRIF
  • 50,000 stRIF
  • 90,000 stRIF
  • higher
0 voters

Should likes be counted for delegate ranking?

  • No
  • Yes, with lower weights
  • Yes, with current weights
  • Yes, with higher weights
0 voters

What’s a good minimum read time threshold to receive delegate compensation

  • 60 minutes
  • 180 minutes
  • 240 minutes
  • higher
0 voters
6 Likes
What do Rootstock builders/grantees need right now from The Collective?
2

This looks great. Glad to see a soft tapering approach land. The continuous distribution addresses exactly the failure mode that I was concerned before.

I spent ~15 years working in managed sales channels and part of those years modeling sales compensation programs in private companies, and found discontinuities or binary cutoffs tend to create issues. Excessive internal competition, toxic dynamics, and all-or-nothing scenarios that look elegant in spreadsheets but punish you in practice. It can be romantic to assume everyone will consistently deliver the “all” performance, but in reality you’ll often be getting the “nothing”.

One quick example: A delegate has a rough early month, or attends a stretch of in-person events that breaks their cadence. Under a binary scheme they quickly conclude they cannot catch up to make compensation that month, and the incentive to keep contributing collapses. Continuous payout structures avoid this because every additional unit of effort still carries real marginal reward, which sustains participation through the bad weeks. That is just one specific example, the pattern repeats in other examples.

I much prefer this model to the previous iterations and fully support moving forward with it.

A quick note on my poll votes:

On likes weighting, I think likes have already proven to be a bad metric for reasons not worth flooding this thread re-listing or restarting the discussion. Giving a bad signal a lower weight is still adding bad signal into the system, just hoping it’s not high enough to cause damage.

On read time, I pulled last months’ numbers. Last month the bottom ranked delegate with any meaningful read time (above 1h) sat at 6.8h. The previous month had one delegate at 3h, but that was unusually low including for that delegate, and all others had 5h+. I feel 4-5h feels like the right floor. I’d suggest 5h. Above 5h is dangerous as we seem to have agreed we had a good delegate last month, with active participation and solid contributions, with 6.8h read time.

On the delegation threshold, this one is trickier. We need to balance keeping the door open to new delegates, against the risk of a swarm of people farming the formula with hands-off AI for free ~$300 per month. I’ll sleep on it before voting. But I’m leaning on 50-90k.

2 Likes
3

In principle, we support this new approach, although with some reservations that we would like to outline below. Unlike @ChronoTrigger, we believe that as we are not dealing with employees of a company, but rather independient professional delegates serving a decentralized organization, we see competition as healthy, since it aligns incentives for delegates to contribute more in pursuit of improving their compensation. There is nothing wrong with competing, nor with wanting to perceive higher incentives, and competition can create the right dynamics to ultimately benefit Rootstock by motivating delegates to contribute more time, effort, and value.

That said, we do agree with adopting a more gradual compensation distribution model, since rigid tiers with abrupt drops can distort incentives. A delegate having a few bad days, getting sick, attending an event, or taking time off should not result in losing all or most of their compensation. A gradual structure aligns incentives much better, as lower performance in a few days in a given month due to temporary circumstances would reduce rewards, but not in a drastic way.

Regarding whether the program should incentivize 6 or 8 delegates, and considering the proposed compensation amounts, our view is that under a larger budget, supporting 8 incentivized delegates could make sense. However, given the current budget constraints, we believe the 6-delegate model is more appropriate. The compensation levels under the 8-delegate scheme are simply too low, and realistically, nobody dedicates significant time and effort in exchange for very limited compensation. People naturally calibrate their level of commitment based on the economic incentives involved, so lower compensation will generally lead to lower engagement and effort. In that sense, while the proposed 6-delegate scheme is already low and modest, the 8-delegate version risks creating insufficient incentives for the level of dedication Rootstock actually needs.

Additionally, unless my math is wrong, there are currently around 8 active delegates in Rootstock. Consistent with what we mentioned earlier, we do not see an issue with fostering competition among those 8 delegates for the 6 compensated spots. If everyone receives guaranteed compensation under a very gradual and low-paying structure, incentives shift toward contributing the bare minimum, since additional effort would not materially change the outcome. For that reason, we believe the 6-delegate structure is the most suitable under the current circumstances, while remaining open to revisiting it in the future if a system is implemented that allows the compensation budget to grow tied to the ecosystem’s success.

Regarding the delegation threshold, the system currently requires more than 10k stRIF of voting power, so increasing it 5x to 50k seems reasonable as a first step, followed by an observation period and a potential future increase if necessary.

