Executive Summary

Demystifying Alternative Data stat bar

Alternative data has been around for at least a decade, but an increase in adoption is making it less “alternative” and a more essential part of portfolio construction. The word is out that these new, unique data sources can add valuable explanatory power to both quantitative and fundamental investment models. Intuitively, investors now understand that alternative data strategies can yield important information. Tracking shipping data, for instance, may offer intelligence about trade flows and corporate performance, measuring foot traffic at shopping malls can inform retail sales forecasts, and vehicle registration data can provide early insight into the automotive sector.

Until recently, the usage of alternative data has been confined mainly to the realm of quantitative investment managers, as these firms were best able to obtain, clean and process this data. Now, however, alternative data is beginning to go mainstream, with increasing interest from fundamental and hybrid asset managers. More traditional financial information vendors are expanding into this market and developing offerings tailored to this new segment.

In this Greenwich Report, we will look into some of the recent trends in the adoption of alternative data and demystify its usage by taking a deep dive into how shipping data can help chart a course for alpha.

Demystifying Alternative Data stat bar

Executive Summary

Alternative data has been around for at least a decade, but an increase in adoption is making it less “alternative” and a more essential part of portfolio construction. The word is out that these new, unique data sources can add valuable explanatory power to both quantitative and fundamental investment models. Intuitively, investors now understand that alternative data strategies can yield important information. Tracking shipping data, for instance, may offer intelligence about trade flows and corporate performance, measuring foot traffic at shopping malls can inform retail sales forecasts, and vehicle registration data can provide early insight into the automotive sector.

Until recently, the usage of alternative data has been confined mainly to the realm of quantitative investment managers, as these firms were best able to obtain, clean and process this data. Now, however, alternative data is beginning to go mainstream, with increasing interest from fundamental and hybrid asset managers. More traditional financial information vendors are expanding into this market and developing offerings tailored to this new segment.

In this Greenwich Report, we will look into some of the recent trends in the adoption of alternative data and demystify its usage by taking a deep dive into how shipping data can help chart a course for alpha.

Alternative Data Usage

The earliest adopters of alternative data began using these new data sources over a decade ago, but only in the last two to three years has it really come under the spotlight. As traditional active managers struggle to beat their passive competitors, many are turning to alternative data as a way to enhance their models and differentiate their product offerings.

We find that nearly half of investment managers are currently using alternative data, with another quarter planning to do so in the next 12 months.

Length of Time Using Alternative Data

"Power users" of alternative data quote

Among current users, most have been implementing alternative data into their models for over four years. But few firms consider themselves to be high-level users, which we define as subscribing to multiple alternative data sources, which form a key part of their investment model. These long-term users are the more niche quantitative firms, sometimes known as systematic investors. These “power users” of alternative data may have it deployed across multiple strategies or have specific investment strategies built entirely around alternative data. These firms will often want the data in raw format, in order to extract as much alpha as possible from the signal within an optimized portfolio-construction process.

Use of Alternative Data

More firms are at the low-usage end of the scale, with occasional ad-hoc querying of online tools or using datasets built into their market data terminal. This probably has more to do with experience than usefulness of the data, as nearly all firms who have been using alternative data for less than four years claimed very low usage. On the other hand, nearly all firms who have been using this data for four years or more describe their usage as high or medium.

We define medium usage as subscribing to at least one alternative data source that is regularly used as part of the investment process. As the market evolves, we believe this level is where most institutional investors will be positioned, or perhaps somewhere between medium and high usage—subscribing to multiple sources but only using one or two per strategy.

The majority of active firms are benchmarked to an index, so it is appropriate that alternative data is used to augment and enhance existing models, as opposed to defining a new investment strategy. Over the next few years, helping investors move up the spectrum, from occasional usage to fully integrated into the investment process, will be the challenge for alternative data vendors.

Our data shows significant increases in alternative data budgets year over year. In 2018, budgets increased by 52%, which comes on the back of an increase of 76% the previous year¹. This is certainly rapid growth, but there is no indication of a slowdown. Digging further into the data, we note that those firms that have been using alternative data for four years or more have budgets three times larger than those of “newbies”—even after removing outliers.

