Software engineering is still a young discipline. Until the Agile Manifesto challenged the accepted waterfall-based development model, software creation borrowed most of its processes from that of hardware. But those older, hardware-like methodologies failed to take advantage of software’s biggest benefit: that it’s soft.
What has emerged in the last 15 to 20 years is the idea of “failing quickly.” Developers want to get products in front of users as early as possible, learn from their mistakes, and iterate better versions quickly. This new process takes advantage of the “softness” of software and ultimately sculpts better products. Mantras like “fail quickly” assume that some large percentage of ideas are bad. The faster you can get those ideas in front of people and find out how they perform, the more “at bats” you’ll get and the better your chances are at finding an innovation.
Pairing Softness and Self-Service On-Demand Computing
Cloud computing and Agile development proved to be a good match for each other. Software could be developed quickly and at much lower hosting costs that match demand to lower scales—in a way that wasn’t possible when deployments were limited to physical hardware. Startups and venture capitalists (VCs) quickly caught onto this, as a 2016 graph by Pitchbook from an article over at GeekWire shows us:
Take a look at how the number of funds closed rose more sharply than the amount of capital received from 2009 to 2013. That is as a result of VCs spreading their bets across a wider set of companies, knowing they could invest less in each individual start-up because of the Agile/cloud combination. What this shows is that at a macro level, VCs now also embrace the “fail quickly” model in an attempt to get more chances to find innovation.
Contrasting Agile and VCs with Traditional Enterprise IT
Most enterprise IT gets funded very differently—and forces organizations to act very differently. Typically, the CFO projects revenue for a particular year. Then the CIO commits to spending no more than a fixed percentage of that revenue, usually somewhere around 3 percent, with some variance for particular verticals. That’s exactly the opposite of how VCs fund startups in the Agile/cloud era. Instead of rewarding risk taking and early identification of failure, this funding model incentivizes IT to optimize efficiency of capital assets—namely servers.
The result is usually some byzantine ticketing process that takes weeks and multiple levels of authorization to get a simple VM launched. On the public cloud, that can be done in minutes . This encourages Shadow IT, as line-of-business teams—who simply want to find software innovation need the same self-service on-demand access to resources their smaller startup competitors enjoy. That’s why claims by Gartner that CMOs spend more on IT than CIOs ring true, because CMOs spend it on public cloud resources when their own IT organizations can’t provide the same access to resources quickly enough.
Change the Funding Model for IT (Partially)
That’s not to say that every piece of technology that IT offers should be billed back to the line-of-business on a per-use basis. In some ways, the Enterprise has headaches that a startup just doesn’t, but that doesn’t mean that change is impossible. Imagine cutting that 3 percent fixed cost in half for things like email, HR systems and file sharing, or for tasks that have legal ramifications or whose metering is so small per user that it doesn’t make sense to bill directly back to the business teams.
That leaves plenty of room for billing out blogs, project wikis, virtual machines, block storage and a host of other things that are easy with private clouds or other SaaS-based solutions. It changes IT’s role from an impediment to an enabler of innovation, instead of worrying about cost efficiencies, security and lightweight governance. Using an Agile/cloud method strikes a balance between the model that startups enjoy and the realities of enterprise scale.
Latest posts by Pete Johnson
- Connected Devices, Remote Security: Data Encryption and Security in the Cloud - January 2, 2018
- Application-Defined Networking Basics: Definitions and reasons to think ADN vs. alternatives - December 8, 2017
- Cloud, Agile and What Startups Can Teach the Enterprise - November 15, 2017