Sid Kalla is CTO at fintech firm Acupay, and a freelance journalist specializing in financial technology, bitcoin and cryptocurrencies. He has invested in blockchain projects including bitcoin, MaidSafeCoin, Counterparty and BitShares (see our Editorial Policy).

In this guest feature, Kalla provides an in-depth guide for those considering investing in an initial coin offering or ICO.

Bitcoin was the first cryptocurrency released to the public in January 2009. Since then, there have been thousands of others modeled after it, with varying degrees of similarity to the original concept.

Even early on, the types of ‘altcoins’ (alternate cryptocurrencies) varied significantly from one another. Broadly, the following categories emerged:

Alongside the initial developments, a fourth type of crypto token also emerged. This differed from the previous cryptocurrency tokens in that the primary application is not for use as a currency. Instead, these tokens were specific to an application, and required to the use of that application.

The value of these crypto tokens came, not so much for their use in the exchange of goods and services, but rather the use of the underlying service that required these tokens to operate. These have been called everything from 'appcoins' to 'protocol tokens'. It is important to note that the economic and financial principles underlying these crypto tokens can be vastly different.

Several examples of these application-specific crypto tokens emerged early on:

Ethereum, the second largest crypto token after bitcoin, was founded on the same principle, allowing anyone to create arbitrarily complex smart contracts enforced and executed by the blockchain. The native crypto token, ether, is used to reward those who run the network's computations.

The flexibility with which contracts could be programmed on ethereum gave rise to many projects listing their own application-specific crypto tokens on the platform. Just in the last year, we have seen all sorts of applications, from digital management rights to in-game currencies, being listed as crypto tokens on ethereum. Some alternative blockchains, like Waves, were built specifically to help new projects issue and manage these crypto tokens.

'Initial coin offerings', or ICOs, are a form of crowdfunding and bootstrapping for crypto projects, where the founding team receives funding to develop the project from backers, in exchange rewarding those backers with crypto tokens.

ICOs are possible for all the types of crypto tokens mentioned above. However, they have become most prominent for application-specific crypto tokens.

One of the first documented cases of an ICO was mastercoin, which raised over 5,000 bicoins in 2013. In a few short years, ICOs have become the dominant form of funding new crypto projects – helping teams raise money for the development and completion of projects that they might not be able to raise via more traditional means like VC or angel investors.

In 2016 alone, ICOs raised over $100 million for their instigating projects.

Bitcoin doesn’t lend itself to easily valuation models. It is part currency, part commodity and part technology. Its price history is too new to reliably backtest models, and, unlike a corporation, bitcoin doesn’t have cash flow or earnings that can be predicted in the future.

Several attempts have been made to provide valuation models for bitcoin, from macro-financial indicators to cost-of-production models. However, no one economic or financial model has been shown to be a reliable indicator of bitcoin’s value in the economy, or has been reliable in predicting its value to the market. No one valuation model is accepted by the market participants, either.

Given these challenges in valuing the digital currency, it is doubly hard to try to value other similar cryptocurrencies (say, for example, litecoin). However, there are other types of crypto tokens that have different economic, financial and cryptocurrency-like characteristics from bitcoin that are much easier to model for valuation purposes.

There are many challenges in valuing bitcoin and bitcoin-like cryptocurrencies as outlined above. However, there are certain tokens, especially the application-specific crypto tokens, that can be modeled, not as currencies, but as economic entities. This allows us to use some of the same valuation methods for these tokens as we do for stocks.

The biggest advantage of treating crypto tokens as stock for valuation purposes is that it allows us to use existing well-established economic and cash-flow models used in equity research today. By building upon these models, we can come up with fair value assumptions of crypto tokens based on assumed input parameters.

In fact, many of these crypto tokens work in a crypto ecosystem that is powered by the token. Projects differ in structure significantly from one another, however.

