I’m cur­rently in the process of try­ing to roll together a few host­ing accounts of mine that have unnec­es­sar­ily been run­ning for the past few years as I’ve dis­cov­ered ‘bet­ter’ ser­vices but not both­ered rolling the old accounts over to. I’m prob­a­bly los­ing around $700 a year because of this, and basi­cally killing any rev­enue from host­ing for other clients by mak­ing the whole process so unmanageable.

It’s frus­trat­ing, because it’s self-inflicted, rel­a­tively fix­able, and a prime exam­ple of ter­ri­ble stew­ard­ship on my part. It trick­les out of my account in minor trans­ac­tions of around $20, and I just don’t miss the money as much as I should. The prob­lem is, if the trans­fer gets screwed up, var­i­ous peo­ple depen­dent par­tic­u­larly on email (less so web­sites — they can with­stand being down for a period of time, but for most clients of mine email is THE cru­cial appli­ca­tion) cease to func­tion until the DNS mends itself.

For this rea­son, I’ve not been brave (fool­ish?) enough to entrust the task to a ser­vice such as ODesk or RentA­Coder, even though either has the poten­tial to totally take the headache away for a pretty min­i­mal cost. I know that the odds of some­thing going wrong between the expo­sure of core host­ing pass­words to strangers, in a process that is the web equiv­a­lent of a heart bypass, are pretty unac­cept­ably high to sim­ply palm off to some stranger for a chunk of change. Part of the prob­lem is the kind of busi­ness con­ti­nu­ity capa­bil­i­ties faced by many of my clients aren’t such that ser­vice noti­fi­ca­tions pro­vide any par­tic­u­lar ben­e­fit: it’s not as though they have any par­tic­u­lar rem­edy for ‘sched­uled main­te­nance’ in terms of noti­fy­ing clients, as, unlike larger busi­nesses, their web­sites are not fre­quently vis­ited as first point of con­tact for vast num­bers of exist­ing and prospec­tive clients of their own, and pub­lic apolo­gies or noti­fi­ca­tions are meaningless.

Sim­i­larly, the scope of client edu­ca­tion for such a minor under­tak­ing is itself quite for­mi­da­ble — this sort of out­age is highly occa­sional and the clients are so var­ied that there are no par­tic­u­lar processes in place for deal­ing with it. Micro web agen­cies aren’t gen­er­ally well equipped to do this sort of thing, sim­ply as a byprod­uct of the nature of the provider/client rela­tion­ship. In my work with larger busi­nesses (espe­cially where SaaS is a core offer­ing) where the rela­tion­ship is less provider/client and more embed­ded (i.e. I don’t end up func­tion­ing as an exter­nal party!) we have of course for­mu­lated plans for con­ti­nu­ity and noti­fi­ca­tion, but this can­not be the case as sim­ply or read­ily for smaller, more frag­mented organisations.

These issues have pro­duced some­thing of a per­fect storm, where minor recur­rent losses are the path of least resis­tance in a sit­u­a­tion that requires a fair amount of (non-financial, tan­gi­ble and intan­gi­ble) invest­ment to cor­rect course. This, com­bined with the fact that I don’t have enough clients to jus­tify writ­ing migra­tion code, and the gen­eral awful­ness (par­tic­u­larly the glacial slow­ness)of WHM/other pro­pri­etary host man­age­ment soft­ware, has meant I’ve yet to embark on an exer­cise with lit­tle vis­i­ble ben­e­fit. Over 50% of this task is stuff you can’t out­source, or at least shouldn’t: client edu­ca­tion and main­tain­ing relationships.

Small busi­ness own­ers: how do you bal­ance this need for process improve­ment and cost sav­ing with the real­ity of day-to-day busy­ness and your oblig­a­tions as a provider?