Showstopper SaaS requirement: Leverage SaaS

Napoleon once condemned overly-cautious generals by saying, "If you're going to take Vienna, take Vienna." You can apply the same principle to software-as-a-service.

Dan Olsen's excellent presentation on product management for SaaS products--a topic dear to my own heart--is really a plea for seizing the opportunity that SaaS provides. Among other potential benefits, Olsen points out that SaaS gives you a window into customers. How do they use the product? What's the normal adoption pattern? Is there any warning behavior that indicates someone might stop using the product?

The problem is, these capabilities don't come for free. Tracking features are like any feature: they take development resources. However, here's a good test of what kind of company you really are. If you think your girlfriend is like a fickle paramour, happy only as long as you keep showering gifts (i.e., features), you'll never invest in tracking features, because customers won't ever see them. ("Why did you buy something for yourself, when you could have bought something for me?")

I hope that you work for a different sort of company than that. At some point, you need to step back and reassess. SaaS gives you the kind of sudden insights that, in a different type of relationship, are only possible through long, expensive therapy sessions--unless you don't care if the other party decides one day to leave.

Comments

[Saeed Khan]

Tom,

Good post. To paraphrase GI Joe -- know what customers use or don't use is only half the battle.

The other, and in my opinion, more important half, is knowing what they need.

Measuring usage is a post implementation activity. Good, but really it's like looking in the rear view mirror.

Understanding needs is very difficult (if not impossible) to do strictly through automation. You need to meet people, talk to them, observe them, their colleagues etc to get a good understanding of needs.

As an example. Intuit's "Follow me home" program gave them insights into how their customers used their product (Quicken) that would have been impossible to learn in any other way.

I'm not decrying usage metrics in any way, but just pointing out some of their limits.