Can “Alternative Data” Sources Complement Your Investment Process?

Written by Gareth Witten via Investor Literature 

I recently finished a workshop for a company where I took the executive team through some scenarios for their industry. I can’t share the exact company or scenarios with you but I hope to give you a sense of how this can be applied in our world of investing. This is a new concept to me but I think may be extremely valuable going forward.

Many asset managers, and particularly hedge funds, have used “alternative data” to attempt to understand how a business makes its money. For example, by using satellite data to monitor crop growth will help give farmers a better idea of current and future crop production and projected revenues for fertilizer firms. This is especially important when companies are using multiple sensors to monitor products and consumer behaviour etc.

This partly came to light during a court case in 2015 by Ashbury Heights Capital. They accused a data partner of underreporting revenue and misusing intellectual property. At the core of the dispute is a methodology that they designed to predict “how stock prices would move, largely based on economic information and extrapolation from company network of customer-supplier relationships”. They claim that they help investors accurately model the impact of disparate events of publicly traded companies by using supply-chain analysis. For example, it might help a company predict what would happen to Intel’s share price if IBM sees sales decline. In a 2010 whitepaper by its founders they said that the system uses similar tools to how Google ranks pages on the internet (alongside graph theory) to provide meaningful and consistent analysis of the economy over the long-term and ultimately provide a basis for alpha.

The real value was its ability to estimate values that were not publicly available. What Ashbury did was form a partnership with a data company called Revere Data. They then licensed Ashbury’s system to several hedge funds with its propriety data set. Ashbury alleged that Revere broke the terms of its contract by not reporting all the revenue it received from licensing and misused their intellectual property. What Ashbury does is use graph theory to index data on the impact of correlations between sales drivers, for example, and the impact on other firms. For example, if Starbucks’ revenue in Canada goes up due to the introduction of a new product line then how does this impact Tim Hortons’ sales and how does it change consumer behaviour etc. If Starbucks has a good sense of all its consumer behaviours (gathered and analyzed by their internal data analytics team) then it is straight forward to see that this is only one part of the puzzle. The second order impact required is by, for example, monitoring social media or parking lot data and its impact on sales etc.

This may seem like short-term trading strategies but it may actually unlock some of the potential strategies for a firm. This is nothing different from how I was trained as an analyst when there were far less firms listed (too long ago now) and an analyst could easily “kick the tires” by visiting the firms or, better still, talking to its customers. Because there are many more firms to analyze, analysts tend to hide behind the reams of data required to make an informed decision about whether to invest in a company. Because data sources are (slowly) becoming more accessible and the links between data sources are improving, there are no more excuses for analysts to access alternative data sources. There are several things that this highlight to me:

  • The need for accurate data is absolutely vital and a real valuable asset. This may be a source of analytical edge for a hedge fund or pension fund. Planning this data strategy will become more and more important over time and integrated further into an investment process.
  • The use of an alternative data set to access a firm’s edge is becoming vital as they navigate an evermore competitive landscape. We can think of Alternative Data as a source of data in addition to financial data or the intangible data that we analysts beat to death every day (see figure below). This became even more apparent when I was running this workshop for this firm. Generally, the firm’s moat is clear to analysts but how it can extend that moat internally goes well beyond just improving its internal inefficiencies.

The workshop usually starts by me running through various short and long-term trends. The first is to provide an “outside view” (following Daniel Kahneman) of potential “clockwork” trends and “cloudy” trends. The challenge with this is understanding how a firm can use that data substantially to make decisions. The key process attempts to unlock the various scenarios and uncover some of the factors constraining firms going forward. This may seem simple to do but it is not and I would rather be approximately right than precisely wrong! The use of alternative data sources is still very much untapped source of value and as we all get more and more “connected”, we will better judge the second order and maybe even third order impacts of decisions and behaviours.

The short answer is very much that alternative data sources can definitely complement your investment process but its up to you to use the relevant data for the edge that you want.

If you are subscribed to the CAASA mailings you will continue to receive these short articles and other topics of interest via email, or subscribe on our Stay Connected page, here.

For longer articles, please visit InvestorLit Reviews via

InvestorLit Reviews is a subscription-based service for reviews of investment literature and topics, with an emphasis on concepts that can help investment performance. The literature reviews are intended to provide brief and readable versions of the original papers or topics but also point out tangible ideas. This is about sense-making and implications for you, the investor.