Typical SaaS Operation
(continued from previous page)

So why does it look unpromising? Being typical, YTSC Operations has the following traits:

  • Operations is under the auspices of Engineering. There is no VP of Operations; there is no Operations
    group. A Sys Admin is managing the production servers and probably doing office IT on the side.

  • The CTO is responsible for uptime, availability and performance. Does the CTO have an Operations
    background? I'll bet my lunch money that he doesn't. Is there a Staging platform? Probably not. Can the
    engineers log into production servers and modify configurations? Yeah. Actually we just fixed that nasty
    bug during lunch break.

  • There is no application-level monitoring in place, or trend analysis.

  • Is Customer Churn being tracked and analyzed? (What? What was that?)

  • There is no 24X7 support, although YTSC claims it is a 24X7 shop.

  • Are the following crucial practices defined and followed?

Change management – the cause of over 60% of downtime is caused by good intentioned modifications to
the platform. Is there a proper process in place? Is there an RFC (Request For Change) form and procedure?
A change committee?

Incident Management – are Support, Operations and Engineering aligned in a well rehearsed routine; roles
and responsibilities defined? Is there an Incident management system in place? How about a
knowledgebase?

Configuration Management – are hundreds of moving parts accounted for? Are they linked into the Change
Management process – actually, we don’t have a Change Management practice.

Availability Management – how do you analyze unavailability? How do you “budget” downtime? Do you know
where to invest your next Dollar to ensure optimal availability? It should be all tied into an Incident
management system. But, wait we don’t have one.

Release Management – how, when, how often, naming conventions. How does it tie into Change
Management and Configuration Management?

SLA Management – Are we providing what we promised? Are we tracking effect of incidents on customers?
Are we compensating them according to our contractual commitments? Is it tied into our (hosted) CRM
solution? Hard to do without an Incident Management system.

No doubt, parts of these practices have been in place with less fancy names. Otherwise YTSC would not have
survived this far. But Excel and Notepad will not suffice for a large scale operation.

Most companies understand that (or maybe that is wishful thinking), but when having to chose between
investing the next Dollar in great features that customers have been begging for, or that ugly, boring,
misunderstood, 800 pound gorilla, they will opt for the former. Pay now or pay later.