Facebook and the downsides of software as a service

The tizzy created by Facebook's page design changes point out some valuable lessons that we should keep in mind as we head more into a SaaS and cloud-based world.

1. Choosing when to change

There are many differences between how shrink-wrapped applications and software as a service (SaaS) work, but one of them is that customers of shrink-wrapped software choose when, and if, they upgrade. They kick the tires to look around at the changes beforehand, download a trial, poll other users, wait for the .1 rev and the kinks to get worked out.

With SaaS, changes get pushed out without those wait-and-see possibilities. Facebook is discovering that this can lead to unpleasant surprises for customers, who have no say in whether they want to adopt them right then.

When something is embedded into the flow of everyday life in the way that Facebook (or Twitter) is for many people, any change, whether it's ultimately better or worse, is going to cause complaint. People get used to patterns of doing things. Even when you change their work-arounds, sometimes they don't like it.

2. Conversation is a double-edged sword

Having said that, on the flip side, SaaS is more responsive when there is feedback. It can turn around updates based on input more quickly, and obviously more universally. But do this too often and you whiplash your users with multiple changes that set and unset particular features, preferences, design decisions, and so on.

Facebook is going to have to tread carefully in the coming weeks as it decides how to respond to the considerable complaining about the new layout. Facebook is quite different from most "applications" because there are such a variety of ways that people use it, and the experiences that each user has are going to be quite different. (All the more so because of the openness of the platform.) That makes it hard to design for, and all the more important to check one's assumptions at the door about what people want to do with it, and what features will support those needs.

On a more macro scale, Facebook (and SaaS in general) are emblematic of the substantially two-way relationship that now exists between companies and customers. The real-time nature of the conversation--and with something like Facebook, the ability of customers to vocalize and organize--is a precursor to what the majority of companies will have to deal with in the future. As the "Cluetrain Manifesto" presciently argued, all markets are going to be more conversational in the future.

3. Don't design by committee

But that doesn't mean everything should turn into a design by (user) committee, or tyranny of the majority. That's not how excellent products get made. There has to be a balance between responding to feedback, and recognizing when you see possibilities that your users, for the time being, do not. Your job as a designer and a company is to create capabilities on their behalf, and not just implement exactly what users ask for. (Not in a high-handed way; users' needs and best interests should always be the focus.)

(It so happens that we had a very lively e-mail thread running around the frog offices about this exact topic recently. Do you stick to the vision or respond to feedback by changing the vision? The answer: "It depends." Not very satisfactory perhaps, but unfortunately basically true. There are well-known examples of hits and flops based on both approaches.)

Watch what happens...

We should all be watching very carefully how Facebook acts in the coming weeks as it responds to the conversation. It will undoubtedly provide lessons for the future for all of us.