For example, a project with an external source of income that gets distributed to existing token-holders will be valued very differently from a project that needs to use the token to drive a specific application, which in turn supports the price of the token. In either case though, we can start from existing valuation tools rather than try and build models from scratch.

Investors and researchers should bear in mind that the market price of crypto tokens can differ significantly from the underlying valuation models, since a large component of price is speculative in nature. This will likely decrease going forwards, as the ecosystem matures.

When we view crypto tokens akin to stock in a corporation for valuation purposes, we can begin to apply some of the economic characteristics of the latter to the former. For a crypto token holder, the value of the token comes from three primary sources, just like a regular stock:

Let’s discuss the price appreciation component first. This is present in many crypto tokens, and is driven by adoption (note that this is different from speculative price appreciation). In this case, the entire economic value of the ecosystem is modeled based on economic assumptions, and then divided among the token holders. Many application-specific crypto tokens belong to this category.

Take the example of decentralized storage startup Storj Labs. Its tokens don’t pay any dividends, and there isn’t a concept of buyback. However, if the protocol is widely adopted, the value of Storj tokens is expected to increase.

A simple valuation in this case would be to model the size of the market that Storj can represent and divide it by the total number of tokens in existence.

Some of the more recent projects have a cash-flow component (earned income) in addition to the adoption-based price appreciation. This is similar to corporations making a profit, and using that profit to grow the business via retained earnings. The ‘excess’ profit is then distributed to shareholders in the form of stock buybacks (increase the percentage ownership of each share of the company) and dividends (provide a cash distribution to beneficial owners of each share).

The more recent trend has been towards using earned income to pay all token-holders, similar to a dividend (due to regulatory uncertainty, it is possible that projects in the future will tilt more towards buybacks). Traditionally, large corporations have used both dividends and buybacks as a way to return capital to investors. However, crypto tokens usually opt for just one of these routes.

Let’s look at the different models adopted by crypto tokens over the years.

The original BitShares (before an inflationary fork) was the first example of a crypto token with buyback-like characteristics. The number of BitShares tokens was capped. Every transaction needed a fee, paid in BitShares. A part of this went to the delegates in its proof-of-stake model, and the rest was 'burned'.

Burning is similar to buybacks in that the total supply of tokens decreases, making everyone hold a slightly higher percentage of the total supply than previously. As adoption grows, the fees (which are always paid in BitShares tokens) increase, so the number of tokens burned increases, thus sending the capital to the token holders.

Augur adopted clean dividend-like characteristics for its tokens, called 'reputations' (REP). All the fees generated by a market get distributed to the REP holders (in ether) that correctly voted for an event (ie they voted with the consensus). As the number of markets rise, the income generated by the REP holders should also rise.

In contrast with some other projects, this capital allocation decision is done via smart contracts, without any one group’s involvement, which makes it easier to value.

Iconomi, a decentralized fund management platform, has adopted  dividend-like characteristics, but with separation into retained earnings and distributable earnings. This most closely resembles capital allocation decisions by corporations today.

The project will generate earned income (in ether) in the form of revenues generated from asset management and fees. A portion of these fees are retained by the company itself for ongoing operations and expenses. The rest is sent to the token holders. How much of the total earnings goes into distributions versus retained earnings is left to the discretion of the team.

We are yet to see crypto tokens with a capital structure consisting of both debt and equity. Such a crypto ecosystem would need strong earnings potential and the ability to scale with an increase in capital.

At this stage, we do not see such characteristics but it is possible and even likely that future crypto ecosystems will use debt intelligently to enhance returns. Such a system would also enable the setting of market-based interest rates for crypto tokens, and perhaps also provide clues to a risk-free rate.

Different crypto tokens lend themselves to the use of different valuation models, but we can use many of the tools that equity research already provides us with.

There are usually two steps involved with valuing a crypto token: modeling the market size and the extent that the project can reach, plus how the market reach translates to returns on individual crypto tokens. Economically, dividends and buybacks are similar. However, investor assumptions about the market adoption can vary widely.