Growth in Alternative Data Budgets

Growth in the alternative data space is going to come from two areas: new users who are rapidly adopting this data (56% of investors have added new alternative data sources in the last 2 years), and from existing users as they get more comfortable, demonstrate the value of alternative data in one or two areas and then integrate it more deeply into their strategies.

A large majority of investment firms, 72%, told us alternative data enhanced their signal, with over one-fifth saying they got over 20% of their alpha from alternative data.

Value of Alternative Data

Integrating Alternative Data

A strong message from alternative data practitioners is that the tools and techniques for analyzing the data are as important as the data itself. Alternative data, by its nature, is different from typical sources of financial data. In raw format, it is often unstructured, requiring extensive scrubbing, normalization and backtesting before it can be effectively utilized.

Evaluating Alternative Data

62% of users prefer alternative data to be packaged in some way

Reflecting this, 62% of users prefer the data to be packaged in some way, as opposed to delivered in raw format. Recognizing that the long tail of investment firms adopting alternative data want it to be more easily consumable, many vendors in the space are creating new products with normalized data—pre-packaged signals delivered alongside more traditional sources.

Data Delivery Preference

When thinking about the preferred level of support alternative data buyers want from their vendors, a large majority, 83%, want some assistance with ingesting and processing the data. Most want access to experts to explain the nature of the data, while others want technical or data science support. However, it is likely there will always be a minority of firms who prefer raw data that they can customize, enabling them to extract as much alpha as possible.

Preferred Level of Support

Taken together, these findings suggest that investment managers, particularly those with a more fundamental strategy, need guidance and support in incorporating these new data sources into their models. This type of collaborative service model will be essential, as the industry seeks to attract the early majority of alternative data users.

Sources of Alternative Data

Over half the investment professionals in our study report that they would seek out specific alternative data sources that they think will augment their models. Others tell us they rely on peer feedback, suggestions from a vendor, or that they search more opportunistically for new and interesting data sources.

We will look for new ways/angles to uncover more alpha

The most popular type of alternative data noted by study participants is web-scraped data, followed by crowd-sourced data, social media sentiment, and credit card or point-of-sale data.

Popular Alternative Data Sources

Usage of Alternative Data Sets

This is consistent with other studies Greenwich Associates has conducted over the last few years. It is clear that the market values these categories of alternative data, and we have discussed them in detail before². The popularity of these alternative data sources may also be related to the fact that they are natively digital—meaning the data is harvested from electronic or online sources and, therefore, more easily integrated with data science tools and techniques.

Data sources that are not natively digital are showing lower utilization quote

As more investors begin to adopt popular alternative sources, such as web-scraped data or social media sentiment, it is possible that the alpha will be diluted. Indeed, 87% of respondents believe that the edge offered by alternative datasets will decline over time, as more firms start tapping into the alpha, weakening its power. However, most people think this will take three years or more.

It is possible, therefore, that we will see a rotation from some of the more popular sources of alternative data into emerging sources that are not natively digital and require different data collection techniques—such as “bills of lading” data or car registration data. As these data types are showing lower utilization, they may possess more unique alpha content.

We have to keep looking quote

It should also be noted that study participants broadly agree that the market is becoming flooded with new alternative data sources, which makes finding unique sources increasingly difficult. This is echoed in anecdotal reports from industry practitioners who maintain that the hype around alternative data has attracted many companies that believe the “data exhaust” from their core businesses may have value to the investment community.

Deep Dive: Charting a Course for Alpha Using Shipping Data

Maritime transport is vital to the global economy, as over 90% of global trade is carried by sea. U.S. Customs and Border Protection requires disclosure of shipment details for all vessels entering or leaving U.S. ports, using documents called bills of lading. This data is often filed electronically via PDF documents, but it may also be filed manually via paper documents.

Sample Bill of Lading

Sometimes third-party logistics companies are listed as the importer/exporter, so careful analysis needs to be performed to avoid misinterpreting the data. In addition, bills of lading data is often augmented with data collected directly from ocean carriers. Alternative data companies must, therefore, invest significantly to ensure expertise in data collection techniques.

