Alt Data Ethics By Simon Troup

In 2020 the SEC (the US Securities and Exchange Commission) announced its examination priorities for the year ahead. It might not have registered with a lot of investors, but there was a sneaky mention of ‘Alternative Data’. They recognised that “advancements in financial technologies, methods of capital formation and market structures, as well as registered firms’ use of new sources of data (often referred to as “alternative data”), warrant ongoing attention and review”

https://www.sec.gov/news/press-release/2020-4

“Financial Technology (Fintech) and Innovation, Including Digital Assets and Electronic Investment Advice – OCIE recognizes that advancements in financial technologies, methods of capital formation and market structures, as well as registered firms’ use of new sources of data (often referred to as “alternative data”), warrant ongoing attention and review. OCIE also will continue to identify and examine SEC-registered firms engaged in the digital asset space, as well as RIAs that provide services to clients through automated investment tools and platforms, often referred to as “robo-advisers.””

But what are these new sources of data, and what are they an alternative to? Alt-data is increasingly used in many enterprises from investing to consulting, even by some of the largest companies we use every day like UBER.

Alternative Data

Traditional Data

What is alternative data? Firstly, at least from an investor’s perspective, what is traditional data (trad-data vs alt-data). Investors traditionally have focused on information made publicly available to the markets through accounting reports (balance sheet and cash flow statements), meetings with management (often facilitated by investment banks and brokers) and trading statements that describe the current environment, whilst not revealing accounting specifics. Depending on whether you are value or momentum investor you would use accounts to ‘value’ the company or use the stock price trend plus ‘earnings’ momentum to pick stocks. For decades the data driving investing clustered in a handful of market data firms, think Bloomberg, Refinativ (formerly Thompson-Reuters), Factset, Standard and Poors. These firms captured accounts through fundamental data products (worldscope, compustat etc) or consensus estimates (IBES, Bloomberg and Factset estimates). This was more or less what an equity investor needs, a forward view tracking consensus and a backward view looking at realised financial performance. Fundamentally equity investors were looking at price multiples like book-to-price, earnings-yield, dividend-yield, dividend-cover to value a company, accounting for growth in profit forecasts captured in consensus estimates. How did alternative data disrupt this established status-quo.

Alternative Data

Simply put, alternative data enables investors to observe our economy in near time; in other words, investors can capture the source of company sales rather than the accounting effect (the profits). Rather than having two or three data sets there are now thousands with some companies being super specific, perhaps even offering an advantage in only a few listed companies. Some impressive startups like Neudata are now advising investors on the latest available data sets that could offer a trading edge. Alternative data is now much less alternative. Before diving into examples of where alternative data has caught the regulators attention, lets think about some examples of alternative data.

Data-As-An-Exhaust

Data exhausts are a common source of Alternative data and have been the centre of the case studies that follow. The key to an exhaust data product is that the data is a derivative of a core business, or perhaps a customer workflow. Let’s get into some high level examples.

Card Payments

A great example of a data exhaust that has been around for many years is the data exhaust from our payments infrastructure. All the major credit card payments companies (think mastercard, American express, visa etc) capture vast amounts of point of payment data. They have insights into the sectors where consumer spend is directed, the companies, georgaphies, perhaps even customer demographics. These data points have been used for years by large brands to track spend, but increasingly by investors to help position portfolios.

Data exhaust products are often derived from transactional data, but more on that later, but the key is that somewhere in your credit card terms and conditions you are signing off the use of your data for insights products. These products are data exhausts because they generate revenue but not from the core business (payments in the case of mastercard for example). But this revenue does cross-subsidise our low cost payments infrastructure; remember if the product you use is free or at a below cost rate, it likely means YOU are the product being sold.

When you think about it, there are so many opportunities to capture you data in an exhaust insights product, your cookies from your web browser, your geo-location data from your phone, your emails. (Yes, some products you grant access to email may be reading e-receipts, so read terms carefully for free email ‘tidying’ or work flow apps).

