The Road Ahead: Predictions for 2019
1. Value stream management will be the key to success in DevOps.
DevOps is now universally recognised as the route to delivering better software, faster. But how often have you heard that cultural change is the major obstacle to making that shift left?
Most people would agree on the key measures of success in software delivery: the number of defects and the speed of delivery, for example. But few firms have built a framework for value stream management that can apply the same metrics consistently across the enterprise, to teams working in the siloed departments of engineering and operations.
A number of software technology vendors offer tools labeled value stream management that visualise the software development chain and capture key indicators of success. But value stream management done properly comprises more than that.
As Nick Jones, head of test engineering at Global Payments, the Atlanta-based payments platform, explained in a recent webinar, produced by QA Financial, value stream management helps deliver operational efficiencies by identifying - on a level playing field - which teams are doing well, and which aren’t. This, in turn, improves collaboration, communication, and co-operation. You can view the webinar on demand here.
2. AI accelerates automation, challenges regulators
Arthur C Clarke, the scientist and writer, is credited with the adage: "Any sufficiently advanced technology is indistinguishable from magic."*
Yes, it's been the buzzword for 2018, but 2019 will be the year AI moves from the realm of magic to the real world and goes enterprise-wide at larger financial firms. Right now, the two crucibles of AI engineering financial software are in mobile app development, and in high-frequency algorithmic trading.
"The mobile app developers are ahead of everyone else right now in terms of assessing the impact AI can have on testing - we want to know what those mobile developers are doing," the head of QA at one leading exchange tells us.
AI is being used, for example, to test mobile apps by emulating the way the human eye behaves when it sees a screen. That’s not necessarily interpreting the world the way a human being does; it’s simply processing a large amount of visual data, faster. In the past, the meaning of AI has been variously defined - clouded even - by the frequent linking of "cognition" to "artificial intelligence”.
We believe the definitions of AI in 2019 will become simpler and more useful. Perhaps: AI improves automation.
However, that still leaves the thorny issue of what view regulators will take on the growing use of AI. They will be focusing on how developers are managing the risk of bias, and how the outputs of AI can be tested to ensure that they match the objectives. For example, how to ensure that what a chatbot tells a customer is what it should, legally and objectively, be telling the customer.
*Our thanks to Sundeep Kumar of JP Morgan for reminding us of this law, during his presentation at the QA Financial Forum New York last November.
3. Risk management is in the driving seat for transition management.
A recent survey by EY on banking risk and technology change found that a key concern of banks is creating “regulatory and supervisory certainty around the deployment of new technologies". As financial firms become more technologically complex, the risk and IT functions are getting pushed closer together and risk managers are increasingly involved in IT decisions.
Basel III and, in particular, the requirements of the Fundamental Review of the Trading book, will be pushing them even closer together over 2019, and beyond. There are many other compliance requirements for financial firms – SOX, IFRS and KYC – that have to be embedded in software development. But the FRTB is proving a catalyst for change at some leading banks because it requires that they capture market risk in real time and render that data accurately to the firm’s risk engine. In short, replacing risk models with data. This necessitates a fresh approach to connecting software across the enterprise.
Despite all the changes in banking culture since the 2009 financial crash, there are still barriers between the business desks and the risk desks. But the scale of this new challenge is prompting a number of firms to change the way they develop and test their trading software, pushing them towards agile ways of working.
A further impetus comes from European open banking regulation, which comes into force in 2019. By next September, all banks will have to comply with the European Union's PSD2 rules, designed to promote competition in payments. The testing of the APIs that will link different payers and payees is already established as a key driver of demand for testing and development services in retail banking.
The challenge is the same for all types of financial firms: the growing complexity of banking technology. The debacle at TSB in April 2018, when the bank’s mobile and online systems collapsed during a transition to the IT platform of its owner, Sabadell, showed just how much can go wrong when managers fail to grasp the requirements of new software technologies, such as micro-services. This report by IBM, produced for TSB’s management a few days into the crisis, explains the challenge, and highlights the importance of test planning.
4. More M&A among specialist vendors, with data driving deals
The Office of the Comptroller of the Currency, the US agency that supervises the federal banking system, has flagged in its latest semi-annual report its concern about "concentration risk" as more business flows to fewer IT vendors.
But we can’t see this – or softening equity markets - stopping the wave of M&A activity among IT vendors; certainly not those that specialise in automation technologies for software testing and quality assurance.
