By KERMAN KOHLI
Compiled by: Frost, BlockBeats
Editor’s note: Crypto researcher KERMAN KOHLI analyzes the success of Starknet’s airdrop from aspects such as the claim and issuance of Starknet airdrop tokens, data, and time.
Following up on my last post about Optimism’s multiple airdrops, I wanted to take a look at Starknet’s airdrop since I pulled data at the same time. I wanted to look at the main differences in the claiming mechanisms between Starknet and Optimism by looking at these two token airdrops. The data is now about a month out of date, but considering the airdrops were done a few months ago, it shouldn’t be too far off the actual numbers.
Application and issuance model
The main difference between the two approaches is that Optimism says "we will personally deliver the airdrop to your wallet", while Starkware says "come to us to claim your airdrop". In the former case, it is easier for the user and saves gas. My personal philosophy is that if you do this on a low-cost chain, then cost should not be an issue and there should be a button to claim the airdrop.
That being said, let’s take a look at the Starknet airdrop. Unfortunately, getting data is difficult because:
Starknet’s post-airdrop analytics do not publicly report details of claiming behavior.
Starknet doesn’t have standard EVM format addresses, which means I have to hack to get the data available on-chain.
Anyway, here’s the official chart on how the airdrop will be distributed:
data collection
To get the data I needed I basically used:
0x06793d9e6ed7182978454c79270e5b14d2655204ba6565ce9b0aa8a3c3121025 is my airdrop address.
0x00ebc61c7ccf056f04886aac8fd9c87eb4a03d7fdc8a162d7015bec3144c3733 as my starting block hash.
0x04718f5a0fc34cc1af16a1cdee98ffb20c31f5cd61d6ab07201858f4287c938d as the contract to get the STRK balance from.
I had to go through a lot of for loops and byte programming to get some interesting snippets of the data I wanted.
Anyway, when extracting the data, I found that only 39.8% of people claimed the airdrop, and the rest of the users were basically used as marketing data - in a sense, this is also a good result! Some people may say that this is bad, but if you can convey your information to the widest number of people without giving everything away, then you have found the sweet spot.
Analysis time
The approach I took was to extract all the addresses that had received the airdrop, and then write a script to query the balances of these addresses at the time (i.e. when the script was run). By dividing the balances into "bins", I can see the amount of balances distributed in different "bins". However, due to the limited data information, it is difficult to understand more about these users. The limited data makes the entire analysis more challenging.
Without further explanation, here are the results! I set a threshold of no more than 100 STRK, because the minimum airdrop amount is 111.1 STRK. The distribution of different amount ranges is listed below:
StarkEx users: 111.1 STRK per person
Open Source Developers: 111.1 STRK each
Starknet users: ranges from 500 to 10,000 STRK, with different multipliers
Starknet community members: range from 10,000 to 180,000 STRK
Starknet developers: 10,000 STRK each
Ethereum staking pool: 360 STRK per validator
Solo Stakers: 1,800 STRK per validator, up to 3,200 STRK for higher risk validators
Ethereum developers: 1,800 STRK each
Protocol Guild Members: 10,000 STRK per person
EIP authors: 2,000 STRK each
Overall, this airdrop did not achieve a very good effect! The retention rate of 13.5% is close to the industry average (and the industry average is actually not high). However, considering that ordinary GitHub users like me received 1,800 STRK, from a deeper level, the effect of this airdrop is much worse than we expected! Only 1.1% of users who received token allocations eventually retained them. Let's look at some other indicators to help us judge whether this airdrop was successful.
A simple proxies indicator is the price trend of the token. Below is the price trend of the STRK token over the past 3 months:
Prices fell by 50%, but the market as a whole underwent structural adjustments during the same period. It’s not a great performance, but at least it’s not a 90% drop.
Let’s look at it from another angle: TVL. At least our friends at DeFi Llama can help with this.
TVL rose to around $320 million and then dropped to around $210 million, which is a pretty good retention rate. However, we don’t know how much Starknet paid to get these numbers. Fortunately, I have the numbers. That number is 67,078,250.942674.
If we assume an average token price of $1.50, we can re-express the equation as Starknet spending $100,617,376 to acquire approximately $300 million of TVL, or in other words, approximately $3 of STRK tokens can purchase $1 of TVL
My next question is what is the number of users so that we can understand the CAC model of this equation. I re-plotted the above chart with percentages of number of users.
Ok, from here on, let’s make a nice case for Starknet and only consider the “under 100 tokens” tier. It cost almost $100 million to acquire 519,282 users. That means it cost about $200 per user. If we recalculate using retained users (those holding > 101 tokens), then the cost per retained user would be $1,341.
This is lower than the thousands and even tens of thousands of dollars in retention CAC we saw in the Arbitrum airdrop and other airdrops. While the Starknet airdrop wasn’t great from a retention perspective, from a CAC perspective it was pretty good relative to other airdrops I’ve seen. My thesis is similar to what we saw with the optimistic airdrops: allocating tokens based on diversified attribute criteria yields strong returns
end
Starknet has been relatively thoughtful in how they distribute large amounts of tokens to different groups. The data clearly shows that they ensure diversity in distribution. This is a common feature I have observed in both successful and failed airdrops.
So why don’t more projects consider the diversity of user attributes when conducting airdrops? The reason is that collecting, analyzing data and drawing conclusions is a very difficult task - especially when there is a large amount of data. However, Starknet uses a relatively simple standard and still ensures the diversity of distribution. In fact, with the right tools, the distribution can be more targeted.