As for counting likes as a metric, we vote NO. It is not a meaningful indicator of contribution, effort, or dedication. Moreover, it creates misaligned incentives, as delegates become reluctant to like other delegates’ posts, even if they find a comment valuable, because doing so may indirectly negatively affect their own ranking or compensation. We support removing this metric entirely, which would allow delegates to freely recognize good contributions without worrying about self-harm to their incentives.

Finally, regarding reading time, it is difficult to establish an exact minimum. However, we agree with @ChronoTrigger that setting a threshold of around 6 hours per month seems reasonable. Even that is still a very low bar, but it could at least function as a minimal filter. For that reason, we support a higher threshold than the ones currently proposed.

We look forward to this, as we believe it is essential to creating the right incentives for delegates for making more impactful contributions and dedicating more time and effort, knowing that this, along with the success of the ecosystem, will lead to increased compensation.

1 Like
4

To clarify this further. The proposed scheme has no fixed seat count. Every qualifying delegate is compensated and the admission bar is raised.

1 Like
5

Understood. In that case, we have further reservations about the proposed program that remove the current 6 slots for incentivized delegates, in which all delegates compete to qualify based on their contributions, creating incentives to increase contributions.

Therefore, if this scheme without fixed slots is adopted, we reiterate our earlier point that such highly gradual with low compensation structure may ultimately risk reducing the level of contribution, time, and effort delegates are willing to dedicate, as people naturally calibrate their effort according to the compensation they receive, and even more, delegates may become less incentivized to go beyond the minimum expected contribution, as additional effort would have only a marginal impact on the final outcome. We therefore want to raise a note of caution on this point and we also suggest closely monitoring the evolution of the program and the level of delegate contributions under this new framework, to see if any adjustments are needed.

And, as we said above, we look forward to the implementation of the mechanism through which delegate compensation can grow alongside the success of the ecosystem, as this could help restore and strengthen the incentives for greater contribution, dedication, and long-term commitment :rocket:

2 Likes
6

Understood and definitely something we’ll watch out for. Given the current budgetary constraints we unfortunately don’t have more to distribute. But this could change with protocol success raising the total amount and with additional opportunities that might open up that are directly tied to value created.

7

Thanks @Raphael_Anode. I just want to add that it’s important for the program to be crystal clear on what problem it is trying to solve. Then work backwards to identify the appropriate metrics.

For example, the problem is not enough of “x” type of project, then business development activities should be incentivized.

With ecosystem revenue being the stated North Star in the updated grant guidelines, then we should work backwords to identify what activities delegates need to work on to help support increasing revenue.

1 Like
8

It’s important to distinguish the two.

Atm delegates function is very clear: check grantees and vote on grants proposals in the spirit of the new grant guidelines.

Ecosystem growth is another function that is currently not measured in delegate compensation.

We can and should figure out what a good way forward here is for engaged delegates that is fair, measurable and mutually profitable.

1 Like
9

While expanding the program to include more delegates sounds great for decentralization on paper, diluting the compensation across an uncapped number of seats is definitely a double-edged sword. We might gain a larger quantity of delegates, but we risk losing the high-impact, proactive contributions that a tighter, more competitive structure naturally encourages. Never the less this seems like an interesting approach, 4-6m down the lane the results of the initial delegate compensation and this new one can be compared side by side to see what actually translated to real ecosystem value.

1 Like
10

Thanks, @Raphael_Anode. I’d like to share my rationale for the survey responses and add one broader point on how we should think about delegate effectiveness going forward.

1. Minimum delegation threshold: 90,000 stRIF

I support a relatively high barrier to entry for compensated delegates. Since compensation would be available to all qualifying delegates rather than a fixed number of slots, the qualification threshold should be meaningful enough to discourage low-effort participation or formula farming.

90,000 stRIF feels reasonable because it is roughly 60% of the delegation active delegates currently have in the Collective. That still leaves room for new delegates to qualify, but requires either meaningful self-stake, credible delegated support, or a demonstrated ability to attract trust from stRIF holders.

2. Likes in delegate ranking: yes, with current weights

I agree that likes are easy to game and can become a vanity metric. They do not reliably measure delegate quality on their own.

That said, likes are also a very common behavior pattern in forum-based governance. People already use them to signal agreement, appreciation, or perceived usefulness, so I do think they can carry some value if treated as one input rather than a primary measure.

The bigger issue is transparency. It would be helpful for delegates to have visibility into where likes are coming from, especially if likes continue to affect compensation. Right now, only admins can see that information. A basic analytics view showing the source and direction of likes would make it easier to identify coordinated or low-quality engagement.

