1. Understand fundamentals – and use them
Indices should give the buyer of financial products that are based on them simple exposure to a particular market, sector or factor. It is therefore very sensible to think through what the fundamentals of your product are. For example, if you use Machine Learning to compile an ever-changing list of coins only, there are no economic fundamentals behind your product. It then becomes difficult for any investor to understand what the product is, and therefore, it will not sell. If you are interested in how to use machine learning in finance I highly recommend to read everything you can find by Marcos Lopez de Prado – https://papers.ssrn.com/sol3/cf_dev/AbsByAuth.cfm?per_id=434076). Simplicity is key. GitHub commits, market capitalisation and daily volume might be boring but much more useful in this index context.
2. If you are not prominent / partner with someone who is
Indices are benchmarks. Ideally, they have a prominent name-tag attached to them. For example, the S&P 500 is backed by one of the biggest credit rating agencies. That is not a coincidence. So if you are a small ICO company and think anyone will pay attention to your index (unless you are perhaps an exchange with real volume) – think again. Recommendation: Partner with someone credible, pay them or share the revenue. Half is better than nothing.
3. There are professionals with a calculation engine
Calculating an index on a realtime basis (anything else, except in the case of extremely illiquid assets, is another guarantee for failure) is no trivial thing. How do you handle corporate actions or forks, for example? Recommendation: Hire someone who knows what they are doing instead of either a) not producing a realtime index or b) even worse, miscalculating it. Also, read the guidance for index products by IOSCO (https://www.iosco.org/library/pubdocs/pdf/IOSCOPD415.pdf)
4. Don’t upset the fund manager
Remember that someone will have to manage the portfolio as per your index rules. Or someone, who pays you a licensing fee, will mint financial products. Those then have to be hedged by the issuer. What does not sell: Anything that makes it challenging to replicate your strategy. For example, very illiquid coins or products. Or products that have wide spreads. If your prospective client can anticipate such issues, they will not license the index from you, and you don’t have a business.
5. Understand market data cost and contractual obligations
Your index probably will need some price data of underlying products. Be careful where you retrieve this data from. If you do so from public sources, this data is probably free but delayed. Not good enough for a professional product. If you want realtime data, you most likely need to pay for it. That is especially important if you use the data to create a derivative product (anything that uses the underlying data). Recommendation: Read the fine print of your market data provider and pay them accordingly.
6. Have your business model clear
How are you planning to commercialise your index? In the old world, you would charge a fee in basis points (not percentage) of the assets under management (AUM) with a minimum fee per year. If you plan to create the products yourself, you do not have anyone to charge. If you want to sell those products, they are considered financial products and you need a fund management license in most jurisdictions. Stating the obvious: But thinking about the business model in advance is crucial. You can avoid wasting your time and investor’s money if you do so.