We interrupt our apparent mini-series on BPM vendor evolution to bring you more exciting reports from the front lines of business process failure…
As you may have read in my posting on testing, I tried, I really tried, to give the AT&T Online Vault back-up service a fair shot at my business. The upshot of that attempt was that I downloaded the software and a patch for the software, and left it running on my computer, which was plugged in and online, as requested for a first backup. I went to bed, and awoke the next morning, to find less than 5 percent of the fewer than 300 megabytes of files I wanted to protect were backed up.
Infuriated, I dug through the AT&T software and online documentation, only to find no clear, explicit instructions for how to deactivate my free trial and the software. So I e-mailed the helpful representative who'd fixed my original software problems, and found out that removing my PC from the list of systems known to and approved by the software would deactivate the service. So I did that. I then used Windows' trusty "Add/Remove Programs" feature to remove the relevant software.
Only it didn't all go away. And I didn't realize this until my system repeatedly, suddenly, and mysteriously slowed down and started generating unusual amounts of online traffic. So I poked around in Task Manager, to find a process related to my terminated AT&T Online Vault account still trying to make backups! I had to restart my system and keep it offline long enough to remove the software entirely, before it tried again. (That software also required that I change the physical location of some of my files, so it wouldn't try backing them up, but did not undo this change upon being uninstalled.)
So I went to my old online backup provider, Backup.com, formerly "@Backup." Within an hour, I had activated a free trial account, downloaded and installed their software, told it what files I wanted to back up, backed them up. I also within that time set the schedule for regular subsequent backups, which have gone off without a hitch.
Now, I don't know how it is that "the new AT&T," with all of its resources and experience, couldn't figure out how to deliver an online backup service at least equal to what Backup.com provides, in terms of clarity of instruction, ease of use, flexibility, and functionality. Frankly, I don't know why AT&T didn't just craft a licensing or co-marketing deal with Backup.com, which has been in the business since 1997. But I suspect there are some broken business processes and/or political considerations behind how such decisions get and got made. I could be wrong. But I doubt it.
In any case, the BPM-related lessons here are that it's not enough to test how new software or business services run, nor how well they do (or do not) install successfully on a variety of systems. It's also essential to test how well they can be de-installed or rolled back, and how well they can restore host systems to their respective states before the attempted installation or invocation. And, of course, this applies to supporting documentation as well. Just as "measure twice, cut once" is always true in construction and repair, "test more, support less" is always true with new applications, services, or technologies.
Regarding NetIdentity, they have for years offered a great domain-sharing service that, among other things, makes having an e-mail address of "yourfirstname@yourlastname.com" relatively easy and inexpensive to have. I had used NetIdentity as my primary personal e-mail provider happily for years as well. Then, the company was purchased by Tucows Inc., a premier site for downloads of freeware, shareware, and "trialware." Now, I have no idea about the internal heavy lifting required when the NetIdentity services were moved from their original infrastructure to the Tucows infrastructure. What I do know, however, is that ever since then, my e-mail access has been far less reliable, and often much slower. I also know this problem is endemic, based on the sometimes incendiary comments made online about Tucows, its customer support, and its management. I haven't given up the service yet, but I'm using Google's Gmail a LOT more.
The BPM lessons here are two-fold. Fold the First: Make every upgrade and infrastructure transition as invisible to users as possible, especially where software as a service (SaaS) is concerned. Salesforce.com, Inc. and WebEx Communications, Inc. are two companies very good at this, and useful examples of how it should be done.
Fold the Second: Assume that every time you try to make an upgrade or infrastructure change invisible, it won't be. That means communicate clearly, before, during, and after the transition, and keep users and other stakeholders in the loop regarding progress. That way, even if there is a surprise malfunction or other hiccup, those users and stakeholders are far less likely to panic, rebel, and cast blame.
Make sure processes at your business address these sometimes-hidden issues, explicitly and effectively. And test those processes from time to time, to make sure they work the ways you think they should. Let me know how process testing works – and doesn't work – at your organization, or others known to you, and let's see if we can't improve things at least a little.
Michael Dortch has been an analyst, 'information entrepreneur,' speaker, and writer about IT and 'the real world" for the past 30 years. After almost 10 years at the Robert Frances Group (RFG), Michael begins a new position with Aberdeen Group in September.
Dennis Byron is an analyst with ebizQ, focusing on Open Source Software as well as Business Process Management technologies.
