Money can’t buy you happiness, but developers just might. According to a new survey from Stripe, companies finally recognize that access to engineering talent is a bigger inhibitor to growth than access to capital. In fact, as fed up as enterprises may be with their outdated IT infrastructure, they’re convinced that if they can just find good developers, most other problems will prove secondary.
But just hiring more developers isn’t the answer. There’s more to it.
Of course, you do need more development capability. Mainstream enterprises now recognize that their biggest threat is likely not another bank or a rival hospitality company. Instead, it’s that a currently noncompetitive technology vendor today will be a clear and present danger tomorrow. Those tech vendors are fueled by developers; everyone else therefore feels they must “developer up,” too.
Plant a developer. Watch her grow
The No. 1 constraint to enterprise growth, according to the thousands surveyed by Stripe, is access to talent, with 55 percent of respondents citing this concern. Immigration requirements—something that rightly gets a lot of media attention—was last on the list at 47 percent. Right at the top of the list, however—and higher than the need for cash to invest in growth (52 percent)—was developers (53 percent).
It perhaps makes sense for developers to think themselves the center of the universe, but this Stripe survey, conducted in partnership with Harris Poll, surveyed not only 1,000 developers but also more than 1,000 C-level executives in the United States, Singapore, France, Germany, and the United Kingdom. This mélange of developers and executives identified a need for more developers (61 percent) as a bigger threat to corporate success than tax rates (58 percent), trade barriers (56 percent), and Brexit (55 percent).
Not surprisingly, a full 96 percent set a high to medium priority on increasing the productivity of their developers, with 56 percent also indicating that they had expanded their developer population over the past year.
Not that they have much choice. With the specter of tech companies entering their markets, companies from seemingly tech-immune industries like farming and manufacturing are running scared, and desperately trying to hire more developers.
This, however, may not be the answer.
Getting more developers isn’t the solution
Not that companies aren’t trying their best to load up on tech to compete, all the same. Across diverse industries, the No. 1 area of investment for the next five years is software infrastructure and tech (43 percent), with the related areas of R&D (31 percent) and recruiting tech talent (31 percent) rounding out the Top 3. More traditional areas like marketing (29 percent), sales (26 percent), and customer service (24 percent) are now see as also-rans in helping enterprises to differentiate. A full 81 percent of those surveyed say that software development must become a core competency over the next decade, whatever their industry.
Again, to compete with tech, enterprises increasingly realize that they must become tech; that they’re not simply a grocery retailer or hotel chain or hospital system. To win in the 21st Century, software must be the heart of your business. That, in turn, heightens the need to hire great software developers—even as hiring talent becomes ever harder.
However, finding new developers may not be the best answer to the talent shortage.
After all, 77 percent of those surveyed by Stripe and Harris are very or somewhat confident that their company already has “sufficient engineering resources to keep up with technology trends.” For those that aren’t confident, “we don’t have enough skilled employees” (42 percent) is a top reason, but it probably misses the point.
With the average developer working 41.1 hours each week, and close to half of this (17.3 hours) is spent on maintenance (fixing bad code, debugging, refactoring, etc.), it seems that the fastest path to developer productivity isn’t to throw more bodies at the problem. Instead, it’s to improve the quality of software.
This, of course, is easier said than done.
If your business is successful, legacy systems and technical debt are unavoidable. Even so, there are countless ways to approach the improvement of code quality (see here and here and here), including better documentation, writing unit and functional tests, and more.
There’s also the obvious-yet-too-often-ignored reality of open source. If you’re writing a custom Kubernetes replacement (yes, I’ve seen this), you’re doing it wrong. Smart companies embrace the best open source projects that get battle-tested by a wide array of companies, rather than roll their own. As consultant Tobie Langel tells it, open source lets developers “write less code” by borrowing great open source code and then “just writ[ing] glue code” to pull it together.
Of course, this works best if you “train your developers by getting them to contribute more to open source,” because new research has shown that open source contributors hold distinct advantages over those who merely free-ride on open source. The more you contribute, the more you understand the code and are better positioned to write and use great code.
Similarly, if there’s a cloud service offered by Google, Microsoft, or Amazon Web Services that fits your needs, use it. If you’re worried about lockin, the bigger problem is your developers’ productivity. Besides, how is it less a question of lockin when your own developers build a system you can’t escape than if you happen to buy it from a cloud vendor? It’s not.
Years ago, Gartner analyst Svetlana Sicular addressed the shortage of data scientists, positing that “organizations already have people who know their own data better than mystical data scientists.” The shortage of developers is much the same. The answer to most companies’ developer shortages is likely to look within, and make those developers more productive. Open source is a key (though not the only) way.
As enterprises focus more on enabling developer productivity they won’t need to make such a fetish of finding new developers. As Fred Brooks’s The Mythical Man-Month once made clear, “adding manpower to a late software project makes it later.” In similar fashion, adding developers to crappy code simply means you have more developers frittering away precious time on crappy code. Yes, more gets done, but at a higher price than necessary.