Wednesday, July 30, 2008

Comments on quality

Dare recently made a comment about quality from my previous post in SaaS. Here is right in what I meant is better accountability in SLA, but I was also thinking to about quality that is not in the software itself, but rather from customer service and accountability.

When Apache breaks in an enterprise, there is no one to call, no one to claim to directly, no one to fire (except the guy who made the decision to go with open source). There is the desire that the feature/bug be addressed in a timely manner. Also, new features are not driven by existing customers as much as when someone is paying real $$$.

If GE goes to Apache foundation and tells them to fix something, they would say: stay in line (or worse: that is stupid, fix your process). If GE goes to a big software company they are paying top dollars to and say the same, the answer will be: yes sir.

Also, if users start complaining that they don't understand how to configure apache, the best solution is to say: hire an expensive consultant or learn it. The quality that comes from other service because of big margins is: call our technical support, or read the tons of books, or heck..we'll even send you an expert on it.

As a matter of fact, this is exactly what Red Hat sells. They may not sell the software, but all that devs they have are either paid by support money they get from their users or from developers working for another big company that pays them kind of in the same way. Either way, the users always has to pay. There is no such things as a free lunch.

SaaS may be similar in that the margins are very very very small. You need tremendous scale to support infrastrucutre or transfer the cost to someone else. Ads don't have a place in productivity applications, so that cost is transfered to the user by subscription. What SaaS as basically done is said: we'll charge you much less for the same thing, and we are able to do this by economies of scale. And we'll also reduce your TCO because we'll do the upgrades and deployment for you. But nothing prevents a shrinked wrapped software company to say: wel'll deploy and fix all your problems for free, and we'll give you free upgrades for life. Just don't burden us with servers costs.

Tuesday, July 22, 2008

SaaS thoughts

Dare, a coworker has posted an article about Software as a Service (SaaS) and the paradox it creates for software vendors. His thoughts are on target when he compares the Innovator's Dilema theory that and cheaper and seemingly worse technology overtakes a higher profit and apparently better technology over time.

Even though I agree with Dare, I sometimes remember the multiple rises and falls of server/desktop environments and the prediction of how free software models will topple the old shrinked wrap ones. With Linux turning 17 years or OpenOffice at age 8, I still wonder how come they haven't fulfilled their rein the world destines. That is because things have to be put in perspective and analyzed case by case. (before you flame me, I know that Linux and OO is pervasive but you can't argue that the prediction of Linux ruling the desktop has not materialized and Windows still dominates even the server)

The main reason I believe that shrinked wrapped software still has plenty of life against SaaS (or open source) is because it brings something that neither free nor cloud can give easily: accountability and quality. Both things go hand in hand and cost a lot of money, things that SaaS has little of (low margins) and OS has none.

Accountability is probably the biggest one. Imagine a manager explaining that the company went down last weekend because your cloud provider was not working. Not easy to point fingers here: most will points to the manager. And this is not the same as electricity. Just because you can outsource marketing, back end processing, etc, doesn't mean you do it. You still keep some things close while you outsource others. Same with computing. You may outsource backup processing, peak load, and similar, but not the main processing.

Then comes quality. Quality is not only in service but also in products. To be fair, Salesforce, Linux desktop (KDE==windows or Mac) OpenOffice (==MS Office) and other alternatives take most of use-and-feel from existing software that were paid by larger profits.

Finally, comes the whole processing power in our hands thing. If we offload everything to the cloud, why do we need all that processing power for? I claim that software needs a bigger transformation, like smart AI, context awareness and Data Mining and use the cloud like a huge database. There is where I feel software + services will go.

I drifted off topic, like always, but anyways...