Bill of lading data comprises transactional level details about waterborne trade for numerous countries. This data includes the ship identifier, origin and destination, description of goods, estimated value and other information. This type of information can yield important information for institutional investors:

Country level forecasts: Increasing or decreasing trade flows are an important component of a country’s GDP growth.

Sector level forecasts: By tracking goods descriptions and mapping to sectors, it is possible to gain insight into global sector performance.

Company level forecasts: When a company is increasing exports, it is generally a good sign of corporate growth. But with bills of lading data, it is also possible to track if a company is increasing imports of raw materials, indicating a strong pipeline of new orders. If an automotive manufacturer increases imports of steel, for instance, it can indicate an increase in production.

In the following graph, we see how the imports of a clothing retailer (Michael Kors), can be a leading indicator for future sales growth.

Michael Kors

Vendors

The growth in the market from quantitative funds to hybrid and fundamental funds is causing a shift in the vendor landscape. While quant funds are more likely to use independent alternative data vendors, overall we are beginning to see the large, integrated market data vendors become the go-to vendors for the wider investment community. Indeed, the top five sources of alternative data in this study were all large, integrated market data vendors, with smaller independent data vendors recording much lower penetration.

With new adopters preferring the data to be scrubbed and packaged quote

This is a sign of how the industry is maturing, and the extent to which large market data vendors have upgraded their products to incorporate alternative data sources. This also makes the process of acquiring alternative data seamless. As investment managers already have licenses with at least one market data vendor, these would logically be the first call for a portfolio manager looking for alternative data. In addition, with new adopters preferring the data to be scrubbed and packaged along with a support framework, it fits in more closely with the market data vendors’ existing business models and delivery methods. Just as FactSet acquired Revere, IHS Markit acquired Automotive Mastermind and Nasdaq acquired Quandl, we should expect further acquisitions of alternative data vendors by larger, integrated data platforms.

Obstacles to Adoption

Analysts and portfolio managers cite a number of obstacles inhibiting their usage of alternative data. Prohibitive cost and difficulty in quantifying the value are the top two. In some ways, these are related. Among respondents, 83% did not have a formal methodology to estimate return on investment (ROI). Without a good idea of the value of something, it is harder to justify the cost.

Key Obstacles to Using Alternative Data Sets

Other obstacles include difficulty in cleaning and integrating the data, and a lack of internal resources. Integrated market data vendors are well positioned to address many of these obstacles: Economies of scale can allow for explicit costs to come down, a more packaged product can alleviate data cleaning, and integration concerns and a robust support model can mean fewer internal resources are required.

Conclusion

Usage of alternative data is expanding beyond the quantitative asset managers who pioneered its use a decade ago. Recognizing this evolution, larger data vendors are expanding their offerings in the space and standardizing the data to make it easily digestible for the more fundamental asset managers. Data alone is not enough. To truly grow this segment, vendors will need to provide normalized data with embedded tags, analytics, integration with existing platforms, and a data support model.

To truly grow this segment, vendors will need to provide normalized data

The types of alternative data being used are likely to shift also. Currently usage is concentrating among the top half-dozen or so sources of alternative data. As more investors adopt these sources, aided by vendors providing more productized offerings, it is possible that the alpha from these sources begins to dilute over the next few years, causing these institutional investors to shift their focus to less utilized sources, including those that are not natively digital.

Despite the rapid growth in the market recently, we believe we are still in the early innings when it comes to alternative data adoption. We expect budgets to continue to increase significantly over the coming years, as more firms come into the market and usage among current users increases. By the time we reach the final innings, alternative data won’t be alternative any more … it will just be data.

Richard Johnson

IHS Markit logo

1https://www.greenwich.com/equities/buyers-guide-alternative-data

2See for example: https://www.greenwich.com/blog/alternative-data-action-web-scraping

Methodology

Between December 2018 and February 2019, Greenwich Associates interviewed 42 investment specialists at quantitative, fundamental and hybrid asset management companies. Respondents were asked a series of questions about their usage of alternative datasets.

Demographics