Identifying Comparable Stocks for SensibleAI Features
Introduction
The purpose of this article is to provide SensibleAI Forecast implementers and architects with a framework to answer the question: “How do I decide what stocks make sense to experiment with as SensibleAI Forecast features?” SensibleAI Forecast comes equipped with the ability to include a variety of publicly traded securities, including domestic and international stocks, ETFs, indexes, and cryptocurrencies, as features. When considering stocks specifically, determining which public companies are relevant to experiment with as a feature is a highly subjective and ambiguous process since no two companies are identical. Below I will outline several items to consider when using stocks as a feature.
Identifying Comparable Public Companies
As mentioned above, no two companies are identical, so we have to look to key comparable characteristics to begin developing a universe of comparable companies that we might expect to perform similarly (i.e., grow at a similar rate, react similarly to macroeconomic trends or news events). There are three specific characteristics to consider when starting your search:
- Industry
- Geography
- Size
Industry
Industry is likely the first characteristic that comes to mind when considering whether two companies are comparable. Here we are looking for public companies that are either direct competitors or have a similar product / service offering and customer base since the stock prices of industry counterparts tend to move similarly. This exercise is akin to creating a competitive landscape for your client – who are they competing against to win customers and increase market share?
Strategies for identifying industry counterparts:
-
Discuss with the client during discovery who their competitors are
-
Look at the holdings of industry-specific ETFs or mutual funds
-
Industry-specific ETFs exist for nearly every industry where there are publicly traded companies
-
Every ETF or mutual fund is made up of individual stocks that will be listed on their website under “Holdings” or “Positions”
- For example, if you are working with an insurance company client, look at the holdings of various insurance ETFs to see what companies asset managers and institutional investors are including in their definition of the insurance industry
-
Large ETF and mutual fund providers include SPDR, Vanguard, Blackrock, iShares, State Street and Invesco
- The ETF Database is an ETF database and a great place to start your ETF search
-
Note: if the ETF is publicly traded on one of the exchanges supported by Financial Modeling Prep, you can consider including the ETF itself as a feature
-
-
If the client company is publicly traded or was previously public, check the Investor Relations section of their website for investor presentations. These presentations often include information on the company’s competitive landscape and direct competitors
-
Search for third party research articles from providers like McKinsey and Gartner or trade publications
-
Do a quick search through the company’s website to confirm that their product / service offering is similar to that of the client
- Two companies may both be in the water industry, but if one is a public utility and one manufactures wastewater treatment equipment, they would have a very different product offering and customer base and therefore may not be very comparable
Once we have an initial list of publicly traded industry counterparts, we can then start to trim it down based on geography and size, which are discussed next.
Geography
The next item to consider after creating your initial list of industry counterparts is geography. Companies that operate in different geographies are exposed to different consumer preferences, macroeconomic trends and regulatory environments that could impact how comparable two companies are. An oil and gas company based in Houston may not necessarily be comparable to one based in Asia or the Middle East for the reasons listed above. It is also important to not just consider the physical location(s) of the business but also the geography of the client’s customer base – do they sell globally to international customers or to one specific region of the US? Answering these questions should help you eliminate companies that are potentially incomparable due to geography.
Size
The final consideration of note is size. There are several measures of size to consider including total revenue, total assets and number of stores / facilities / plants, with companies that are similar in these categories generally being more comparable. Note that there is some overlap here with geography when considering number of stores / plants / facilities. Though an extreme example, consider whether a local fast food burger chain is comparable to McDonald's, the largest fast food chain in the world with over 40,000 locations. Though both are in the fast food industry and may compete in the same market where the local chain is based, the sheer size and scope of McDonald’s operations make these two businesses very different.
Conclusion
When evaluating whether two companies are comparable for the purpose of FOR feature experimentation, one should consider if these companies share a similar industry, geography and size. The framework above will help you to first identify a list of companies that operate in a similar industry and then whittle that list down to businesses that also have a similar geographic footprint and size. Additionally, this framework is easily explainable to a client if they provide pushback or want additional detail on why a given stock was chosen as a feature.