3. Minimum read time threshold: 180 minutes

I support 180 minutes and feel strongly that we should not incentivize higher read time for its own sake.

With AI reducing the workload for delegates, we should be leaning into better leverage, not rewarding people for artificially spending more time reading forum posts. I do not think there is evidence that higher read time necessarily produces higher-quality governance outcomes.

AI is making it easier and faster to do more with less. The goal should be better evaluation, not more hours logged.

More broadly, if the role of delegates right now is to “check grantees and vote on grant proposals in the spirit of the grant guidelines,” then we should be investing in AI governance agents and delegate tooling that help delegates do that work more effectively.

An AI governance agent can help delegates evaluate grantees, triage proposals, detect monthly reporting variances, maintain institutional memory, check for conflicts, and keep vote rationales consistent. These are areas where humans often struggle because the work is repetitive, data-heavy, and context-dependent. AI can be especially useful for data analysis, identifying inconsistencies, and surfacing irregularities in grantee and ambassador outcomes.

Finally, I think we should have delegate tooling for tracked metrics that clearly shows the source and direction of all likes and read time. Basic data transparency would help identify gaming, improve trust in the compensation system, and give delegates a stronger basis for evaluating whether the scheme is producing the behaviors the Collective actually wants.

2 Likes
11

First of all, we truly appreciate @Raphael_Anode for trying to find the best possible way to improve the program, even though the system alone can’t solve everything, because at the end of the day it also depends on how people respond to it. It’s like solving a wicked problem if people don’t act in good faith.

The purpose of the Recognized Delegate Compensation is to incentivize contribution, which naturally leads to “how do we measure contribution?” Most metrics have flaws and can eventually be gamed, there’s no perfect answer. But the real question shouldn’t start from “what metric do we use”, it should start from “what is the role and responsibility of delegates?” Which is simply: review grant proposals, cast votes, and provide constructive feedback. From there, we work backwards to find the most practical signal we have, even if it’s imperfect.

We mostly agree with the new system, but wanted to share some analysis to ensure we as a collective are making decisions based on actual evidence, not just what feels right, or what reflects our own individual work style. Because everyone performs differently, right? Some love spending a lot of time on the forum, drafting long replies. Others don’t spend as much time but stay engaged through shorter, focused responses. Either way, we need to find the middle ground that reflects what we actually have in common.

Here is our analysis:

Read Time Per Delegate Per Month (hours) — Since the start of the program

Delegate Nov Dec Jan Feb Mar Apr Avg
ChronoTrigger 9.00 5.21 5.17 11.67 17.69 14.58 10.55
SEEDGov N/A N/A N/A N/A 8.00 10.64 9.32
DAOStar 0.08 0.01 8.74 15.02 15.35 15.46 9.11
Curia 11.00 3.56 5.25 4.77 5.27 9.46 6.55
Axia Network 7.00 3.56 5.51 5.90 5.41 6.77 5.69
Tané 9.00 4.13 5.01 5.95 5.41 0.00 4.92
DAOPlomats 1.00 1.96 2.53 8.19 2.82 8.71 4.20
Ignas DeFi 2.00 1.70 1.83 2.78 3.07 9.44 3.47

* Axia Network includes data from 404Gov. November figures are based on a 60-day rolling window, not strictly monthly.

Monthly group averages (Overall program average: 6.49h/month)

Month Avg Min Max
November 5.58h 0.08h 11.00h
December 2.88h 0.01h 5.21h
January 4.86h 1.83h 8.74h
February 7.75h 2.78h 15.02h
March 7.88h 2.82h 17.69h
April 9.38h 0.00h 15.46h

Note: The December and January dip (2.88h and 4.86h avg) may reflect the holiday period.

To answer: What’s a good minimum read time threshold to receive delegate compensation?

Read time may not be a perfect measure of contribution quality, but it’s the most consistent observable “signal” we have so far.

We respect this perspective, and we agree that AI tooling can genuinely help delegates work more efficiently. But read time is a signal that a delegate is actually following the conversation. AI can help you process information faster, but it can’t replace the judgment that comes from actually being present in the discussion.

And the data backs this, across 6 months, the group is already naturally averaging 6.49h/month. That said, delegates like @ChronoTrigger and @DAOstar_gov consistently read well above the rest, which pulls the average up. Strip those outliers out and the realistic center sits closer to 4–5h.