Open Banking

Open banking has unlocked a huge amount of value by unlocking payment capabilities to startups (think settleup, currensea). Its also the entry point for a range of subscription management apps to ensure we get the most from the online products we use https://www.cnbc.com/select/best-subscription-trackers/ But buyer beware! These apps may track subscriptions by introspecting your direct-debits in your current account, assuming you have granted access. Indeed most money budgeting apps need access to your current account to track spend an make recommendations. https://www.moneyboxapp.com/

has thrived with this strategy; one of their sources of revenue is the ability to track direct debits across customers and provide data insights in a similar way to our the payment provider use case. However, your current account is even closer to home and has different information. Portfolio Managers have used data insights to track direct debits for specific companies after info-sec (information security) events. Are customers cancelling services from a provider with leaked user names and personal data? Needless to say this is very valuable information when forecasting quarterly revenues for companies, and positioning a portfolio into earnings announcements.

Data Ethics

That all sounds a bit scary, but hopefully most readers are aware their data is often used as an exhaust product, in particular for apps that are free or super cheap. Rest assured your gmail account is free due to the data advantage it affords Google. So where could this all go wrong? There are two well-known examples of alternative data startups being caught either on the wrong side of the consumer directly, or due to the interest of financial regulators.

Unroll.Me

What companies were involved –
- Me >> “Unroll.Me gives you the tools you need to manage a cluttered mailbox full of pesky subscriptions emails.”
- UBER – data insights buyer
- LYFT – data insights target
What regulators were involved
- FTC – The US Federal Trade Commission

The Case – https://www.ftc.gov/legal-library/browse/cases-proceedings/172-3139-unrollme-inc-matter

The Story –
- https://www.nytimes.com/2017/04/24/technology/personal-data-firm-slice-unroll-me-backlash-uber.html

Where did it go wrong?
- me is clear isn’t special at all, it was clear in its t&c’s how it worked, it seems viral internets stories caused their challenges, coupled with tribal loyalty in San Francisco for Lyft vs Uber
- But it seems EU GDPR is too strict for them to operate comfortably within……….“The EU is implementing new data privacy rules, known as General Data Protection Regulation (GDPR)…….as a result, we will temporarily stop providing our service to EU customers, and we will stop providing service to all EU residents on May 23.”
Take Aways
- Beware of the Data-as-an-Exhaust – If a service is free you are likely the product. Don’t be surprised if you find your data being used in unexpected ways.
- Don’t underestimate how data could be used – Data is super valuable, rest assured that your data is probably being used in ways you couldn’t possibly imagine!
- Customer loyalty – Customer loyalty is real. In this instance its wasn’t so much that the data was being used, it was that fiercely loyal Lyft customers felt violated that UBER was learn of their habits and preferences. The impact on Unroll.me brand was material despite most tech folks being thoroughly unsurprised.

BIG Tech Sharp Practice – Needless to say UBER are super innovative and have a history or sharp practice, https://www.theguardian.com/news/2022/jul/12/uber-paid-academics-six-figure-sums-for-research-to-feed-to-the-media ! Its great to work in BIG tech, but apply with eyes wide open to how they operate and grow.

App Annie

What companies were involved –
- App Annie (now data.ai) – “data.ai’s mobile market estimates and analytics platform delivers actionable insights to analysts, helping them drive client trading activity.
What regulators were involved
- SEC – The US Securities and Exchange Commission

The Case –https://www.sec.gov/news/press-release/2021-176

The Story –
- https://www.neudata.co/alternative-data-news/sec-breaks-new-ground-in-settlement-with-alternative-data-provider-app-annie

Where did it go wrong?
- App Annie Representations – Fundamentally App Annie represented that it anonymized and aggregated app data to protect privacy. Furthermore, they represented to customers of insights that app user customers had consented to the confidential data be used for data insights and had the policies and process in place to comply with SEC compliance rules.
- App Annies Insights Process – Contrary to their claims, App Annie used non-aggregated data to provide insights and most importantly did so in an identifiable way.
- The Action – App Annie were filed $10million USD and the CEO removed.
Take Aways
- Leadership Accountability – The CEO was removed, this was a serious breach of corporate governance with the CEO being held accountable.
- SEC Optics -The SEC have taken a hard line via their first Alternative data prosecution.
- Definition of an ‘insider’ – The traditional (general) definition of an insider is the possession of “material non-public information”. Material in that it could affect the price of a security, and non-public in that its unavailable to the wider market participants. Alt Data vendors may have material non-public information, and investors are increasingly incorporating strong due diligence processes to ensure they remain within the regulatory requirements.
- Consent – It is extremely important for the correct consents on data sold to insights customers are robustly enforced. Furthermore aggregations must protect the identity of individuals, in particular given through the use of multiple data sets and individual could be identified with a high confidence level.