3 Ways to Value Land in The Metaverse

Photo by Mo on Unsplash

Real estate in the metaverse is going nuts. The big four metaverses for real estate Sandbox, Decentraland, Cryptovoxels and Somnium are the top destinations for investors. But, before you yourself buy a piece of the action, how do you actually value a piece of the metaverse real estate?

There probably is no one size fits all sort of method. But, here are 3 ideas that could help you value the metaverse. The first method is traditional valuation, the second method is machine learning and the last method is financial modelling. In this article, I will use Decentraland as an example.

1. Traditional Valuation

This is probably the go to method for most people. Essentially, this valuation is all about location. (In real life real estate investing, type of house, number of rooms etc. matter too bit not in this case. Maybe it will in the future?) In Decentraland, close proximity to popular areas such as high traffic roads and plazas generally determines value. For example, a parcel in between other parcels may cost 5700 ETH whereas a parcel facing the road may cost 6250 ETH (as of February 2022). This method has both its advantage and disadvantages. If you are well verse in meta-verse real estate, you can more easily recognise price mismatches intuitively but if you are new to investing in the metaverse, you will probably be taken advantage of; especially, if you apply too much ‘real life real estate rationalisation’ to the metaverse.

2. Machine Learning

Given that the data is freely available (eg. you can call the Decentraland API), you could easily build a K-means model to tell you how much a piece of real estate should be valued. Essentially, land prices can be clustered together to help you identify the which clusters should be worth roughly how much. For instance, if your K-means model has 3 clusters, you can colour-code on the Decentraland map, which parcels fall in the <1000 ETH, 1001–2000 ETH and 2000> ETH clusters. The advantage of this is that the machine learning algorithm will find the ‘rules’ for you and you can more easily spot potential mispricing; furthermore, this method is more ‘automated’, so you save cognitive space for more difficult tasks such as negotiations. Perhaps, the disadvantage to this method is that you are susceptible to falling for ‘noise’ since machine learning results are seldom explicit in their reasoning and usually account for ‘noise’ as part of the calculations. (Noise in this context means unrelated variables that lead to correlations with the desired outcome)

3.Financial Forecasting

This last method relies on the cashflow of potential real estate. This method is uses a lot of inference and it susceptible to being wrong quite easily, especially with cryptocurrency fluctuations. However, a good model can estimate accurate cashflows at least in the same order of magnitude. (This means to the closest 10, 100, 1000 dollar etc). For example, in Decentraland, a parcel in front of Genesis Plaza may cost you anywhere between 300,000 ETH to 1,000,000 ETH. With such a price range, how would you truly know if the 300,000 ETH is truly under or fair priced or just anchored to the 1,000,000 ETH to make itself look cheap.

To do this well, you’ll need to do some ‘Fermi-izing’ intially. (A way of thinking to calculate difficult numbers with limited information) For example, a quick Google Search that Decentraland usually has 18,000 daily users. Without knowing what their interests are, we need to take a guess of proportions. Let’s say we are selling women’s clothing and want to buy a piece of the metaverse as a form of advertising, we can take a guess that half of the daily users are women. Furthermore, with our current business knowledge, we know that for every dollar of revenue generated, 90 cents is spent on operating expenses including advertisements, so we gain 10 cents in operating profit. So, we can guess that out of 9000 women daily users, we could gain $900 in profit (9000 * $0.1) a day. This doesn’t sound like much but this leads to an annual profit of $328,500.

Next, let’s say that we want to be on Decentraland platform for the next 10 years, we can measure the annual user growth and add that to our calculations. For instance, in this case let’s guess that the annual growth of Decentraland users will be 20% for the next 10 years and begin slumping from year 11. Next, we need to take a guess at the ETH discount rate. This is probably near impossible since the fluctuations are so high and it has not be in existence for such a long time. Instead, we’ll use the USD discount rate since our sales at this point in time are still in USD. As of February 2022, it is at 0.25% but I doubt it will be like this for the next 10 years, so I’ll raise it to 5% to be cautious. If this above scenario is true and we plug our numbers into a discounted cash flow model, then the intrinsic value of the piece of real estate could be worth as much as 32 million USD. Today’s price (Feb 2022) would be 10,136.75 ETH.

Discounted Cashflow Model

The question now is: “Is this 32 million a trust worthy number?” I would say definitely not. Metaverse advertising is probably not going to be like ‘real life’ advertising. My numbers are nothing more than guesses. Finally, discounted cashflow works well when consistency is guaranteed. Cyrtocurrency and blockchain are still quite new, so consistency is unlikely. As a matter of fact, I would definitely round down my number. In the book, “Range” by David Epstein, it tells us that in a study of professional investors, the professionals were often wrong in their predictions usually by twice of the actual sold price. Using this knowledge, we can half our prediction to 16 million. This is still a large number but it gives us an idea if paying millions for a piece of metaverse real estate is actually worth it. For example, paying 5000 ETH for a piece of road facing real estate close to Genesis Plaza may be a fair price after all. (10,000 ETH = 32 Million USD, so 5,000 ETH = 16 million USD roughly)

The Main Pitfall With All of This

All of these form of valuation rely heavily on supply and demand, which tells us that when something is in short supply, the prices go up and when something is in abundance, the prices go down. With regards to metaverse investing, you need to be on the constant look out and know how much things are going for.

Unfortunately, supply and demand doesn’t lend itself well for forecasting. You can predict demand in the next few days but it’s much harder to predict demand in the next year. For example, a time series graph can tell you the increasing number of users in Decentraland but it won’t tell you what happens if an even better metaverse created. For example, when Meta releases their own metaverse, how would the current metaverses fair? Would Decentraland’s users rise because Meta’s metaverse just isn’t that good or would it fall since Meta already has a large userbase from Facebook?


Before I recommend going out and using the methods, I do like to point out that I haven’t actually bought any metaverse land myself – just being frank as a disclaimer and metaverse land is too damn expensive for me. I have tried the methods as experiments and do believe they are better than guessing. Finally, more accurate forecasts tend to be the average of results, so you’re better off combing the results of the three methods rather than relying on just one or cherry picking the method that gives you your most desired results.

Schedule a DDIChat Session in Blockchain and Cryptocurrency:

Experts – Blockchain and Cryptocurrency – DDIChat

Apply to be a DDIChat Expert here.

3 Ways to Value Land in The Metaverse was originally published in DataDrivenInvestor on Medium, where people are continuing the conversation by highlighting and responding to this story.