Crisis? What crisis? No, I'm not talking about the 1975 Supertramp album, but research from a global leader in the test automation space demonstrating the “acute pressure” facing businesses to deliver apps in the Internet of Things (IoT) and digital era. It reveals that half of companies in the US and the UK admit to releasing apps before completing “quality testing.”
And, enterprises are being urged by the British CEO spearheading the test automation firm’s endeavours to rethink test automation to avoid the “app scrapheap.” But are matters getting to a crisis in the apps space?
In the survey commissioned by Testplant, a UK headquartered firm with a R&D presence in Boulder, Colorado, which provides what is touted as user-centric, digital automation intelligence solutions to enhance the quality and performance of the digital experience, canvassed 750 development team leaders in Britain and the United States to derive its findings.
The sample of canvassed ranged from telecoms companies, financial services groups, retailers and manufacturing firms - from listed to reasonably large companies - as well as government organizations. And the rationale for conducting the exercise was to validate what Testplant felt was the reality on ground and current thinking by companies.
The results come a month after Testplant, in which Carlyle Group’s European TMT-focussed fund invested €657 million (c.$696m) last year, launched its AI-powered Digital Automation Intelligence Suite, which brings to testing and the promise of “full automation and predictive analytics” to ensure their customers have a robust digital experiences.
Clearly, improving user experience (UX) and productivity are the keys to success in a digital world. But it would seem not all companies are up to speed given the survey’s results, which were conducted on its behalf by Kickstart.
Given that the manual testing market today is worth around $33 billion (bn) globally, if one could use Artificial Intelligence (AI) to augment humans to make them more productive the spoils on offer could be sizeable.
Roy Cornelius, Principal Architect (Digital Apps), IT Centres of Excellence, at BT Group, reflecting in the wake of the recently conducted survey said: “Test Automation is now a critical business requirement in the world of Digital and Mobile Apps. Rapid time-to-market and delighting customers are imperatives to business success.”
He added: “Expanding automation to assist with not just test automation, but also test case creation and test results analysis is a vision I share. And this must be done ensuring an optimal digital experience on any device or network.”
John Bates, CEO of Testplant, who urged enterprises to step up to the plate and commenting said: “To win in the Digital World and avoid the ‘app scrapheap’, enterprises must go beyond the current approaches to test automation. Testing must transition to become a profit center - testing not just code quality and app performance, but delivering predictive business metrics like satisfaction and retention.”
Cambridge University educated Bates, who has been recognized as a fintech and Big Data visionary and will move between TestPlant’s London HQ and their U.S. office in Boulder, added: “Only companies that adopt intelligent, end-to-end test automation approaches for software and apps will be able to keep pace with customer demands and the DevOps teams that support them.”
Productivity Gaps
Given that almost 1 in 4 people who download an app only use it once, and 51% of users do not download any apps in a month, things might well need to change and on testing front. However, at the same time, 86% of test teams say they are meeting their test objectives.
This it turns out is because organizations have “made testing more about code compliance checking rather than about delighting users” according to Testplant. The upshot is low user adoption, engagement and revenue. And, in turn this creates a gap in user experience.
Concurrently with all this happening, DevOps, mobile and IoT, digital, and consumerization are hugely increasing the scope of testing, but shrinking time to delivery.
Teams are in consequence unable to keep up with the pace of DevOps, or deliver what the business wants - even with the budget they have requested. This creates productivity gaps in time to market and efficiency.
App Market & Economy
Last October, market intelligence firm Newzoo released its first Global Mobile Market Report, which included global app revenues and forecasts, device brand market share and smartphone penetration, which estimated that global app revenues will reach $44.8bn in 2016 and grow to $80.6bn by 2020. But the figure rather depends on which markets and segments are included.
More recently in June 2017, a report from app analytics firm App Annie projected that the global app economy would be worth $6.3 trillion (trn) by 2021 - against $1.3trn last year. And, over this time horizon the user base using apps will see an almost doubling to 6.3bn from 3.4bn people.
Mobile commerce was cited in this report as being the single largest driver of growth for the app economy, increasing from $344 per user to $946 by 2021. Asia will witness the quickest growth, reaching $3.2 trillion (trn) in 2021, followed by the Americas hitting $1.7trn and the EMEA region at $1.0trn.
