Decentralised applications are hailed as the pathway to massive adoption. It seems the hype around much of them has been slowly dying out. The present status of DApps are very similar to the state of the internet before the launch of Netscape and sadly enough, Brave’s rapid growth may not be compensating in terms of bringing onboard new users to the broader DApp ecosystem. I have been wondering whether DApps are dying a slow but definitive death much like other trends like ICOs. As the token ecosystem transitions focus to new trends like DeFi, staking and DAOs, it may be worth looking at where the previous “hot thing” in town has been trending towards.Data Source : Dapp.review I have not included Blockstack in this. Will do a follow-up piece to cover them.
1. Number of Active DApps See Little Growth
Active DApps here refer to any DApp that had seen a transactional activity on its smart contracts in 24 hours. I had taken this to be a “daily” metric to gauge how often people engage with DApps. The reasoning for this being that in more traditional app ecosystems like software or mobile apps, the higher the open-rates of an app, the higher its utility and therefore stickiness. If individuals open a DApp only once a quarter it would be perfectly justified if the transaction size is large and profit % of that transaction is large enough. This is the case with many DeFi applications. As it stands today, there are only about ~600 DApps that are used daily. For a scale of comparison — there are over 2 million mobile apps and over a billion web sites. Granted building a DApp is hard for even seasoned developers and websites and mobile apps have seen considerable scale due to the healthy ecosystem around them (iOS and Google Play as customer on-ramps, no-code movement in websites), this number gives an estimate of where in the growth cycle for DApps we stand today.
According to Statista, there were 130 websites on the internet in 1993. A figure that jumped 30x to 3000 in 1994. By those assumptions, it is fair to suggest that we are in the early 1990s for DApps. While the comparison to mobile and internet growth rates may not be fair it is safe to suggest that broad trends in how mobile devices are being shipped with private key management (Samsung, HTC) and browsers (Opera, Brave integration of wallets) — we may see explosive growth in DApps in the future. But those days are not here yet.
Note: There are a lot of graphs that point straight to the right on “Dapp count” as a metric. The reason why I believe that is a bad metric is because DApp count is bound to constantly be increasing over time as it takes a cumulative figure since 2016. Or in other words — accounts for thousands of dead DApps that see no activity over time. A (rather sad) parallel to draw here would be to suggest planet Earth’s population is over 10 billion by including the count of everyone that died in the millions of years of our existence. Not smart, is it?
2. We Need To Talk About Developer Incentives. No, seriously
There is an active decline in the number of “active DApps” over time. What I refer to as “DApp Health” is a percentage metric I come to on basis of comparing the total number of DApps in the space with the number of active DApps at any given point in time
The way I interpret this data is as
- Ethereum’s percentage of active DApps may seem to be flatlining but likely refers to the fact that there is an increase in the number of Active Dapps over time. Why? Because the total number of DApps as a figure is bound to increase over time. If the number of active DApps doesn’t increase proportionally, it will decline as the other two have
- The figures show a clear down-trend for EOS and Tron due to their relatively younger ecosystem. While the number of DApps in both may be increasing, the retention of activity on both are in drastic decline. Too many DApps being built, deployed and ignored over time.
One explanation for this is the fact that developers often build gambling and gaming DApps to learn how to build them and scale it before moving on to more complex use-cases. Whatever be the case, it is quite clear that on a long enough time-frame, the number of active DApps declines drastically due to the lack of incentives individuals have in maintaining and marketing them. The rare exceptions being the likes of LocalEthereum that crack product-market fit fairly early on. To increase the likelihood that DApps survive and scale there need to be more incentives (in terms of grants, salaries, and direct investments) for developers to have a runway as they crack profitability. We are beginning the early stages of this trend with the pace at which firms like UniSwap and InstaDApp have been able to raise funding but there needs to be more work done in this regard. More on that later.
