Alternative data for financial services, such as satellite imagery, logistics data, and social media feeds, has been getting a lot of attention recently. Like any trending topic, its uses and benefits can be hyped up a bit but if the right plumbing and creativity is in place, those benefits can be realized. This post helps to understand the plumbing by describing the component parts of the alternative data ecosystem. This framework will help you to navigate your way from an inkling of an idea to an actionable strategy while the window of opportunity is still open. See the webinar, “Put Alternative Data to Use in Capital Markets” for a demonstration of this process.
The Ecosystem in Four Parts: Who, What, Why, and How
Putting alternative data to use in financial services is not just about the end-user experience. There are dependencies that only a larger ecosystem of partners can fulfill. An understanding of these interconnected elements will help you to understand your options. There are four parts that answer the who, what, why, and how of alternative data (but not in that order).
Part 1: The beneficiaries of alternative data strategies.
Why consider an alternative data strategy? Much of the attention paid to alternative data has been around creating trading strategies with an edge (generating alpha). It also applies to risk management and enhanced regulatory oversight. Two categories of beneficiaries come from this, those whose responsibility it is to mitigate risk and those whose remit it is to generate revenue.
- Risk mitigation: A compliance officer correlates suspect trade flow patterns with newsletters and messages that have been analyzed for misleading statements. Benefit: suspicious trade activity is identified inclusive of an evidence trail.
- Revenue generation: A fund manager, portfolio manager, or a chief investment officer enriches traditional financial analysis with multiple alternative data sources to identify unique and localized industry trends. Benefit: a novel strategy is founded on a comprehensive knowledge base.
Part 2: The originators of the data.
What data is used? Alternative data types that ultimately provide value are ones that answer distinct questions. In other words, the selection of data should be evaluated from a fit-for-purpose point of view. A strategy responds to multiple questions or scenarios, therefore you will most likely pull together multiple data types. Here is an example relating to the evaluation of retail stores.
- Satellite imagery: How full are the parking lots? Are they getting busier or less so?
- IoT mobile phone tracking: What is the turnover of pedestrian traffic into and out of the stores? Where are the customers from?
- Social media: What do people think of the experience (positive or negative sentiment)?
The originators of the data also prepare that data, which is a key point because the raw data has been analyzed, prepared, and packaged for you. For example, you would not use NASA photos alone. You would use the data generated by a provider’s neural networks that determined the boundaries of a parking lot, that the rectangles within that boundary are probably cars, and over time, the volume of cars have gone up or down. Likewise, individual tweets don’t tell you what people think of a brand, it is the sentiment analysis on large volumes of tweets, Facebook postings, and blog posts that enhances your decision-making.
Alternative data sources are fit for purpose and the quality of each is determined by the company that originated the data and prepared it. Some are better at it than others. In the above example, three different types of data were cobbled together for one strategy.
Part 3: The providers of modern big data platforms that ingest, store, and process the data.
How does this all work? The mechanics of putting alternative data to use can be categorized into three general areas. This part of the alternative data ecosystem belongs to the providers of modern big data platforms.
- Data ingestion: The information you depend on will be coming to you via an array of live data streams and in large batches.
- Data storage: Huge volumes and high varieties of data from disparate sources will need to be combined and stored, some of which you may not use in the end but you need it just the same.
- Data Analysis: Machine learning and advanced analytics will be key to gaining deep insight into the data regardless of its original purpose. Collaboration requires flexibility of deployment so that any type of analytics tools that suits the team can be used.
It is crucial to understand that an alternative data strategy is about enrichment and additional insight. You are using new and alternative assets to provide additional signals that can be evaluated and tested. It is not an end in itself. Advance analytics described above is key to making those connection and putting alternative data to use.
Part 4: The subject matter experts that turn an inkling of an idea into an actionable strategy.
Who puts it all together? In an earlier blog post, “How to Efficiently Put Alternative Data to Use”, I concluded that it is the subject matter experts (i.e., portfolio managers) who are best equipped to derive meaning (and thus value) from new types of data. They understand the historic relevance of a topic and quickly discern nuanced disruptions that could flag an opportunity. Collaboration with subject matter experts from other disciplines will increase the efficiency and efficacy of their work. This includes client investors, whose feedback and questions will provide additional signals from which to test assumptions that either support or debunk the strategy.
In summary, you need to get an understanding of the alternative data ecosystem and technology framework so that you can evolve an inkling of an idea into real value while the window of opportunity is still open. For more information, visit Arcadia Data’s alternative data resource page for insights gained by working with clients and partners across the alternative data ecosystem through blog posts, expert interviews, and demonstrations.