In today’s world, it’s very difficult to gain an extra percentage point of return on your investments. Traders and investors on Wall Street are aware of how every trader has access to the same information at the click-of-a-button. The fast access to vast amounts of data has led to the rise of algorithmic trading that has automated the execution of trades with little or no human intervention. So, traders have turned to alternative data to gain the edge in the marketplace. It is just a matter of time before decision makers across all industries rely on alternative data to augment their decisions.
Large companies are beginning to see alternative data as a business opportunity and have lent credibility to the trend. NASDAQ recently acquired alternative data provider – Quandl.
What is Alternative Data?
Alternative data has been around for a long time. Traders looking to gain an edge in oil trading before the release of U.S. government data on oil supplies have purchased and used grainy images of oil storage facilities in Cushing, Oklahoma to estimate the amount of oil stored in its massive storage tanks. It turns out that the lid on top of these storage tanks float on top of the oil and thus the aerial images of tanks showing the lid position or depth can be an accurate predictor for oil in those tanks. Similarly, investors in retail companies have analyzed satellite images of mall parking lots to estimate the number of shoppers in the mall.
So, what is alternative data? According to AlternativeData.org, alternative data refers to data used by investors to evaluate a company or investment that is not within their traditional data sources (financial statements, SEC filings, management presentations, press releases, etc.). AlternativeData.org offers datasets that cover many industry segments.
How Retailers Use Alternative Data
Alternative data gives key decision-makers a fuller, more accurate picture of their operating environment. For example, retailers looking to open new stores have relied on a traditional set of data such as customer segments, trade area extents, locational convenience, etc., to select a site for their new store. The alternative data in this context could be an analysis of buying preferences over the past year in a specific location using credit card transactions. Retailers could gain an additional, more granular perspective into the purchasing power of consumers within their trade area by using data on the growth of debt within the community in the last year. If debt levels had grown rather than fallen in the community overall in the past year, that could be a sign of tepid demand when the store opens. Alternative data can further reduce the probability that a new store would under-perform and lead to closure.
Retailers could use alternative data to customize and localize product selection at the store level. For example, sentiment analysis of product postings on social media platforms can help drive inventory levels at the stores. This type of analysis would have enormous ramifications to the overall supply chain. Retailers would have to build a stronger, deeper relationship with their vendors to ensure they could be nimble in responding to changing tastes. Vendors may have to reinvent their manufacturing processes to reduce batch sizes and turnaround times. Retailers and vendors may have to share more data with each other than ever before to ensure better decision-making.
Traditional Data Vs. Alternative Data
The question is not where alternative data can be substituted for traditional data. The appropriate question is how alternative data can be combined with traditional data to improve the quality of my decisions. First, you need to answer these questions:
- What are some of the alternative data that could be applicable to my industry?
- How can the data be customized to meet the needs of my business?
- What levels of granularity should the data be gathered?
For example, a multi-national oil company may use gasoline tax rates acrosseach country to forecast demand for its product. The oil company’sgranularity could be at the country level.
- A retailer’s granularity could be each zip code in its trade area.
- Which statistical models have reliable predictive power?
- Should alternative data be bought from vendors or could the company effectively gather the data that they need?
Alternative datacan be very expensive to procure from vendors. So, companies maywant to assess the effectiveness of the data before buying. It mayalso want to investigate how the data can be customized for the specificneeds of the company. A company could also put together a data team that could help gather some of the data from open sources withoutpaying a dime.
Importance of Statistical Models
Not all alternative data will be useful in decision making. It will take companies a lot of time and effort in identifying sources of alternative data and then putting them to the test via building sophisticated statistical models. The alternative data needs to be customized for your business needs and then combined with traditional data to build those models. Once you have those models, it must be put to the test in the real world by making decisions based on the model. For example, a retailer may use a combination of traditional and alternative data to open a new store or localize inventory for a store. The retailer may have to put the model to the test multiple times before deciding on its effectiveness.
Once a company has found a model that is effective in predicting outcomes, it now has a standard. But the company must work hard to keep the data and the model fresh and be on the look-out for changes in the environment that would change the effectiveness of the alternative data used in the model.
All companies need to look for alternative sources of data that could improve their decision making and give them the edge in growing revenue and profits. In short, alternative data can open new opportunities but needs to be used wisely.