Assessing the Impact of Trading Rewards Reduction on dYdX's Active User Base: A Quantitative Analysis

Governance/DAO Discussions
1

Following the 45% reduction in trading rewards during Epoch 21 and Epoch 22, sufficient data has now been accumulated to assess the impact on the user base of the decentralized exchange platform, dYdX.

Methodological Approach:
In order to develop a methodology that is not influenced by market conditions, I propose an approach that, while not ideal, serves as a starting point for analysis. I identified two groups of “active users”: accounts that received more than 500 dYdX rewards in both Epoch 19 and 20, and accounts that received more than 100 dYdX rewards in the same period.

The analysis revealed that there were 291 accounts with over 500 rewards and 794 accounts with over 100 rewards. I then determined the number of these accounts that received zero rewards in either Epoch 21 or 22. The results showed that 117 accounts from the first group and 347 accounts from the second group experienced this decline.

Acknowledging the possibility that some users may have created new accounts or that new users may have joined dYdX, I established two additional groups to account for this factor. The first group comprised accounts with zero rewards in Epoch 19 and 20 but with over 275 dYdX rewards (500 minus 45%) in both Epoch 21 and 22, while the second group included accounts with more than 55 tokens (100 minus 45%). These groups consisted of 50 and 146 users, respectively.

Resulting Insights:
Considering these numbers, I estimate that dYdX lost 67 users (117 minus 50) in the first group and 201 users (347 minus 146) in the second group. Consequently, the loss percentage for the first group is approximately 23% (67/291), while the loss percentage for the second group is around 25% (201/794).

In conclusion, the reduction in trading rewards led to the loss of at least 23% of dYdX’s active user base.

I kindly request @Callen_Wintermute to provide feedback on my research findings

7 Likes
2

Great work Vovochka I have a few questions.

Group A = 500 dYdX rewards (or more)
Group B = 100 dYdX rewards (or more)

Is there a possibility the data within Group B can be affected by market conditions much more than Group A? Due to the potential increase or reduction in overall trading affecting whether they cross the 100k volume threshold to receive rewards?

As a follow-up to this would it be beneficial to backtest the data on Epochs further back to get a control group of usual turnover/onboarding?

1 Like
3

Thanks, @NTinc , for your follow-up questions!
Personally, I don’t believe the market conditions were unfavorable during the last two epochs. In fact, the overall crypto market cap increased from $930 billion to $1.28 trillion and is now at $1.14 trillion. So, I’d say the market conditions in the last two epochs were relatively positive

I believe that, generally speaking, users who received 100+ rewards in both epochs probably had a trading volume much bigger than $100,000. It’s only 791 users while dYdX stated that in Epoch 19 there were over 13000 active traders.

Backtesting data is awesome. Since the rewards distribution for Epoch 22 was only released today my research is just an initial step in understanding the real impact

1 Like
4

Thank you. I know you have been hard at work on your own time nonetheless and I appreciate what your doing. No rush on the backtest I look forward to others’ opinions and talking about this more.

1 Like
5

This is really interesting @RealVovochka and a data driven post mortem is something we should probably do for all proposals like this.

Some comments from my side:

I think the success metrics of churn of high and medium volume traders is an interesting one and I really like it. Personally I like the idea of optimising for the dual objectives of growing volumes (measured appropriately) and growing the user base, which this measures perfectly as volume growth over the long term is what will increase liquidity, fees, brand, prominence, etc. This means adding an unbiased volume metric.

If you like I can help add a modified version of the volume measure @0xcchan and I used in the fee research. The idea is to remove the market condition impact by including a variable for Binance volumes in the same markets.

Out of interest how did the total number of traders receiving rewards change? I’m curious if the rewards from the churned group went to new traders or was concentrated more amongst existing traders not churning?

I think @NTinc was alluding to this, but if it is possible to track this churn rate over previous epochs to get a baseline to compare to it would be really helpful?

