The End of Upgrades: A Manifesto

December 10, 2007
By

Joshua Greenbaum

• Submit Feedback »
• More by Author »

Then there are the hidden dangers of upgrades. My favorite comes from a conversation I had with a client of IntelliCorp., which sells software to help minimizing the hassle and maximize the success of this charming upgrade process. This company, which will remain nameless, found out that in the course of its upgrades it was potentially exposing valuable partner trade secrets buried inside the ERP system it was upgrading. Nothing in the standard upgrade kit from its vendor allowed it to safeguard these trade secrets, which if they leaked out to the outside world or were re-implemented incorrectly during the upgrade process would have basically ended a lucrative business relationship for the upgradee.

Luckily IntelliCorp actually has tools that can keep this kind of upgrade on the up and up, and by the way, in regulatory compliance as well, which is another “at risk” problem inside more upgrades than most of us would like to imagine.

Of course, this “end of upgrades” manifesto needs to end with a little reality check: upgrades aren’t going to end soon. If you’re an Oracle customer, Oracle has promised that you don’t have to upgrade anything, but chances are you will anyway. There are some compelling reasons to upgrade your technology stack, particularly to make good use of the new “edge” applications that will provide competitive advantage to your otherwise competitively dying core ERP system. And someday, off in the distant future, a Fusion Applications upgrade looms for every Oracle customer.

If you’re an SAP customer, SAP is actually trying to fix this problem with its Enhancement Packs, which are basically mini-upgrades that are intended to provide a major-upgrade’s worth of new functionality. The EP’s are just starting to hit the market, and early conversations with users have been very positive, so this may be one way in which the problem gets solved. While it’s still an upgrade, the EP’s go a long way towards making the upgrade a relatively trivial event.

The other way to end upgrades, of course, comes from deploying on-demand solutions, which of course still need lots of upgrades, all of which, at least theoretically, are transparent to the users. But don’t think you don’t pay for upgrades in the on-demand or SaaS model: you do, it’s just built into the per user per month price you pay. Which is why some calculations show that after five years of paying per user per month, your total cost of ownership for a SaaS/on demand solution starts to resemble the TCO of an on-premise system.

So, the end of upgrades? Not in our lifetime. But I do believe an end to the sycophantic acceptance of the burden of upgrades is long overdue. The disruptive nature of the upgrade process, combined with the disruptive cost and the generally disruptive risks therein, should be seen as a problem in desperate need of relief, if not resolution. It’s a sad day when we neglect the need to question the obvious and fail to demand that something truly egregious be made to be significantly less so. Upgrades suck, and it’s time that they sucked a whole lot less. Way past time, if you ask me.