Friday, April 14, 2006

shift: the power of software usage analysis

I've given two speeches on this topic in the last two weeks, and I can't stop thinking about it.

You can argue that the reason all of us software producer types are in a lather over "software as a service" (SaaS) is because we think customers prefer it over packaged software. The economics are incredible - but whether anyone other than can make any money doing it remains to be seen.

But what has me all in a lather is what this shift will enable in the area of usage analysis, which I maintain is the product marketer's Grail.

With few exceptions, packaged software isn't able to provide the publisher with anonymous data regarding how it is used. So those of us who have to figure out what to build next, what to promote, what to position, etc do so with little to no objective data around whether what we've built/promoted/positioned worked or not.

Compare this to what the web crowd has had from day one - rich, comprehensive and immediate feedback on how content is used, by whom, for how long, and how often.

When the buzzword was "sticky", as in making pages more sticky, web analytics were king. With today's SaaS meme in high gear, analytics are back with a vengeance.

Closing the loop between manufacturing, distribution and customers requires some information on how they use the product. This information should be used to guide development, support, marketing- frankly, just about everything the publisher does.

Now instead of page views, imagine you had bear-realtime access to a complete picture of how your customers value your product. What would you do with this data?

I could do a lot with that sort of information.

Anyhow - both at SW2006 and SLAM, my payoff slide had three points I wanted to make to the packaged software producers in the crowd who were starting to make the shift from physical to digital distribution. My thesis was that by putting these three points into their roadmap, they could get ready for the shift in a non-disruptive way.

  1. Customer care comes first. Digital distribution will mean customers can get access to your new hotness nearly as soon as you've done shoving it through QA. If you can't support the uptick in support that comes with this, don't start. A bad customer experience is death.
  2. Put usage analytics at the heart of your business. I believe you need to measure everything associated with how customers use your products. Not just how often and for how long, but what modules they use, the navigation paths, even how they use help. When everything starts to move really fast, your only friend is your metrics, so be prepared for the flood of data and plan for how you'll use it. The day someone asks you "what was the impact of that marketing campaign" you'll be able to answer them definitively.
  3. Embrace new opportunities for pricing and packaging your solution - and test, test, test. If you accept from day one that your product's pricing and packaging will change - in profound ways - you build your products from day one with this in mind. This gives you the flexibility to try out new things, see how they work, without crushing your development organization. It also puts marketing in the spotlight, since they should be the ones coming to the table with fresh ideas for serving customers better.

These three things build on each other - but usage analytics are the most critical. They make your work in support more focused and effective, and they do the same for your pricing and packaging decisions. Without usage data, you're flying blind.

Reliable, comprehensive usage data is not something that packaged software vendors have ever enjoyed in great numbers. Putting all of the hype and customer buzz around SaaS aside, I think once producers figure out that digital distribution gives them this data by default, they'll realize the competitive advantage it gives the SaaS crowd, and look for ways to get the same advantage.

In the meantime, as the SaaS space continues to mature, smart publishers of packaged software should:

  1. Start to implement technologies for gathering usage data around their packaged software and actually use it to drive development/sales/marketing/support decisions
  2. Implement ways to flexibly price and package their solutions independent of the development cycle
  3. Make sure their support organizations are prepared for when the usage flood comes

No comments: