Sometimes the best business intelligence comes from the least likely data
Alternative data has been a buzzword among investors for several years now. By leveraging the insights available in non-traditional datasets, hedge funds can reap massive profits off insights that aren’t easily available to the average investor or anyone looking at traditional markers. Hedge funds so value alternative data that, according to JP Morgan, asset managers were already spending $2–3 billion on alternative data in 2017 and those investments have increased 10–20% a year since.
But it is not just hedge funds who profit from alternative data. All kinds of businesses can use alternative data to increase returns — and the most innovative companies have realized that alternative data holds the key to maximizing their competitive advantage. In fact, data science analyses that leverage alternative data outperform benchmarks about 13% better than traditional approaches to analytics.
What exactly is alternative data?
Put simply, alternative data is any data used to make an analysis that would not traditionally be used to make that decision. Alternative data includes proxy metrics that can stand in for factors that are usually difficult to measure. They can also include information originating from unofficial sources that individuals can use to gain insight into an analysis. Alternative data provides both new types of business intelligence and new ways of understanding intelligence that could be gained from traditional data.
Alternative data draws from non-traditional data sources so that when you apply analytics to the data, they yield additional insights that complement the information you receive from traditional sources.
Krishna Nathan, CIO of S&P Global
Alternative data sources will vary widely depending on the industry and type of analysis being done. Investors have used everything from credit card transactions to location data from cell phones and scraped from the web. Even tracking the private jets of companies has been used to assess whether or not to invest in a particular company.
Alternative data can be leveraged just as effectively outside of the investments world. Retailers have used data taken from satellite images to decide where to open new locations. Fintech companies have used cashflow markers and academic history to assess the creditworthiness of people without a credit history. Even travel companies have used alternative data scraped from the internet to decide which amenities to offer and where.
At Evo, the supply chain and pricing AI company where I work, alternative data is a significant part of why our supply chain tools increase inventory efficiency by at least 10%. We use everything from scraped web data to weather to improve our analyses. One of my favourite examples of alternative data that we had great success with was using store manager opinions on trends. Essentially, we allowed store managers to request particular items they believed were most likely to be popular in the upcoming sales period. It was a way to get a more local, granular measurement of trends and popularity, using the managers as a proxy — and it worked. This increased the accuracy of our forecast by an additional 5pp over the 20pp improvement over the original replenishment system.
How alternative data fuels digital transformation
Data-driven decision-making is vital for businesses that want to compete in today’s economy. That’s why a majority of companies use big data analytics to collect business intelligence. Few of these, however, have discovered how to leverage data effectively enough to succeed in company-wide digital transformation.
Why? Because they are still looking at the same data. True digital transformation is about more than integrating AI and machine learning into your current decision-making processes. It’s about rethinking your entire approach to the problem by leveraging new technologies. Business intelligence may be collected more efficiently when using new tools to analyse traditional data, but overall gains will be limited unless alternative data is also incorporated.
Including alternative data in your analysis allows you to consider new strategies that would not have been informed by traditional approaches. You can fill in the gaps in your analysis for a more granular, more real-time, and more accurate recommendation. Only this can deliver the expected dramatic improvements promised by digital transformation.
When digital transformation is done right, it’s like a caterpillar turning into a butterfly, but when done wrong, all you have is a really fast caterpillar.
George Westerman, Research Scientist with the MIT Sloan Initiative on the Digital Economy
Finding the right alternative data
As a data scientist, simply knowing that alternative data in general helps improve analysis isn’t very useful. You have to understand which data will help you achieve your business goals and deliver useful business intelligence. So much data is available, yet most of that data is just noise — and ultimately useless.
Every day, three times per second, we produce the equivalent of the amount of data that the Library of Congress has in its entire print collection, right? But most of it is like cat videos on YouTube or 13-year-olds exchanging text messages about the next Twilight movie.
Nate Silver, Statistician and Founder and Editor-in-Chief of FiveThirtyEight
Some of filtering out the right alternative data is simply trial and error. You choose a source of data that is likely to apply to the analysis you are making, assess the risks of choosing that data, and make the best guess. The results will allow you hone in on the most appropriate choice after running some tests.
Ultimately, however, you choose the right data by asking the right questions. When you always prioritize the true business goal and not just KPIs, you can work back to what data you need to fill in the gaps much more easily. I tend to use a common-sense approach. It is all about context. When I’m working with clients, I make sure to find out their motivation. Why are they looking to make their supply chain more efficient? Those answers will help direct me to a reasonable source of alternative data.
Maximizing returns from digital transformation
Big data has flipped the challenge for data scientists. We no longer struggle to find enough sources of data; we struggle to find coherent and useful patterns in the massive amounts of data available. Considering alternative data only amplifies that problem.
Yet we must meet the challenge. The best business intelligence today often comes from non-traditional sources, so it is a data scientist’s responsibility to pinpoint and analyse that data. Alternative data is the future of digital transformation, and failing to embrace it as a critical part of your analysis will lead to your company falling behind. If you want to maximize your ROI, you must first invest in the right information. That is now alternative data.
Big thanks to Kaitlin Goodrich.
About the author
Elena Marocco joined Evo as data scientist in 2016 after a very successful internship experience. A cum-laude graduate in Mathematics at the University of Turin, she defended an MSc with an innovative solution for Fashion Inventory Management.
She is excited about the world of probability, statistics and, more generally, in discovering useful maths that can have a significant impact through real life applications.