My last comment is that traders clearly care about getting money back (a la poker rakeback) and not about owning DYDX tokens given that almost all end up on Binance within a day or two. The way I see it for now it is the dollar value of rewards that is important and not necessarily the number of tokens. Obviously the emissions is what the DAO controls, but epoch 19 in particular was a bit of a local high in the token price so we must be careful not to ignore that. My belief is that is why the initial drop in trader rewards had no negative long term effect.

Dollar value of emissions:

image
image911×44 4.61 KB

1 Like
6

That’s interesting about the user base. We would be interested to see total number of users epoch to epoch. This is important data to see, and the average volume traded per user.

7

Hi @RealVovochka, thanks for the post!

I think there are multiple ways to assess the impact of the reduction in trading rewards on both users and trading volume. Your analysis seems to directly look at user base.

A reduction in user base was always expected from this large change and so the figure your analysis ended up on isn’t too much of a surprise. With such high incentives, there was going to be a level of inorganic volume/users that trade to make a profit between fees and dydx rewards.

Volume metrics for Epoch 21 and 22 look on par at first glance: Metabase

I’m also not sure if ignoring market conditions is a realistic way to measure changes, for example, it’s important to look at:

  • relative dYdX trading volume across the DeFi dex sector, CEX sector, Perp DEX sector etc.
  • same for users where possible

We have also seen a large focus on memecoins/coins that are not offered by dYdX which could have an impact on dYdX users. But @0xCLR has some good ideas as to where to look next!

Happy to help out on further research if needed

8

Thanks a lot, @Callen_Wintermute!
My original post was intended to examine the reduction in rewards from the viewpoint of a user base. However, I’ll certainly incorporate the insights from @0xCLR and your suggestions. This will enable me to conduct a more thorough investigation into this issue from various perspectives.

I don’t believe that inorganic volume or users pose a significant issue anymore, given the current price of the token. However, the top recipient of the rewards does seem quite questionable, an issue I’ve addressed in another discussion.

1 Like
9

Interesting data, appreciate the insight here.

As everyone knows, liquidity begets liquidity – creating a flywheel that leads to great growth when liquidity / users are abundant, but can lead to precarious positioning when liquidity becomes thinner. Analysis like this is extremely important.

For some context I’m Cole, one of the cofounders of Elixir: the industry’s decentralized algorithmic market making protocol (@Cole0x on Twitter, also a dYdX token holder / genesis user since the first few months of launching!)

A bigger area of analysis we should also explore is expanding the capital efficiency of existing token emissions. The amount of liquidity on the books is actually a two-sided question: how many emissions are there, and what is the actual capital efficiency of those emissions? If we assume that roughly 1/4th of the total open interest is actually on the orderbooks at a given time, these rewards equate to roughly a 60%+ APY paid out in dYdX tokens.

As one example, at Elixir we could unlock retail liquidity for algorithmic market making on dYdX (similar to what we’re doing with Injective). This way anyone could easily supply assets to pairs via Elixir, and earn a share of the liquidity provider rewards dYdX is already distributing. Market making rewards would remain constant, but tens of thousands of users could supply liquidity (instead of only a handful of centralized firms), freeing the market for these rewards.

Potentially will create a more formal proposal down the line re: a potential solution to bring in more liquidity for each dYdX token the exchange spends on liquidity.

Good to meet everyone and excited to get more active here.

~Cole

3 Likes
10

Hey @cole_elixir, great to see Elixir protocol here! Been a fan of the objective with decentralised market making since the testnet was launched earlier this year!

It’d be great to see how this can play out in encouraging users to provide liquidity. Previously, there was a collaboration with Hummingbot to update the connectors and it attracted decent volume over the past few months. Would be a great addition with Elixir protocol, esp with your experience in integrating with Injective.

If we assume that roughly 1/4th of the total open interest is actually on the orderbooks at a given time, these rewards equate to roughly a 60%+ APY paid out in dYdX tokens.

As for this statement, I’m just curious how did you derive the 60+%?

11

Appreciate the kind words re: Elixir @0xcchan!