With that in mind, we’d suggest setting the baseline at 4h (240 min), grounded in actual behavior, not arbitrary. Low enough not to penalize delegates during natural dips like the holiday period, but high enough that it can’t be gamed with minimal effort.

To answer: Should likes be counted for delegate ranking?

Previous months have already shown that “Likes Received” can be easily gamed.

And even with full transparency, the problem runs deeper than that. Everyone can already see where likes come from by checking manually, but there’s nothing stopping someone from creating another account to work around it. It’s a behavioral problem, not a transparency problem, no analytics dashboard fixes that. That’s why we don’t see the benefit of keeping likes in the formula.

To answer: What should the minimum threshold of delegation be for delegates to qualify?

As Rootstock works toward onboarding more stakers, particularly across regions like SEA and APAC, active delegates could potentially grow their voting power over time. That makes the delegation threshold even more important as the entry gate. It needs to be meaningful enough to ensure that new delegates who enter are genuinely committed and of quality, whether through self-stake or by demonstrating enough credibility to attract delegation from RIF holders.

That’s why we’d support 90,000 stRIF as the minimum threshold. It’s roughly in line with what the least-delegated active delegate currently holds, as Axia already pointed out.

This also leads us to another point we think is equally important, as all three questions above. With @SEEDGov and @Eren_DAOplomats already sharing their concern on uncapped seats, as Rootstock aims to onboard more stakers, more delegates could potentially qualify each month or even attract more delegates to join. And with a fixed pool of $3,750, that means every additional qualifying delegate dilutes the payout further. At that point compensation stops being a meaningful incentive and starts feeling like a token gesture, which defeats the purpose of the program entirely and could discourage delegates from putting in the actual work, ultimately diluting the quality of governance itself.

We think this needs to be carefully considered. A fixed seat count is still needed, and a soft cap of around 6–8 delegates would keep the compensation meaningful enough to actually incentivize the quality of work the program was designed for in the first place.

5 Likes
12

Thanks for this @Curia. Really appreciate your data-driven approach here!

1 Like
13

Thank you so much @Raphael_Anode and @tamlerner for working hard to offer a revised program while navigating these budget constraints , and also to @Curia for providing the historical data to help ground this discussion, great data, thk u!

As coffee-crusher from the DAOstar team, I wanted to share a few of my own thoughts on the updated framework:

Now that the baseline threshold consensus is clear, we need to focus on protecting the structural integrity of the program. If we want this program to successfully incentivize high-utility governance, we must address two critical areas: seat dilution and the behavioral role of “likes.”

1. The Dilution Risk: The Case for a Competitive 6-Seat Cap

My primary concern with the current v2 design is the uncapped nature of qualifying seats drawing from a fixed $3,750 pool. As @Curia , @SEEDGov , and @Eren_DAOplomats have already pointed out, under the current model, success actively punishes top performers. If the pool dilutes too far, compensation degrades into a token gesture, which completely defeats the purpose of the program.

I strongly advocate for a strict 6-seat cap.

Governance compensation should be competitive. Limiting the pool to 6 seats ensures that the payout remains meaningful enough to justify the deep, analytical work required for complex proposals. A cap doesn’t lock people out; it sets a clear, competitive bar. It ensures that the rewards flow to the delegates putting in the real, heavy lifting on proposals, while providing a clear path forward for rising governors to step up their contribution quality to claim a spot.

2. Killing “Likes” in the Formula: Restoring Intention to Engagement

I fully support completely removing “likes” from the ranking formula. When you attach a dollar value to a metric, you inevitably incentivize gaming and performance art over substance.

Likes should never be an algorithmic optimization tool. Personally, I hand out likes highly judiciously - only when I strongly agree with an author’s positioning, wanting to signal alignment, or wish to show genuine appreciation for a well-reasoned argument (and there are a lot of great comments written by other delegates!). Engagement must be driven by true intent, not by a desire to manipulate a mathematical distribution model. Stripping likes from the compensation formula completely removes the noise and restores the integrity of forum interactions.

By securing a 6-seat cap and eliminating gamified engagement metrics, we can keep this downsized budget highly effective, sustainable, and entirely focused on rewarding professional-grade evaluation. Great work! Tks!

1 Like
14

Thanks to everyone who participated in the votes.

The thresholds to participation will be:

  • 90k stRIF min delegation
  • Likes will no longer be counted towards the delgate score
  • Tie break between 180 min and >240 min of read time (counting @DAOplomats only once everywhere)

Thanks to Curia for their aggregation of read time data. we will set the baseline to 240 mins for now.

All of these metrics aren’t set in stone but represent the ongoing evolution of the program.

