3 Comments

Great post, John!

In a follow-up interview, the CEO of SWA explained they were implementing a new release of the purchased app that cratered under the load--and that they'd implemented EIGHT releases in the prior year. NINE releases of a large, on-prem, app suite--with custom mods, integrations, and undoubtedly UX/CX/report/database changes in one year??? He wanted to sound reassuring, but nine releases of a complex, critical system in a year should scare the heck out of Board members!

And that FAA database error was also scary. A critical system shouldn't allow an employee to just overlay a production database. Again, it looked like they were saying, "Nothing to worry about; just a dumb contractor." And again, where are the IT checks and balances that ensure components are validated before migration to production?

Expand full comment
author

Thanks for the added context Wayne. As a former CIO, you know from first-hand experience of the many potential gotchas when it comes to enterprise software, from not updating enough (obsolescence) to updating too much (added complexity, less stability). Modern, cloud-based apps can help with some of this (i.e. minimizing the risks of software upgrades), but the challenges of establishing high-performing, highly-reliable IT environments seem as great as ever.

Expand full comment

Definitely agree that today's SaaS apps, especially industry vertical apps, can minimize technical debt because the vendor controls the update cadence. But adopting a SaaS app with its rapid update cycle requires an organization to transform how it tests and configures software, and especially how it documents & trains its users.

Expand full comment