Here’s how we calculated the capital efficiency of existing DYDX token incentives:

$275m in open interest * (1/4) = $68.75m on orderbooks (assuming 1/4 of platform TVL is actively deployed on the orderbook)
Monthly emissions (from earlier in thread) = ~$3.5m in monthly DYDX token rewards
($3.5m*12)/$68.75m = 61.09% APY

Let me know what you think about these assumptions or if you can help hone them in a bit more!

12

@cole_elixir There is no Open interest in trading rewards formula. You can have zero OI and pay fees and get tokens. So this APY is kinda of artificial indicator in my opinion.

13

Yep @RealVovochka - dYdX rewards both liquidity on the books as well as trading volume. If dYdX pays out $3.5M a month for $68M in orderbook liquidity, then it’s effectively paying 61% APY to market participants to subsidize its liquidity (as well as its volume).

This doesn’t mean that the full 61% APY fully accrues to liquidity providers (since some of these go to market takers as well). It’s worth noting that since half of the volume-based rewards are paid to liquidity providers (on top of their LP incentives), liquidity providers should have a much larger proportion than takers.

While imperfect it’s important to get a sense of the capital efficiency of the rewards distributed. In our opinion the exchange should aim to increase the efficacy of these existing rewards, e.g. doubling the amount of liquidity on the books for the same amount of dYdX tokens distributed, targeting roughly 30-35% APY

14

~ personal thoughts as community member ~

Active number of traders is an important metric to look at, but likely more important to look at are total trading volume and fees paid. Do we have any good data on how the decrease affected those?

It’s also very interesting to look at the fees paid vs $ token incentives for trading rewards. After the decrease the fees paid has been almost double the incentives (just looking at trading rewards). From that perspective I would personally consider the reduction in trading rewards to be a huge success

3 Likes
15

Hey @cole_elixir, @RealVovochka,

  • Yeap, the OI component was removed from the trading rewards formula, and its now solely based on fees.
  • For the historical OI and volume data, you can retrieve it from this endpoint!
16

Great work to all for the discussion - I think we’re able to get a clearer picture of the effects based on the post above (Great work @RealVovochka ) and the discussion below.

My personal thoughts:

I’ve been following this topic for a while now and have read the multiple posts in relation thereto. Initially, I must admit that my conclusion was that the reward reduction was a net-negative due to the fact that I based my conclusion merely on the user-base (which is important).

However, with that in mind, I must admit that my view on the above has changed. The decrease in user-base was probably caused by the reward reduction (on this, I agree). However, let’s also keep in mind that our main user-attraction should not be reward-disbursement or some other financial incentive, but rather, it should be the development and operation of a best-in-class decentralised derivatives platform. Hence, my reasoning is that a user-base propped up or attracted merely by reward incentives is not sustainable for the longer-term as it would deplete our treasury in a non-sensical manner when those funds allocated for rewards would be better used in other areas.

Furthermore, I also think that the reward reduction proposed by @Callen_Wintermute will be looked at in a couple of years’ time as a responsible and wise move from a treasury management perspective. As highlighted by @antonio , fees paid are now almost double rewards disbursed (we should not only aim to maintain this, but come up with ways and means on improving upon it and increase the fee-margin in comparison to rewards disbursed). The dYdX DAO will need to be self-sufficient and, to achieve that state of self-sufficiency, we need to be very careful in regard to how we utilise and allocate funds from our treasury (which is our life-line). Furthermore, we must come up with ways and means of bolstering our treasury (this proposal was a good first step in achieving that goal).

In conclusion, I do believe that if our focus remains on developing and optimising the Platform, the % user-base we lost as a result of the reward reduction will inadvertently return in a more organic manner - which would be more sustainable for the longer term. Let’s focus on the longer term.

Thank you all for the discussion!

17

Dear @antonio ,

I appreciate your insightful feedback and concur with your assertion that our user attraction strategy shouldn’t merely rely on the dispensation of rewards.

