Recently I stumbled across this question on Quora: What are some good metrics to track for an enterprise SaaS app?
And I loved this response by Omer Gottlieb:
While the obvious ones are Signup, login, purchase and important features on your app, you should also remember to measure import data, app customization and sharing content activities, since they are implying on a higher level of satisfaction and engagement from your customers.
At a previous SaaS company where I worked, we discovered that interaction with support was a key indicator. Specifically, if a trial customer interacted with our support team, their likelihood to purchase was went up to 50% compared with about 16% overall – a 200% increase in conversion rate! While we couldn’t truly pinpoint cause vs. effect, we nevertheless decided to make contacting support our second-most desired action (upgrading being number one). So all of our auto-responders and in-app communication emphasized support. We also had our outbound sales team treat their initial follow-up calls with new trials as support calls, not sales calls.
It took us quite a long time to make that realization. Not because we chose not to track and analyze it, though. We suspected that support might be a big selling point from the beginning, and we emphasized it somewhat in our communications. But we never had empirical evidence until much later in the company lifecycle. So, given how important it might be to our positioning, why didn’t we gather that data? Here’s why: because tracking it was such a hassle!
Many times over the years I’ve seen insightful, well-meaning articles with titles like “Why you should track this one super-specific thing that there’s no way in hell your web analytics tool is tracking.” And I’d think yeah, that’s a great idea. Then I’d realize what a pain in the ass it is to 1) capture the data and 2) build a report to display it. So usually we either wouldn’t track the data at all, or we would log it somewhere and pretty much forget about it until sometime down the road when it came up in a random meeting. (“Say, didn’t we decide to track the daily number of BlahBlahs compared to WhatsIts?” “Oh yeah, we did. I’ll get one of the dev team to run some queries.”)
And, sometimes we’d try to shoehorn it into Google Analytics. And every time, we failed to make use of it.
Which is why we created StatsMix: to turn metrics tracking and analysis into a painless process.
Later in that same Quora thread, Cody Crnkovich of GoodData recomends that
…to successfully measure your business, it is critical to start small, and iterate what, and how, you measure rapidly.
Internally we sometimes refer to StatsMix as “bottom-up BI.” (BI is industry-speak for Business Intelligence.) Instead of trying to analyze all your web traffic or connect to and import all the data in your database, we advocate a one-at-a-time approach. Choose something you want to measure, quickly set it up in StatsMix, turn on the data spigot, then let it run for a while. Once you start viewing the results, you may think of new things you want to track, or metadata you might want to attach to the baseline metric.
Or not. Sometimes to see the forest, you only need to look at one tree. Just make sure you choose the right tree and follow through.