Featured Image

Build Versus Buy: How to Evaluate Your Software

Data Science and AI

By FactSet Insight  |  February 28, 2023

For as long as the investment management industry has used software, there has been a debate about whether asset managers should build or buy their tools. Jonas Svallin, global head of quantitative research and product development at FactSet, argues that the decision is not one or the other—but rather a combination.

Since the exact approach is likely to evolve over time, the right strategy to better inform the build versus buy decision is to adhere to a philosophy that works for the organization through market cycles. This leaves a significant opportunity to adjust the employed strategy when technology significantly improves (such as cloud computing or when vendors introduce material enhancements). This article posits seven questions that can facilitate data-driven decisions about which approach is right for your business.   


1. Is your firm an investment manager or a software company?

Bringing a substantial part of the software development effort in-house can shift a business model from investment management and client service toward data and engineering. This can be distracting during periods of major investment platform development (which requires substantial investments in technology and people) or performance turbulence.

At many investment management companies, software development teams are not part of the front office (in effect, the investment team), which means they are not identified as core to the organization. This can lead to difficulty attracting and retaining high-caliber talent. At software companies, however, development teams lead the organizations. To attract the right talent, firms need to ensure that they are asked to work on enhancing and developing new solutions—“caretakers” seldom deliver the cutting-edge solutions needed to produce sustained alpha (positive risk-adjusted active or absolute return).  


2. Is the platform unique? 

It’s not difficult to think of reasons to build an in-house research and portfolio management platform. The most common argument is flexibility, so teams can produce unique research that leads to sustained alpha. However, building a state-of-the-art platform involves first building the 70% that is mostly generic. Tools that allow efficient data consumption, back-testing, portfolio simulation, optimal weighting and attribution are now developed by third-party providers with increased levels of sophistication. Therefore, the challenge is to find software providers whose solutions can be seamlessly integrated into the investment platform so investment teams can focus on the proprietary aspect of investing (the 30% portion associated with sustained alpha).


3. Can your company manage design mistakes? 

One of the most important inputs when building software is feedback—ideally continuous and in great quantities.  As a result, software companies have developed concepts such as sprints to quickly incorporate feedback. Third-party providers get thousands of daily enhancement requests, which are documented and incorporated via sprints. This leads to quick course corrections that ensure robust features at a steady state. Since this requires significant specialization (software, cloud and database engineering, and database product development) it’s important to have in-house know-how that can turn feedback into practical solutions.

In some ways, the software short-term mindset is the opposite of the long-term investment approach practiced by many investment managers. This highlights the potential challenge of being both a high-caliber investment management firm and a software company.


4. Can your company maintain the software it has created? 

When big software and data projects are launched, most organizations do a good job of budgeting the associated cost. However, once the main specifications are completed, the platform starts to depreciate, and companies must add meaningful resources for upkeep. In many ways, building the platform is often the less difficult part, while keeping it relevant and competitive is the challenging and costly one. It is therefore important to set aside a realistic budget to develop the platform and for annual maintenance and enhancement requests. When done accurately, it often highlights why building software is expensive and requires amortization over a large user base.  


5. Will the platform meet milestones or be abandoned? 

When sizeable internal platforms are developed, companies need to carefully consider that the technology employed will not last for as long as they believe or that the company’s strategy could change through a merger or an acquisition. If it’s supposed to be depreciated over three years, a two-year what-if scenario should be built in, for example. This means the cost to develop the internal platform could be dramatically higher than originally planned (two instead of three years). If a combination of build and buy is considered, that shifts uncertainty onto the software provider, which is ideally the generic portion of the investment platform.  


6. Are fixed/variable costs optimal?

The worldwide economy is incrementally driven by specialization as the level of sophistication keeps accelerating. Good examples of this can be found among cloud providers and data warehouses such as Amazon Web Services (AWS) and Snowflake (the market dominance among cloud providers is also a good indication). Many users, including investment management companies, now buy annual consumption from AWS and Snowflake in return for knowing they can worry less about data computation, data management or storage. The same argument could be made for many of the investment tools needed to run a sophisticated investment management company.

Another key argument to consider is the level of uncertainty in the world of investing. If a company launched a quantitative business in 2006 or 2017, they most likely incurred sizeable performance losses over the following years (“Quant Quake” and “Quant Winter” were some of the more challenging periods for quantitative managers in the last 15 years). These performance losses, combined with substantial upfront development expenses, can significantly challenge a viable investment strategy. If the cost involved in building a competitive investment platform is $5 million, for example, then a more appealing proposal would have been to buy all the basic services for $500,000 per year, avoiding material startup expenses and performance challenges.   


7. What new features are being worked on and what will they cost? 

During the software evaluation process, it’s critical to evaluate not only current features but also additional features being worked on and how the software provider plans to price them.

If a company can see a clear road map that contains enhancements as well as new functions that are part of the current expense, that becomes an important argument for buying. 

For example, when a consumer buys an iPhone, they know they will likely be able to continue using it over a period in which Apple may release three, four or five new operating system updates. That is a free option the consumer benefits from during the lifetime of the phone.  


Conclusion

If the seven questions in this article can be answered with confidence, then building a bespoke investment platform may be the right answer. This is especially true if a company can continue to resource its investment platform during times of adverse investment environments. One way to increase the likelihood of success is to make sure technology is a key pillar in the investment process and elevate the CTO to be a peer of the CIO.

As more of the economy gravitates towards specialization, one must be cognizant of what is cutting edge versus what is generic and how companies could potentially use a third-party platform to outsource generic aspects while retaining access to bespoke and sophisticated solutions.

 

This article was originally published on WatersTechnology.

This blog post is for informational purposes only. The information contained in this blog post is not legal, tax, or investment advice. FactSet does not endorse or recommend any investments and assumes no liability for any consequence relating directly or indirectly to any action or inaction taken based on the information contained in this article.

Subscribe to FactSet Insight

Comments

The information contained in this article is not investment advice. FactSet does not endorse or recommend any investments and assumes no liability for any consequence relating directly or indirectly to any action or inaction taken based on the information contained in this article.