The primary economic indices we employ are trading fees amassed and the USD equivalent of rewards disseminated. As it stands, approximately one-third of trading rewards are being accrued by a single account. This essentially means that this account is responsible for one-third of our entire trading volume. However, in my view, this poses potential risks to the overall stability of a project.

In contrast, I advocate for a more balanced distribution. I believe that having 100 users, each contributing a volume of 10,000, is significantly healthier and more sustainable than having one user monopolize 1 million of the volume. My assumption is that smaller users might display elasticity towards trading rewards, and fostering a large, loyal user base should be a priority before advancing to V4. Smaller users can serve as highly effective brand ambassadors.

Trading rewards do function akin to a rakeback system, as observed in poker. Users benefit from fee discounts and obtain a ‘loyalty bonus’ at the end of each epoch, which can be considered a form of gamification. Personally, I perceive this as a superior strategy compared to offering direct fee discounts.

A significant concern to address is the prevalent practice of users instantly dumping their tokens. Therefore, the stkDydx factor, which encourages users to retain their dydx tokens, was a commendable solution. It might be worthwhile to consider initiatives such as trading leagues for wallets holding specified amounts of tokens, such as 100, 1k, 5k, or 10k, to incentivize users to maintain their dydx tokens. Moreover, understanding our user base, their motivations, and their engagement with the dydx product is paramount.

Looking forward to V4, I acknowledge that its innovative design might not appeal to all, especially some of ETH traditionalists who may resist using the Cosmos chain. However, I believe that the degree of their loyalty will significantly influence the successful conversion of our user base.

While I wholeheartedly agree that our primary focus should be on the development of a superior product, it’s also essential to bear in mind that neither we, nor the DAO, can predict with certainty how users will respond to the product. As such, I propose extensive research into the user base’s reaction to previous changes in the dydx product. This approach will provide valuable insights, guiding our future marketing strategies. I would be grateful if you could share any insider information you have in dydx trading about user base that might help in this regard.

Best Regards. Vladimir

18

Hi @antonio,

Thank you for sharing your valuable insights on the recent changes in trading rewards on dYdX. I have been reflecting upon your perspective, particularly regarding the doubling of fees relative to rewards, and as a trader and community member, I believe there are some concerns worth addressing.

We’ve observed a considerable decline in trading volume and active traders following the 45% reduction in trading rewards. This development doesn’t merely affect the ratio of fees to rewards but also has broader implications for the platform’s liquidity.

As you’re well aware, decreased liquidity often leads to larger bid-ask spreads and price gaps, particularly on shorter time frames. This situation contributes to a challenging trading environment, making it more difficult to execute trades at the desired prices and leading to increased slippage. This enhanced slippage can result in more substantial losses for traders during volatile periods, a risk now exacerbated on dYdX due to the reduced rewards.

Furthermore, this decrease in liquidity and the subsequent increase in trading risk could discourage traders. This could potentially instigate a negative feedback loop leading to a further decrease in trading volume. When compared with centralised exchanges, which offer competitive taker fees as low as 0.01% and are less prone to significant slippage issues, it seems crucial to ensure that trading on dYdX remains an attractive option.

An additional concern relates to the current reward distribution system. In my view, the concentration of rewards into a single account could discourage broad participation, further threatening our user base and the platform’s liquidity.

While these observations reflect my personal opinion rather than definitive facts, I feel the long-term implications of decreased incentives, declining volume, and reduced liquidity are potentially serious issues we need to address.

I cordially invite you to join us in a discussion we’ve initiated about innovating and revitalising the dYdX rewards system. We are brainstorming strategies to enhance user adoption, foster platform growth, and increase trading volume.

The thread can be found at this link: https://dydx.forum/t/revitalising-the-dydx-rewards-system-innovative-strategies-for-enhanced-user-adoption-platform-growth-and-increased-trading-volume/56

Our discussion explores several key strategies and principles:

  1. Crafting principles for an equitable rewards system.
  2. Developing strategies to prevent reward concentration, such as caps and limits on rewards per account, and a tiered rewards distribution system.
  3. Proposing incentives for user growth and participation, including bonus rewards for new users and affiliates.
  4. Suggesting additional rewards incentives like incentives for long-term users, competition rewards, activity-based milestone rewards, loyalty tier rewards, and early adopter rewards.
  5. Enhancing the user experience and accessibility of the rewards system through a user-friendly interface.