3. The need for breakout applications is becoming evident
EOS and Tron have substantially more users than Ethereum. This may have a lot to do with how fees are designed for transactions on them. More on that later. A “user’ here is counted as any active wallet that interacts with a smart contract over the day. The problem with this assumption is that it could very likely be bots interacting with gambling or gaming DApps. Until self-sovereign identity becomes the norm for DApps, there is no real way to measure this metric. At a little over ~500,000 users at peak when combined, DApp developers are competing for a small niche within which acquiring a large market-share comes with relative ease but scaling beyond becomes difficult. For founders, this could mean two things
- Less advertising dollars spent early on as word of mouth and organic marketing could likely be more effective given the niche nature of the space.
- High possibility of saturating markets in terms of how much market share they can acquire due to the small size of the industry today
The exception here would likely be applications that expand the existing userbase past what it is today. There used to be a time when users used to “subscribe” to AOL and Napster. For many in India, their first experience on the web was “Google”. Basically, instances where an application outgrows the ecosystem around it due to the substantial increase in value proposition it brings the end-user. Until we see those breakout applications in DApps, odds are high that this does not grow past a point.
My bet is on a content/ social media platform that sees traction due to Brave/Opera’s traction. Micro-transactions in the context of blogging combined with reputation could be the next big thing in an age riddled with fake news and mental health issues induced by toxic social media. Cent is doing some interesting work in that regard.3. Lets get transactional without robots getting in the way.
Here’s the good news. On a good day, DApps could see some 5 million transactions. Here’s the bad news — the vast majority of these transactions are likely being done by automated bots looking to game Ponzi schemes and gambling-related DApps. Keep in mind that as with all things — there’s a very clear power law at play here. A handful of wallets are responsible for the vast majority of transactions on DApps and there needs to be more research done in this regard. However, purely from the point of view of scalability, it is good to see these bots test the upper limits of what is possible on each of these individual chains. As a metric of how much dApps eat into the mainchain’s activity, it may also make sense to consider the % of total transactions on each chain that a DApp contributes. On basis of rough calculations — the figures I could come up with are for Ethereum, DApps contribute ~10% of all transactions, for EOS it is at a measly 5%, but for Tron — the figure is as high as 50%. As much as I’d like to jump to the conclusion that Tron’s DApps ecosystem is the most vibrant and healthy, a few things make me wonder if the entirety of the system is being run by bots on Tron. To understand why we have to look at two metrics. One — is the average transaction done by each user on these chains, and the other is the average volume of a transaction on these chains. Keep in mind that bot-run accounts bring the average substantially higher.
Eth is an outlier with a mere 4 transactions per user. A lot of it has to do with Gas requirements. Fees on Eth are considerably higher than those of EOS and Tron, making it a bad pick for instances where transactions are handled by a bot. However, it is in sync with what an average user would use a mobile app like one catering to food delivery or e-commerce. It is safe to assume that an average user would open a dApp 4 times to do a transaction. EOS and Tron however, live in a (botted) world of their own. Notice how Tron begins with 200 average transactions in January? I believe a lot of it has to do with (i) organic users testing the limits of the chain or (ii) a high number of bots engaging with the chain early on. As the DApp ecosystem around Tron matured, it came increasingly in sync with what a low fee chain like that of EOS looks like.What’s interesting is that at the low end, both EOS and Tron, without fee show around 20 transactions for each user. While both chains have considerable capacity to accommodate for more transactions, it makes one wonder the extent to which “adoption” is being put on display by bots instead of actual users. In not restricting bot activity in web 3.0 ecosystems, we are doomed to repeat the mistakes of early variants of the web where bots still dominate how ad dollars are looted by click bots and key metrics to gauge the health of individual businesses are gamed by inorganic, programmed bots. We will lay the case for why a higher transaction count may not be a better gauge of the health of the DApp ecosystem when we look at the volumes, but before that — consider the average number of transactions each DApp receives daily.
