Parallels between investing and software development

At first glance, software development and investing seem to be poles apart. But we find a common and interesting thread in the process that investment managers and software developers follow.

The first public release of the Android operating system (OS) for mobile phones was in 2008. The most recent version is its sixteenth, meaning an upgrade to the OS has averaged three times every two years! The iPhone, currently in its thirteenth year, has seen major upgrades to its iOS every single year since launch. In fact, most software products now undergo frequent updates and upgrades. Contrast this with the nostalgic era of desktops, where the Windows OS would see a new version every 5 years or so. And where upgrading the OS was a ritual, requiring the help of an IT support specialist. Now the OS and the apps on your phone get updated while you sleep!

Frequent updates, or continuous iteration and improvement, is an essential feature of modern software development. Testing the iterations, getting customer response and incorporating the feedback in the next round of iterations is a process increasingly followed by software developers, who call this the ‘Agile’ methodology. The pervasiveness of Agile can be gauged from the fact that in the 250-odd page 2019 annual report of one of India’s leading IT services companies’, the word finds a mention over 100 times. Product developers, especially in tech start-ups, have their own derivative concept termed Minimum Viable Product (MVP) – an early version of the product that lets you get the best customer validation with the lowest effort/cost. You start small, and once sure of the product fit to the customer’s wants, commit more resources (capital, human) to keep building it further and better. Swiggy, for example, started in just 1 neighbourhood in Bangalore, with 25 restaurant partners and 6 delivery executives. Once the model was proven, it’s taken them just 5 years (and piles of cash!) to get to where they are today.

We routinely see the agile method at work in investment management too. The sensible (and that is why the obvious and usual) approach to investing is to undertake thorough research on a stock/business and convince yourself if it is worth committing money to. How much time to spend on the research is a function of one’s own investing style and more crucially, the fund’s mandate. The manager of a close-ended fund with an initial lock-in can, and probably will, spend weeks on a company before buying the stock. However, a relative return fund manager, expected to be a top quartile performer every quarter, will not have the same luxury. In either case, it is a practical difficulty to get a grip on all moving parts of a business, especially as an outsider.

What the ‘agile’ fund manager then does, is evaluate whatever information she has on hand, assess its ‘minimum viability’ and price the risk of the unavailable or unknown bits, before making a move. The ‘MVP’ is then launched by committing a small proportion of the fund to the new stock. This is followed by ongoing research, the results of which determine the strength of the manager’s conviction in the business. This conviction is then expressed in the position sizing of the stock in the overall portfolio – commit a larger proportion as the conviction rises. If one observes carefully, most successful active managers follow ‘agile’ investing in some form. In fact, when you are building something complex, where more variables are at play than what the mind can fathom, it makes sense to build it gradually. This way, the risk of making catastrophic mistakes is mitigated.

What is critical in this process is the ongoing research, and needless to say, it is critical irrespective of the time spent in the initial analysis. A small holding hardly adds value to the overall portfolio and therefore, conviction building assumes significant importance if a stock were to have a meaningful contribution to portfolio returns. In one of our portfolios at Marcellus, even after having spent a considerable amount of time on initial research, we’ve increased the weight on a stock by as much as 30%, after results of subsequent efforts bolstered our conviction in the quality of the business and its deepening moats.

One pitfall to avoid while undertaking ongoing research is the dilution of focus, from tracking structural changes in the competitive position of the business, to forecasting and monitoring earnings. Although earnings are reflective of competitive advantages, the frequency of result announcements subconsciously makes us miss the woods for the trees. How much time should one spend in tracking raw material prices to forecast the quarterly profitability of a business as accurately as possible? Instead, all that time could be better spent in identifying the factors that can improve the business’ pricing power, making the cost of raw materials nearly redundant – the company passes on cost increases to customers and retains decreases for itself.

The biggest mental hurdle to ongoing research is the stock price performance. A rising stock price immediately after building an ‘MVP’ position could be purely a stroke of luck. But this can give a false sense of conviction. In turn, this leads to complacency in continuing to dig deeper into the business. And committing more funds to the stock without adequate research could set the portfolio up for trouble in the future. At the same time, inadequate research will not convince one of greater value in a stock even after the recent price movement.

Where the gradualist approach to investing diverges from agile software development is that in the former you are dealing with a ‘noisy’ system whereas the latter is a predictable system. In noisy systems, it is harder for the system builder to know whether she’s made a serious error or whether noise has polluted the system. This is what arguably makes investing much harder and more interesting than software development.

How then, can an investor deal with noise? One way is to distance herself from it – TVs in the research offices, Bloomberg terminals on desks, WhatsApp messages, and other tools are constantly bombarding the fund manager with information, much of which is really noise. Second is to be patient. When it comes to making investing decisions, the more patient an investor is, the less likely it is that she is responding to noise.

So, next time you meet your software engineer friend, do explore the commonalities in the way you work.