These measures aim to create a more equitable system that fosters a strong user base, maintains robust liquidity, and ensures competitive trading conditions.

As a fellow member of this community, I believe we should concentrate on nurturing a strong user base, preserving robust liquidity, and ensuring competitive trading conditions to safeguard dYdX’s standing in the DEX market. Our ultimate goal is to put forth a proposal that the community can vote on, ensuring we collectively build a sustainable and thriving platform with a better rewards system.

I look forward to your engagement in this crucial discussion.

Kagan

19

Hi Kagan,

May I ask where you are looking at data that suggests there is a significant decline in trading volume?

Data from here and the dYdX Epoch Review blog posts don’t indicate a significant reduction in trading volume:

image
image838×233 19.4 KB

image

Once again, these stats should be looked at in aggregate across crypto/defi’s trading volume. But at first glance, fees paid, trading volume, and traders eligible for rewards look pretty in line with previous epochs.

20

Hi @Callen_Wintermute, we have been examining data from the dYdX Epoch Review blog posts. To provide a comprehensive view, I’ve compared key metrics from four recent epochs - two before the rewards reduction (Epochs 19 and 20) and two following it (Epochs 21 and 22).

Before the rewards reduction, during epochs 19 and 20, the platform demonstrated robust activity:

  • Trading Volume: $30.7B and $38.0B, respectively
  • Average Daily Trading Volume: $1.1B and $1.4B, respectively
  • Net deposits and withdrawals: $195.1M and $178.9M, respectively
  • Unique depositors: 4357 and 4114, respectively
  • Fees Paid: $7.5M and $8.6M, respectively

Following the rewards reduction, during epochs 21 and 22, we see a decline in these metrics:

  • Trading Volume: Dropped to $31.5B and $30.2B, respectively
  • Average Daily Trading Volume: Decreased to $1.2B and $1.1B, respectively
  • Net deposits and withdrawals: Turned negative at -$47.5M and -$97.4, respectively
  • Unique depositors: Dropped to 4334 and 2906, respectively
  • Fees Paid: Fell to $7.7M and $7.1M, respectively

Following the rewards reduction, we observe a distinct downward trend across all key metrics. For instance, the trading volume decreased by approximately 20.5% from Epoch 20 to Epoch 22. There’s also a drop in unique depositors by around 29.3%.

This decline is reflected in the average daily trading volume (a fall of about 21.4%), and in net deposits and withdrawals, which shifted significantly from positive to negative. The fees paid to the platform also decreased by approximately 17.4%.

While we understand the importance of looking at these stats in the context of the broader crypto/DeFi’s trading volume, these changes indicate that reducing rewards has negatively influenced user behaviour on dYdX.

Judging by the first two weeks of the current epoch, these metrics are looking increasingly concerning. We’re observing an average daily trading volume of around $500 million, possibly even less. This represents an alarming further decrease of approximately 50%.

These changes lead to less liquidity, making the platform less appealing to traders. The increase in net withdrawals and decrease in unique depositors suggest a decline in user confidence, which impacts the platform’s growth and reputation. Lastly, the decrease in fees paid negatively affects revenue and profitability.

In light of these observations, we propose reconsidering the reward reduction. dYdX is in the early stages of product development and is vying for a significant share in a highly competitive market. Given this scenario, reinstating and redesigning the rewards system entirely will help mitigate these negative effects and improve user engagement.

I have initiated a discussion on the forum around redesigning the rewards system entirely. You can find the discussion here.

Our priority is to drive platform growth and increase trading volume. I believe revisiting the rewards strategy could be a significant step towards achieving this goal. I hope this analysis provides a valuable perspective on the post-reward reduction landscape, and I look forward to discussing this further.