Advisor Perspectives welcomes guest contributions. The views presented here do not necessarily represent those of Advisor Perspectives.
We all have blind spots; that’s why cars were made with side and rear-view mirrors. We’re humans and can only see and do so much. When it comes to data, I have noticed firms still have noteworthy unseen areas that could cause operational risks. However, once those firms are made aware of their issues, there are some great things they can do to course correct.
To determine if your firm has gone off-roading unintentionally, here are five steps to ensure your systems are operating on the best and cleanest data set possible:
1. Minimize how often people interact with the data
We all know that human involvement in data inevitably leads to errors; this is unavoidable. Whether a bad rule was created or bad logic was used to decide how information should flow between systems, you can only focus on what is within your control. Minimizing the number of times people interact with your data simplifies and eliminates additional errors and setbacks. It will leave you with less anxiety that someone made an innocent but costly mistake.
2. Be proactive, not reactive
Be proactive versus reactive in finding issues with your data. Certainly, when there is a problem, you should respond to it. But that’s the bare minimum of what ought to be done. If you aren’t constantly checking to ensure the highest level of data quality and that everything is functioning as it should, you are leaving yourself and your firm open to risks. Firms are deploying artificial intelligence via pattern recognition and machine learning so that once an error has occurred and been assessed, if a similar one happens it can be resolved automatically. There are various controls and dashboards that overnight teams review and address daily. This resolves any potential issues ahead of time before it makes its way to accounting and reporting.
3. Avoid file transfers
It’s hard to believe that in 2022 we’re still talking about needing data in standard formats that are accessible in digital filing systems, as opposed to being written down somewhere or on a PDF, but here we are. In addition to having files in digital formats, they also need to be as up to date as possible. Avoid doing file transfers of data if possible. APIs are much better because they allow you to get the most current data, and there is less of a chance of something going wrong like a file freezing.
4. Establishing a golden source of data
To get what I call the “golden source of data,” you must get this information from at least two sources, with an option to override the data and input manually when needed. Your data should be tested and validated through a checks and balance system. And the best kind of data should have a track record that can be proven over several years.
5. Do your calculations once and correctly
One best practice I have found particularly helpful is that rather than take a data point and calculate it multiple times for each report/dashboard, use a precalculated number that is stored in a data warehouse. In this case, once it’s updated in one place, that number is updated in every instance it is used. Not only does that ensure accuracy but it also makes maintenance a lot easier, and the response time is much quicker.
Bad data leads to bad decisions by investment teams. Timely and accurate data is of utmost importance to all firms. The biggest operational risk is inserting human interventions into your data. No matter what approach you take to safety proofing your firm’s data, minimize human interactions with the data wherever possible. By finding the most efficient and secure way to manage data as well as being proactive with your controls, you can be assured of data accuracy that allows your operational processes to be scalable and cost effective.
Saneesh Surendran is senior vice president and head of systems at STP Investment Services.
Read more articles by Saneesh Surendran