On seat count

Last month we correctly identified that the harsh cutoff at seat 6 penalized delegates who clearly did put in the work, yet didn’t make the cut. Maybe because they’re quick readers, maybe because of AI usage. Many factors can contribute here.

A cutoff at a specific seat will inevitably lead to a gaming arms race, because no one wants to go home empty after showing up for a month.

We will keep the seat number uncapped for now.

Since the inception of the delegate program we had 6 delegates drop out and 2 joining the roster of those that achieve compensation on a regular basis. There seems to be little risk of additional accounts swamping the program, especially given the high bar we’re setting now.
Delegates can also vote to remove another delegate from the roster, if they agree that the delegate in question isn’t bringing value to the table. This can happen via a Forum poll with a minimum of one week of duration, with voters set to visible.

Uncapped seats will also find their natural equilibrium when teams drop out once they feel compensation isn’t worth it. But once more, we see low risk in that regard in the current data.

Important to keep the actual Collective in mind when making these decisions and not optimizing for a theoretical optimium.

Compensation this month

With these metrics now firmly in place we will calculate the compensation eligiblity using the thresholds that have been voted on here and an uncapped seat counts.

Thanks to everyone who contributed constructively to this discussion. Onwards to new heights for the Collective and RIF.

3 Likes
Introducing Recognized Delegate Compensation 🏦
15

We regret that the initial decision to remove the cap of 6 incentivized delegate slots has been maintained, despite the majority opinion expressed by delegates on the matter (out of the 5 delegates who commented on this issue, 4 were against the change and only 1 supported it).

The combination of cutting the delegate program budget in half while simultaneously removing the 6-slot cap (effectively increasing it to 8 incentivized delegates) will significantly dilute compensation levels into what are, in practice, merely symbolic gesture. We believe this will negatively impact the quality of delegate contributions, as a near-flat compensation range of roughly $386 to $551 clearly does not adequately reward the level of time, effort, and responsibility that Rootstock requires.

Moreover, such a flat structure creates very little difference in compensation between delegates who contribute substantially more and those who aim only for the minimum threshold. In our view, this misaligns incentives and may encourage lower overall effort, as delegates may conclude that significantly increasing their contribution results in only marginally higher rewards.

The difference in compensation between a delegate who merely reads the forum at the minimum threshold, votes, and justifies their vote, and a more committed delegate, who, in addition to reading the forum, voting, and justifying their vote, analyzes proposals, provides feedback, and remains engaged even outside the forum, is a mere $165. It is easy to see what kind of involvement the current system encourages and how misaligned the incentives are.

For that reason, we strongly suggest closely monitoring the evolution of the program and remaining open to revisiting this framework, including potentially returning to a 6-delegate incentivized cap, if negative effects become evident.

From our side, once the first results under this new structure are available, we intend to promote a non-binding community survey to better understand community sentiment regarding this change and whether there is support for reconsidering a return to the 6-seat cap delegate structure, which would increase compensation levels for qualifying delegates (while maintaining the total budget) and help realign incentives toward healthy competition focused on delivering greater and higher-quality contributions.

3 Likes
16

Just a week before the majority commented on the unfairness of the cutoff. And rightfully pointed out that was a driver for gaming and other unhealthy behavior. Just something to keep in mind because this discussion happened on the delegeate Telegram chat instead of the Forum.

The only way to keep similar levels of compensation as before would be to limit seats to 4. This doesn’t seem like a good outcome to me.

17

Just to clarify: As for us and the delegates involved in the discussions, we never suggested limiting the number of seats to 4; rather, we proposed keeping it at 6, under the new more balanced and soft established distribution method. This way, the compensation would be adjusted downward in line with the budget cut, but It wouldn’t be diluted under the current flat system, which misaligns incentives and levels down.

18

Thanks for the update @Raphael_Anode. I’m cautiously optimistic about the soft tapering approach. The main thing I’ll be watching is whether the taper is strong enough to continue rewarding meaningful contribution beyond the minimum requirements, (as @SEEDGov alluded to), while still avoiding the unhealthy competition that came with a hard seat cap. That balance is tricky, but this feels like a reasonable iteration to test.

Overall, I support trying this direction and revisiting once we have actual data from May.

1 Like
19

Thanks, we’ll be monitoring the situation. The problem with the fixed seat count is that we don’t have a measuring stick that works. Only solution would be for me to not reveal the formula, so it can’t be gamed.

Goodhart’s law is strong here. Let’s see what the new approach does. Evaluate, Recalibrate, improve.

2 Likes