Testplant’s survey titled ‘Application Crisis Research’, which was conducted amongst development team leaders on both sides of The Pond, revealed that 68% plan to build more apps during the next twelve Months. At the same time as reporting increased volumes of development, 91% of developers agree that user expectations for innovation and quality have also increased.
The ‘App Complexity Crisis’
The research, confirming that organizations are not only planning to build more apps in the next year, noted that they will be more complex. Over the next 12-18 months, 50% of the sample said their companies will develop more apps with IoT components, more than half (58%) indicated that their companies will develop more apps with Artificial Intelligence (AI) components, and 62% of enterprises will increase the amount of apps deploying machine learning/deep learning components.
In addition, 42% believed that they are expected to design, develop and test apps in an “unrealistic amount of time”, with over a third (36%) admitting they are not given enough time to ensure apps are properly tested before deployment.
Just over four fifths (81%) expressed the view that with more time, apps deployed by their team would have a greater impact on the business.
Over half of the respondents (56%) agreed that use of outdated techniques and tools is “holding them back” from meeting the demands of the digital world. And, 75% agreed that with better tools, apps deployed by their team would also have a greater impact on the business.
Pressure Cooker & Crunch Point
Of the survey respondents, 60% said the majority of pressure comes from within the company, whereas nearly half (49%) indicated pressure comes from competitors.
Two-thirds (66%) of the companies canvassed felt pressure from their company to innovate quickly and worryingly according to the Testplant’s survey, this resulted in 49% putting out apps “before they go through ideal testing.” Just under half (45%) knowingly put out apps that “will perform below initial requirements”, it was found.
And when it comes to specific app testing, automation beyond execution was ranked as the “most critical challenge” by 67% of the sample, though 70% claimed they are now more focused on automated testing than they were in the past.
Bates, who was previously CEO of PLAT.One, a company with an enterprise-grade IoT platform acquired by SAP back in September 2016, asserted: “It is reaching a crunch point for businesses, under pressure to provide the most amazing services and apps to consumers, and development teams facing internal and competitive pressures, as organizations rush towards digital transformation.”
He added: “Companies are struggling to keep up with the pace of digitalization and there will be a quality toll unless businesses see the imperative to completely re-think their traditional development processes and move towards more automated, intelligent solutions and tools.”
Digital Automation & Testing
Speaking on the sidelines of a Digital Automation Intelligence Roadshow at Absolute on the 29th floor of the Millbank Tower in London hosted this week, Dr Bates said: “It has moved from testing being about checking that the code [for an application] works, to now that in order for an application to be successful, does it delight the customer, does it get 5 stars in the App store and does it have a high net promoter score?”
He added: “An app success is not just about does it work, but does it win, does it delight the customer and does it get 5 stars? And. that is all about the user experience.”
As such Testplant’s goal is to use AI and analytics to automate it all, and according to Bates “figure out the system, to automatically create the test cases and coverage, then to analyse the results and to learn using AI.”
Making an analogy in his presentation in London, Bates said that in terms of where we currently are and mapping historically for where testing is: “We are not exactly in the dark ages, but probably in 1815 in terms of the use of real-time data. We are not learning from testing, not feeding it back and not using AI. Although we are automating tests, they are manually created just like algorithms. So, we are not considering what to test but automating the test process.”
The “real opportunity” for disruption in the testing space according to Bates is to “move to an AI-powered” learning approach to testing the system. Effectively, this involves analysing the system, automatically building the script and running them.
A tester’s job will never be easy given that the systems they work with vary in scale, complexity and cost, and systems development projects differ in timescales and the pressures facing participants in the space.
As Paul Gerrard, Principal of Gerrard Consulting and a member of the Working Party that produced the Component Test Standard (BS 7925), who attended the event in London, stated in ‘The Tester’s Pocketbook’: “Process-heavy technologies, technically dazzling tools and agile approaches work well in the right context but all have their shortcomings: high ceremony, cost, maintenance and ineffectiveness; poor productivity, reliability, accountability and return on investment.” But with AI and enhanced analytics we might just get there.
Testplant convenes its next digital automation intelligence roadshow in Los Angeles on Tuesday, 17 October, 2017, which will hear speakers including Anthony Edwards, Testplant's CTO, Michael Silverman of FIS and Samir Shah from UCLA Health. Thereafter it moves to Dallas on October 19 and Philadelphia in early November.
No comments:
Post a Comment