5. Volumes tell stories transaction counts don’t.
Although the number of transactions each DApp sees on EOS and Tron is considerably higher than that of Ethereum, when you consider the total volume they see on average — it almost converges at some points. This is not to suggest that EOS, TRON is lesser than Ethereum by any means but it could be a gauge for how organic use affects DApp ecosystems. Ethereum can move the same amount of volume to DApps with a fraction of the number of transactions EOS and Ethereum takes. This could be interpreted in two ways
- Whales dominate Ethereum There is some figures that indicate to this. As shown in my piece on InstaDapp, ~80% of activity on a DeFi product could originate from a handful of whales. This reflects on the average amount moved by each user on Ethereum too. It is ~5x higher than those of Eth and EOS
- Bots dominate EOS and TronI’d like to clarify here that the reason why I believe bots are dominating EOS and Tron is that the number of transactions done by each user on average is considerably higher than what you could expect from a normal individual. It is rare to see 50 transactions being done on average for web 2.0 apps, let alone DApps. The only way it could be justified is if the whole thing is automated. A likely possibility given that DApps in Tron and EOS are heavily focused on gambling and individuals likely have thousands of transactions on these as a way of doing low risk, low reward bets. During the early days of Bitcoin, it was common to see a high number of transactions trending towards dice rolling gambling sites.
The reasoning for my assumptions here becomes even more clear when one studies the average volume each user moves on these chains. For Eth, it is as high as $50 while for EOS it is as low as $3. Of course, these figures vary depending on the use-case. A ledger focused on enabling large-value, collateralised loans will very likely see average transaction per user that is considerably higher than one focused on micro-transactions. As such, these metrics should not be seen as stand-alone values and be looked at as part of a broader picture. Pixels that are part of a larger piece of art.
A metric I could compare to here is the average order size an amazon order would see. Today that stands at $47. The figures EOS and Tron are seeing are more in sync with what gaming applications like Fifa, PUBG and GTA 5 likely see for their in-game economies. As a parting note, what I found most interesting is the average volume each user brings to these DApps
To me, this represents the big opportunity founders have when it comes to DApps. Although the user-base is rather small, the $ amount each user brings is substantially higher daily. For those unlocking the large wallets that interact with these DApps, being creative on business models could unlock substantial value early on. The user journey from being on-boarded to transacting in large volumes is considerably shorter in comparison to what you would see with web 2.0. Businesses like InstaDapp and Juno (yes am being partial here) are already seeing the benefits of this. It will be interesting to see how user stickiness and transactional behavior converts to profitability for them in the future.Parting notes. The amount of capital moved in each transaction on a chain could be taken as a metric for how secure individuals storing value in it believe it to be. By those lens, Ethereum stands as a lead. However it still needs to solve for scaling as we have seen in the past with the likes of CryptoKitties. As for EOS and Tron — bots on their own are not unhealthy. It is in fact a requisite for testing the TPS capabilities of these chains. It is the fact that developers game their own DApp data with bots that leads to bad results. There is no way to limit it from happening in a free market, but perhaps we should remind ourselves that progress is not made on fake numbers.
I believe we are at a point in time in history where either DApps evolve to be an entirely new product category of its own. It may see certain elements of centralisation to increase utility but will likely survive. Trends in how NFTs are being used by gaming companies point towards this being the case. Other catalysts could contribute to the same. To begin with, the dev tooling environment around building DApps has evolved since 2016. Tools like Fortmatic and Arkane Network remove the headache of handling user wallets and make it relatively easier for developers to go to market. Besides, the on-ramps around DApps will evolve considerably with the likes of Wyre allowing DApp interactions soon through payment methods set up with Google Play already. Then there is the bit that investor appetite for DApps has not boomed yet. It is still focused heavily on the API layer as we search for the “AWS” for DApps. Lastly, there is also the bit that Ethereum is set to push its scaling solutions in the future. The work Connext Network and Loom have been doing will be crucial in increasing the number of transactions individuals do on Eth and reducing the fees associated with them. I wouldn’t write of DApps yet. However, there needs to be considerably more ingenuity in how business models are structured around them. For developers looking to build DApps today, there is a clear indication that more value is transferred on Ethereum and as such it may be more attractive to them. The ones that crack them, will be in a good enough position to bring value to the lives of individuals while solving for profitability early on.
On a ligher note — I leave you with this interesting clip from Mainframe on the state of DApps currently
— — — — — — — — — — — — — — — — — — — — — — — — — — — — — — —
Originally published on Decentralised.co