The estimation of the size of the market can be top-down or bottom-up.

As an example of the former, consider Storj from our previous example. In the top-down approach, one would start with the total size of the addressable market, which in this case is file storage. Then estimate what percentage of that market can be captured by Storj in a given time frame.

In the latter approach, bottom-up, one would start with the existing market size of Storj and make assumptions on the rate of growth of that market. Sometimes a combination of these approaches will need to be used.

Addressing the size of the market itself can be challenging. Investors should look at information from existing companies. In the example of file storage, one can look at data from Dropbox and Box, two existing multi-billion-dollar companies in the space, and estimate the size of the market. From there, we can make an assumption about what percentage of the total market can be captured by a solution like Storj.

As an example of a bottom-up approach, consider SingularDTV, which raised $7.5m in its ICO last year. The project's investor communication (since removed, but archived) broke down revenue projection numbers per project. For example, they assumed a documentary would generate 200,000 paid views, with each view paying $2.60.

This is a typical bottom-up approach where individual components are valued, and then the value of the whole enterprise is estimated based on the sum of the parts. Investors should generally not agree with the numbers provided by the teams themselves, but can still use the framework for their valuation purposes.

Once the market size and earnings estimates are obtained, investors can use basic valuation frameworks like a dividend discount model, which gives a present value based on projected earnings and dividends, or a relative valuation model that compares a crypto ecosystem to existing companies that have well-defined pricing, for example due to the company being traded on the public markets (for example, comparing Storj to Box).

A sum of the parts valuation is also quite common. As an example, Iconomi has three distinct products that generate revenue: a crypto index-like fund with management fees, a crypto hedge fund-like product with both management and performance fees, and a fund management platform with usage fees. Each part needs to be modeled separately to arrive at the final valuation of the tokens.

An important factor to realize when valuing crypto tokens is that they naturally present higher risk than even regular startups. The data doesn’t support very long-lived crypto tokens. The markets are new, and demand can be flaky. The going-concern assumption may need to be revised to instead value these tokens based on a finite lifespan.

Investors should therefore demand a higher risk premium for investing in these tokens, and demand a higher margin of safety.

Due to the very short history of such crypto tokens and crypto-economic systems, there are several challenges that investors face when trying to value these projects.

First, the short history of crypto tokens has generally shown an even shorter lifespan of many of the projects. This is especially true because projects present a big principal-agent problem due to most of the tokens being given away through ICOs.

This is different than a startup that usually raises money in a series of different rounds over several years. In due time, crypto tokens will likely be sold to the public in several stages, depending on different milestones being hit by the project.

Second, the success rate so far for most crypto projects is not very encouraging. This makes projecting cash flows and earnings into the future very difficult, because the going concern assumption may not be valid. Models like the dividend-discount model assume dividends into perpetuity. If the project has a lifespan of five years instead, the valuation can be very different.

Third, there is no set risk-free rate for the crypto markets. This makes it harder to discount future cash flows to the present. There are some market-based interest rates on exchanges that lend to short-sellers, but that data is hard to come by, and it is hard to infer a risk-free rate from the data.

Fourth, there is usually some level of systemic risk associated with the crypto markets that cannot really be diversified away. The industry is too nascent for that. Therefore investors take on both project-specific risk and market risk when they invest in this sector.

The systemic risk is very hard to predict, due to the short timespan, and is unique to the industry. Everything from hard forks to new crypto attacks are a source of systemic risk that traditional investments don’t suffer from.

Fifth, many projects are interdependent, which causes dependency risk to projects. For example, a crypto project built on ethereum will be affected by things happening in ethereum, like a bug found in a compiler, or an attack on the ethereum network.

And, further, as layers of the ecosystem build up, this dependency risk deepens.

For more details about the developing 'ICO' market, read CoinDesk Research's latest report.

Measuring spoons image via Shutterstock