In my years as a cloud computing consultant, I am always taken back by my client’s request to have techs with vertical-specific knowledge help with their cloud projects. Moreover, they seek vertical-specific clouds and cloud services, or clouds that are known to do well in a particular industry or sector and have services aligned with those industries.
Although this may make sense at a broad level, there are some things that should call this industry-specific approach into question as you move to the cloud. Here are two caveats that I’ve been considering lately.
Capabilities matter more than industry
The capabilities of the technology are more important than the canned vertical-oriented processes and/or prebuilt industry specific metadata.I’m finding that most technology that is vertically specific, in the cloud or not, is typically not the best of breed for that technology.
For example, I’ve seen companies pick cloud databases largely due to fact that they are more vertically aligned. However, they often are poor cloud-based databases compared to others on the same cloud or other clouds, and selecting them limits the enterprise’s abilities to do advanced tasks that they most likely need now or will need in the future.
By picking a vertical database, may get your data up and running a bit faster but give up best-of-breed capabilities. The end result may not be optimal and could well cost you way more than the value of the vertical-specific cloud database.
And of course, cloud databases are just an example; the principle applies to any technology.
Industry-specific tools can be a poor start
Often, using industry-specific processes or metadata don’t provide a sound starting point.The advantage of building something from scratch is that you have complete control over the resulting solution. But in many instances, enterprises moving to cloud pick a cloud service or a cloud due because they have prebuilt processes for their industry—only to find out that these processes need a good deal of changing to adapt to the enterprise’s actual or needed processes.
It’s actually harder to change existing processes and data than to build something from a blank page—and starting from scratch is less risky if done right. You’re also much more likely to be using better technology as a result as well.
This seem counter-intuitive considering that we’ve been buying industry-specific technology for the last 30 years, so it may seem to make sense to carry that approach to the cloud. But ERP systems went through this expensive recustomization dance in the 1990s, and you should not repeat that dance in the cloud in the 2010s.