Last year, I researched the impact of the proliferation of cloud computing on business. I know it sounds like a dry topic, but if you are a CIO or an application development manager, the results can have serious meaning.
Cloud computing has deeply penetrated the enterprise and SMB ranks, and 46.3 percent of respondents said they had four or more cloud applications in use, my partner, Esteban Kolsky, and I discovered. Some have dozens.
That’s great for the cloud because it validates what Kolsky and I (and many others) have been saying for more than a decade: The cloud is just flat out a better way to deliver application functionality at a reasonable price to a growing universe of companies.
Of course, there was an underside side to the results as well. With so many cloud apps in use, many IT departments found that one of their major responsibilities had become developing and maintaining integrations among many separate clouds.
The integration technology available today is quite good at making cloud integration straightforward, but it’s not free of effort. When something changes in one cloud app, IT staff often are compelled to understand how the change ripples through the rest of the company’s cloud infrastructure.
Apps built on the same platform have an advantage because a common platform provides standards that make communicating between them easier; nonetheless, developers still need to patrol the interfaces. Additionally, while the population of cloud apps on many platforms is large and growing (they’re referred to as “ecosystems”), it is still finite, and their granularity or fitness for a particular business purpose is variable.
Also, at the very smallest level, an off-the-shelf app still will need some editing, tweaking or customization — if you can find an app that’s close to your need, that is. When all else fails, lines of business have resorted to spreadsheet apps, which often bring more trouble than they are worth.
One of the inescapable conclusions I draw from all this is that we need better ways to ride herd on all our cloud apps with far less effort.
By extension, businesses also need more than individual cloud apps, more than integration services, and even more than high-level standards or advanced development tools. If you have a business today and you are trying to keep up with user requests for new or revised cloud apps, you really need a standardized app platform with tools that you can standardize on and that works well with the rest of cloud-land.
Ideally, there also should be a universe or library of app specifications that you can download and trade so that every development project, no matter how small, doesn’t have to start from scratch.
Notice the use of specifications because code should be generated, not written. We’ve seen this before with code libraries, but I’m thinking that code won’t work if you’re trying to address the needs of line-of-business users, many of whom don’t write code. Those users need a simple facility that lets them describe in a nonprocedural way the apps that support real-world business processes.
Having an accessible library of apps also means that there are potentially no app requirements that are too small for IT, meaning we can seriously consider eliminating spreadsheet apps.
Digital Disruption Marches On
Now, you might think that this has always been the case and that we’ve always been striving to have enough useful software to run our businesses, so what’s the big deal?
There are two simple points. First, not having enough of the right software that’s easy to use and maintain has been a drag on business historically, but all businesses grew together more or less so the drag was a relative burden and not absolute.
When there was no CRM we got along with a variety of paper files, faulty memory and spreadsheets. Once CRM became available, it was such a step forward that virtually every business simply had to have it.
The second point is this: The so-called digital disruption has entered a new phase in which businesses will need all kinds of software functionality to advance. However, there isn’t enough software on the shelf to satisfy the demand, and old-fashion approaches to developing and maintaining it already are insufficient. That situation only is going to get worse.
To net this out, what Kolsky and I saw in cloud computing last year is really just the beginning of a strain that will rupture the current software paradigm. What comes next has to be better, faster and cheaper — and no, you can’t pick just two out of three.