Fintechs that measure software efficiency may threaten tech jobs at large banks as management grapples with the shortcomings of legacy software, according to Jason Maude, chief technology advocate at Starling Bank.
“I think junior engineers will be very worried and probably rightly so because they will now not only have to deliver the scope in the time allotted which was miss-estimated up front, but they’ll also have to deliver it to a certain quality standard … I would be concerned for their engineers’ welfare around the end of the project development phase,” says Maude.
On January 22, efinancialcareers reported that JPMorgan’s Corporate and Investment Bank (CIB) was undergoing a trial with BlueOptima, a company that introduces metrics to manage software resources and the productivity of individual developers. JPMorgan did not comment on the trial before publication of this piece.
BlueOptima works with nine of the top 16 universal banks as well as companies within healthtech and telecommunications, according to its website. The company leverages Actual Coding Effort (ACE), Analysis of Relative Thresholds (ART), the Cambridge Performance Benchmark, and predictive assessment tools. It claims to provide a “first-of-its-kind strategic benchmark to monitor software development productivity, quality and cost to understand global positioning.”
Banks aiming to measure efficiency using software may struggle, according to Maude.
“The problem that many large banks are facing is that they have been developing software and trying to digitise for some time. What that means is they’ve had a lot of time to develop large amounts of software and what they’ve found is they’ve accumulated a lot of legacy software, which is essentially software that has been developed and is now very difficult or in some cases almost impossible to maintain. They don’t have the ability to change it anymore.
“They are increasingly worried about the costs of that technology and how much it’s costing them to maintain their systems. This is basically an attempt by them to make sure that the code that gets written for them is maintainable so that next time they try and change it, they won’t have to spend a lot.”
Measuring programmer efficiency is far from a straightforward process, according to a senior engineer at a peer-to-peer lending company.
“A legacy codebase is defined as legacy because it’s unmaintainable. You generally can’t just tinker with a legacy codebase until it becomes a good, maintainable codebase. Often it’ll have to be completely re-architected from the ground up, so even if the metric is accurate, it’s not going to be measuring productivity,” he says.
“And the main problem is that by measuring developers in this way you’re incentivising them to do busy work and nibble around the edges of a problem without actually solving it. We’re meant to be problem solvers for a living and the metric doesn’t even try and measure how effective a programmer is at problem solving; it measures how active a programmer is at adding and removing lines from a codebase as long as you don’t make the codebase less maintainable, which is really easy to do. If they’re making firing decisions or bonus decisions based on this metric, then programmers are going to game it, and any programmer would love to game this metric.”
According to Maude, the deployment of software efficiency solutions indicates an attempt to solve a tech problem which is more accurately a cultural problem.
“I think [big banks] are going to have to go through a long, painful process of gradually changing the system if they want to stay efficient and relevant.”
Despite moves by banks to optimise their tech stack, larger market participants are spending more on tech than ever before. According to a 2019 report by UBS Labs, JPMorgan increased its tech spend by 5.6 percent between 2018 and 2019, projecting an $11.4bn tech budget. Next was Bank of America at $10bn, Wells Fargo at $9bn and Citi at $8bn.
This increase in spending is occurring alongside significant job cuts, with many affecting tech roles. In December 2019, CNN reported that Morgan Stanley had cut 1,500 employees, with technology and operations roles being hit the hardest.
“I think that big banks have realised that they’re spending a huge amount on technology and they don’t want to. At the same time there is a massive trend towards them wanting to digitally transform and invest in all sorts of technologies like blockchain and AI, so it’s a somewhat bizarre congruence of things,” says Maude.
Performance testing software has gained popularity across several markets. Companies such as Neotys, Eggplant, and Load Impact provide solutions for testing software efficiency, but many industries rely on peer assessment for software and individual developers.
“Essentially, if all of the companies that were buying this [solution] were creating software in a much more agile way, they wouldn’t need this because they would be developing much less legacy code. So that sort of low maintainability would go away,” says Maude.