Subscribe / Unsubscribe Enewsletters | Login | Register

Pencil Banner

What your SaaS vendor doesn't know about you

David Taber | May 2, 2014
You'd think your SaaS vendors would know a lot more about you and their customers than IBM, Ma Bell or even the NSA. They know every click, login and data entry, right? Not necessarily.

Software as a Service (SaaS) vendors have a huge advantage over traditional enterprise software vendors, because they host and process all your user interactions. They not only know how many licenses you bought, they know how many you use, how often and even which browser you run. This makes for a huge advantage in marketing and customer support, and it could provide the basis for some powerful information products based on aggregate usage.

Google most obviously uses this SaaS model: It knows what you're searching for, reads your Gmail and knows the timing (and maybe even the content) of your Google Phone calls - all in the name of service of well-targeted advertising. I should know. As a consultant, the only ads that ever show up are ads for my clients. Google's algorithms know I'm very interested in these companies, so they figure I should find out even more about them through their advertising.

As I pointed out more than five years ago, SaaS vendors don't need usability labs or focus groups. They can just monitor how customers use their product in real-time, in real situations. SaaS, like no other business model before, can be instrumented to provide powerful insights into user behavior and market preferences.

In theory, that is. In practice, however, this level of user measurement is typically limited to smaller companies and earlier-stage products. Google is the exception that proves the rule: No "normal" company can afford to instrument its flagship SaaS products to get tons of usable information.

More Customers, More Computations, More Complications

Why? Think big data. To make the information from clickstreams useful in understanding user problems, customer preferences and market trends, you must link the individual clicks across a long series of user entries to give the clicks meaning in context. You'd need to subdivide the aggregations by user language, product version, geography and user demographics. You'd need to separate "idiot user" actions from "expert who can't find her way through our mess" actions. You'd need extensive A/B testing.

If a SaaS vendor has 1 million users, then it has 1 billion clicks and 1 trillion analytical permutations. Just the act of collecting a comprehensive data set could affect product performance, let alone storage. I've heard people say they can't find anything off the shelf that can scale up to the task.

That's where Wall Street comes in. It won't be big boosters for all this data collection and digestion; it has no short-term payoff but plenty of short-term costs. In other words, it's poison.

It's conceptually possible for SaaS vendors to lower support costs, facilitate the ultimate user-centric design and have much more effective marketing, but the commercial reality is much more primitive. Many SaaS vendors we've worked with barely have the customer master synchronized between their operational system, their accounting system and their CRM. (Seriously, we've seen our share of SaaS vendors who "sync" this stuff by redundant typing.)


1  2  Next Page 

Sign up for CIO Asia eNewsletters.