2018 saw SmartBear’s acquisition of Zephyr and Perforce’s acquisition of Perfecto. SQS, the German-based, testing services provider, was acquired by Assystem Technologies. Tricentis and QASymphony merge their brands (into Tricentis) in January 2019 (though they were both already owned by the same parent, private equity firm Insight Venture Partners).
The larger-scale IT industry deals included IBM's acquisition of Red Hat, the cloud and open-source integration specialist, for $34 billion and chip-maker Broadcom’s acquisition of CA Technologies, at a cost of $19 billion.
While a deal that will compete in size with IBM’s controversial swoop on Red Hat is unlikely, we can confidently predict that just as many specialist tool vendors will be bought or merged in 2019. That’s in the nature of their (mostly) VC and private equity ownership, of course. There has to be a level of deal churn.
But what also drives activity is, firstly, the need for scale. Combining tools with management platforms helps marketing.
The second key driver will be the well-founded interest in software automation technologies and, in particular, AI. It’s a land grab.
So there will be deals. Our prediction for 2019, though, is that we will see deals that combine test automation and management businesses, especially those data businesses that focus on virtualisation technologies. Even the largest financial firms say that their biggest QA headache is making decisions on the right investment for test environments.
5. Consultancy is King, and smaller customers are needed.
The leading systems integrators (SIs) had mixed results in 2018, but one consistent theme was… that none of them want to be called a systems integrator anymore.
Their revenue growth has come – and will continue to come - on the back of demand for cloud services, IoT and data management services, and automation transition management.
Fairly or unfairly, however, SI remain lumbered with the association of armies of offshore manual testers. And while the leading systems integrators are seeing good revenue growth overall, growth in revenues from the financial services sector is lagging behind others.
Tata Consulting Services' quarterly results announcement in September showed year-on-year revenue growth of 11.5% with 6.1% growth in revenues from the financial services segment of its customer base.
For Cognizant, financial clients are clearly problematic: while it's overall revenues were up 8.3% year-on-year, financial services revenue growth was only 4.4%, well behind sectors such as healthcare and media.
Infosys bucked the trend, reporting 4.2% overall revenue growth with financial services growth out-performing the average at 5.8%.
In April 2018, Infosys announced it was putting Panaya, its Israel-based specialist test automation platform business, up for sale.
While its acquisition of Panaya had in any case been problematic, Infosys’s decision to sell it was also billed as a move to position itself more as a pure-play consulting business, unencumbered by the need to push proprietary solutions. In particular, financial services firms still need guidance when it comes to decisions about transitioning to the cloud.
In any case, the marketplace for software services among the world's largest banks and insurance companies is saturated. Big contracts are increasingly dependent on customers making a decision to drop one key provider for another.
For the leading SIs and the specialist quality assurance vendors alike, there is more growth to be had from winning business from smaller clients in sectors that have been behind the curve of digital development: credit unions, private, alternative lenders, and smaller asset management businesses, for example.
6. Somebody will be blown up by Brexit
With only three months to go before the deadline for Brexit, it would be a cop-out not to make a prediction on what it might mean for IT. All the banks and trading venues we’ve spoken to say they are ready, whatever the outcome of the UK’s deal/no-deal see-saw.
In essence, while there may well be havoc in financial markets, the central software testing challenges posed by Brexit are straightforward and have to do with load testing and network testing, they say. If there is no deal, there could be a huge exodus of German and French investor money from London. The derivatives markets will be plunged into legal uncertainty and the pound sterling could fall through the floor. More and more business will be done in Paris, Frankfurt and Luxembourg. But this is all predictable. Firms’ systems simply have to be ready.
But while nobody is actually saying they are unprepared for these strains, the UK government and the British banking regulators have been flagging their concerns about the resilience of the financial sector.
The UK Treasury has launched an official inquiry into why IT failures such as TSB’s (see above) happen, and what the consequences might be. The Bank of England has published a discussion paper on IT resilience. All very prudent. But we also know that UK banks carry an additional burden of older and more complex IT infrastructures than their European and US peers.
This study by Cast Software quantifies that problem. For example, software at UK banks contains on average more than one million lines of code, compared to around 440,000 lines for European and US banks. British banking software is more likely to break, and we believe it’s a sure bet that someone (we hope not someone too big) will be broken by the strains of Brexit.
Of course, Theresa May could secure a brilliant deal for Brexit any day. The pressure will be off. The nightmare could go away. But it’s not looking likely, is it?
Happy New Year!