IN 2018 A field technician working for Telstra, an Australian telecoms firm, built an app that unified 70 messaging systems for reporting phone-line problems. The technician did this despite having no coding experience. The interface may look cluttered: the landing page jams in 150 buttons and a local-news ticker—the app equivalent of an airplane cockpit, quips Charles Lamanna of Microsoft, who oversees the software titan’s Power Apps platform that made it possible. But it has been a hit. Some 1,300 other Telstra technicians employ it, saving the firm an annual $12m.
Professional developers (pro devs) might poke fun at the technician’s DIY app. But the trend it exemplifies is no joke. Since well before 2017, when Chris Wanstrath, co-founder of GitHub, a coding-collaboration site, declared that “the future of coding is no coding at all”, so-called low code/no code (LC/NC) tools have burgeoned. They allow anyone to write software using drag-and-drop visual interfaces alone (no code) or with a bit of code creeping in (low code). Under the hood, this is translated into pre-written or automatically generated code, which then whirs away.
Such tools are in hot demand. Just 25m people around the world are fluent in standard programming languages, reckons Evans Data Corporation, a research firm—one for every 125 people in the global workforce and 1.4m fewer than needed. That shortfall will rise to 4m by 2025, says IDC, a research firm. LC/NC products expand the pool of coders to “line-of-business” employees who seldom speak C++, Java or Python. And beyond. Cheryl Feldman went from a junior position in a hair salon to a technical career at Salesforce, a software firm, thanks to LC/NC. Samit Saini changed jobs after 13 years as a security guard at Heathrow to become an “ IT solution specialist” at the airport after making software on Microsoft’s Power Apps.
Overcoming language barriers
IDC reckons the low/no coders numbered 2.6m globally in 2021. It expects their ranks to swell by an average of 40% a year until 2025, three times as fast as the total developer population. The number of organisations using Power Apps more than doubled in 2021. It now has 10m monthly users. BASF, a chemicals firm, uses it to let 122,000 workers write software. A study last year by Aite-Novarica Group, a consultancy, found that over half of American insurers have deployed or plan to deploy LC/NC. Unqork, a no-code startup valued at over $2bn and backed by Goldman Sachs, is convincing other financial firms to take the plunge. Mr Lamanna envisages a global population of a billion low/no coders.
The dream of codelessness is not new. Tony Wasserman of Carnegie Mellon University’s branch in Silicon Valley dates it back to the concept of “automatic programming” in the 1960s. Since then successive waves of simplification and abstraction have made life easier for programmers by distancing coding languages further from the machine code understood by computer hardware. In the early 1990s Microsoft tried to simplify things further by launching Visual Basic, an early stab at LC/NC. In the next decade firms like Appian, Caspio, Mendix and Salesforce began offering products aimed expressly at line-of-business types.
Recently LC/NC’s potential has been unlocked by the cloud, which lets people connect to data easily and collaborate in real time, says Ryan Ellis, who leads LC/NC products at Salesforce. Last year Amazon Web Services (AWS), the online giant’s cloud-computing arm, introduced Amazon SageMaker Canvas, a set of tools that lets people deploy machine-learning models without writing code. It also offers Honeycode, a no-code app builder, in beta version.
LC/NC used to be chiefly about making pro devs more efficient. Now it is also about pulling more humans into creating applications, says Adam Seligman of AWS. In terms of impact, he says, the latest wave “will race higher up the beach”. For one thing, firms in a hurry to digitise appreciate that when line-of-business people design software, it speeds things up. “A field worker making something for other field workers is hugely valuable as the feedback loop is faster,” says Adam Barr, a former Microsoft pro dev and author of “The Problem with Software: Why Smart Engineers Write Bad Code”. As digital natives enter the workforce they are also demanding automation of repetitive or manual data-entry tasks, often on pain of walking out.
In addition, LC/NC is fast becoming the secret sauce in modern software development, notably in machine learning, says Arnal Dayaratna of IDC. The mastery of Python or Java required for this type of artificial-intelligence (AI) software is daunting even for pro devs. Bratin Saha, who oversees AWS’s machine-learning services, wants SageMaker Canvas to empower regular business analysts—from marketing or finance, say—to deploy machine learning. That could increase the number of AI specialists available to businesses by an order of magnitude, he predicts.
Some scepticism is warranted. Just because non-programmers are able to build an application with LC/NC tools does not mean it will be any good, says Mr Wasserman, just as bug-ridden spreadsheets yield faulty results. They could also become a headache for corporate IT departments if citizen developers collect customer data that are worthless or, worse, that violate privacy. Especially with no code, businesses can find that the functionality they need does not yet exist. No-code platforms make the first 90% of delivering a useful application easy, and the last 5% often impossible, says Tim Bray, a pro dev formerly of AWS. And many pro devs remain resistant. Although they turn to LC/NC to simplify some tasks, plenty of pros see it as the programming cousin of pin-it-on neckties, in the words of one commentator. Some worry that specialising in LC/NC makes them look like dilettantes, reports Mr Barr.
LC/NC will not displace “full” coding altogether, as its evangelists insist. Pro devs will continue writing their firms’ core products and mission-critical enterprise systems. But they will increasingly be complemented by legions of enterprising line-of-business workers with a software-development string to their bow. For employers, this means greater productivity. For employees, it could be life-changing. In 2019 the Telstra technician became senior business specialist for field digitisation and has since been promoted again. ■
For more expert analysis of the biggest stories in economics, business and markets, sign up to Money Talks, our weekly newsletter.
This article appeared in the Business section of the print edition under the headline “Going codeless”
https://www.economist.com/business/2022/01/29/what-if-all-workers-wrote-software-not-just-the-geek-elite