Trust issues: The data is error-prone, and I have been burnt before.

 

Trust is hard to build but easily broken. In our recent eBook about the challenges facing investors using Alternative Data we described a situation where a tagging issue led to millions of dollars in realized losses for hedge funds that opened short positions in a certain retailer. The trade was based on an indication of tanking credit card transactions that turned out to be a mistake. A very costly one.

 

The knee-jerk reaction was to blame data. Indeed, some managers, frustrated with the loss, decided to abandon Alternative Data altogether.

 

Blaming data for the unfortunate trade is akin to blaming construction bricks for a wall that collapsed due to poor masonry. Should we never use bricks again? Clearly, it’s not the bricks themselves, but the process that failed to identify the cracked ones. And a process can be improved.

 

Not all managers that made the wrong-footed trade abandoned data, and some turned the pain into a stimulus to do a much-needed retrospective. Examining the “data bricks” led managers to identify and understand the issue and consequently get a deeper appreciation for the process of tagging data correctly. Tagging is just one crucial step of several in the flow from raw data to insight. The manager or their analysts must pay individual attention to each step in this flow to better understand how and where issues may arise.

 

As with most things, a deeper understanding helps build trust and confidence.

 

This is not to say that Alternative Data is perfect and completely free of errors. There are a vast number of things to consider when using this data for investment purposes, and most managers operate with the old maxim of “Trust but Verify”.  But just because you have been hurt before, it’s not a reason not to try again, albeit with a more sophisticated approach. The many benefits of integrating Alternative Data into your investment process are more than worth it.

 

Time to Value: It will be months if not years until I get any value from this data.

 

Painkillers always sell better than vitamins. Some managers look at Alternative Data as a requisite “CYA” investment and a cost center rather than a profit center that improves every part of their investment process. They believe that because they feel it can’t help them right this second.

 

Time to Value describes how long it takes for your investment to start paying back, and the myth among managers is that the time to value with Alternative Data is necessarily months or even years. It’s important to acknowledge that this belief is grounded in truth – that is, if you go about onboarding Alternative Data in a traditional way.

 

Think about the traditional process of data discovery and onboarding. To find out what data is available and applicable to their portfolio, managers must reach out to multiple vendors, and request coverage lists and (where available) backtests for their names. After making some assumptions and decisions (best guesses) on which datasets to onboard, they must contend with different formats, data lags, frequencies, and delivery mechanisms. They then must build an ETL process, with error detection, mapping, tagging, and finally, staging the data for analysis. Finally, they must build models and validation to relate the data to their companies’ reported KPIs, or “ground truth.”

 

All of this can take vast amounts of resources and precious time, and that is just for one data vendor. Now, imagine you need to evaluate and onboard five. Needless to say, managers have good reasons to believe the time to value for Alternative Data is a concern.

 

Thankfully, technology can help with this problem, and today, managers can take advantage of partnering up with a company that has already done all the heavy lifting. Working with a technology provider that has already onboarded hundreds of datasets and established processes for normalization, anomaly detection, mapping, and staging, can reduce the time to value from years to days. Access can become a flip of a permissions switch, and you can go from zero to generating reports and receiving data-driven alerts in just a few days.

 

It’s a good time to revisit your beliefs on time to value for Alternative Data because even long-term investors can enjoy some instant gratification.

I am good enough: My current process works fine, and I do not need to change it.

 

Managers need to continuously convince investors that they are the best stewards of their capital. It’s no surprise that at some point, they start to believe it themselves.

 

Even as most fund managers embrace Alternative Data, there are still holdouts, convinced that they do not need it, because their process is “already great”.  They may also shy away from data for any of the preconceived notions we have mentioned thus far. Even though it’s tough to look critically in the mirror, no matter how great your process is, there is always room for improvement. Let’s admit it, the mechanics of fundamental research aren’t always ideal.

 

Think about your typical research of a company. Sitting in a boardroom with the CEO or COO, have you ever wondered if management is 100% truthful and transparent with investors? Data can help provide a different, unbiased, and independent perspective. When calling one of your peers and chatting about their latest idea, couldn’t you swear you heard that same pitch at a conference you were just attending? A screening process that relies on Alternative Data can yield unique ideas that do not come from the usual circuit of hedge fund conferences and offer critical differentiation from the crowd. Managers love to underscore that their process is “data-driven,” but can one really claim that when a huge portion of data available on their securities (by volume) is ignored and left out of the process?

 

Download eBook on the challenges to managers using alt data

 

What do you think about Alternative Data? Talk to us about your current process and see how Alternative Data can fit in, or how to improve the value of your current data investment.