Berachain Testnet Activity -- What's Real?

    Developer Activity
    One of the most important metrics to look at in a testnet is meaningful developer activity. We want to see developers actively deploying contracts, upgrading them, etc.
    Understanding the Smart Contracts
    • However, only looking at top level numbers is misleading. Testnets are always plagued with useless contracts from tutorials or even worse, points farmers trying to disguise themselves as developers. To get a more accurate sense, we must analyze these testnet contracts to learn how to best filter away the noise.

    • Most of the top deployed contracts found on Berachain had no verified code, so some basic function signature matching was done first. Unsurprisingly, all top 10 of the most deployed contracts on Berachain are dummy contracts of various varieties, ranging from counters, variable holders, Hello Worlds, and more. Going further down the list, one starts to find ERC20s and oracle-like contracts, but the vast majority seems to be meaningless. We can continue this analysis, but a broad assumption is that most identical deployed contracts are likely not "meaningful" contracts. Edge cases like upgradeable proxy addresses are also fine to exclude because the implementation contract is likely still captured.

    • Additionally, looking at the frequency of interactions with these contracts, a simplifying assumption can be made -- contracts with less than 10 interactions are probably dummy.

    "Real" Contracts

    Applying both filters, we come to the following statistics.

    Conclusion

    Some key observations:

    • Daily contract count dropped from a few thousand to ~400 daily. The consistency of the amount of new contracts created remains suspicious, but could just be a result of very diligent builders.

    • Deployers dropped sharply from 500-1000 daily to ~200.

    Despite the filtering we've done, we still see a vibrant builder community on Berachain testnet. The ~200 deployers roughly check out against the 158 projects listed on Berachain's website, as well as the numerous unannounced Build-a-Bera projects.

    Diving Deeper with Growth Accounting
    Day-to-day retention, averaged across all users, was 29.2%. This means that of the users active in August 11, 2024, 29.2% were active in August 12, 2024. Most developers are resurrected, meaning after the first month of the testnet, we were mostly seeing developers come back every few days to redeploy contracts.
    db-image-image-T2dt
    User retention is a key explanatory factor for LTV. Below, we show absolute user retention by day. For example, x=3 days shows what proportion of the original cohort was active in day 3. This is preferable to day-over-day retention because each day is measured independently. The 7/14/28-day consumer retention rates were 40.1/33.8/29.5%.
    db-image-image-k_xK
    The heatmap shows some interesting day cohorts that have significantly higher retention than most others. Initially, I thought this steady retention was a sure sign of farmers deploying daily. Yet, the high retention cohorts came in late June/beginning of July and every weekend after, which doesn't coincide with any large ecosystem news.
    db-image-image-jB01
    Disclaimer: Flipside AI is here to help but it can make mistakes. Always review outputs and use the upvote/downvote buttons to help us improve. This content is not financial advice.