Friday, October 17, 2008

Bessemer SaaS Law #9

Bessemer SaaS Law #9. The most important part of Software-as-a-Service isn’t “Software” it’s “Service”!

One of the most valuable and least appreciated assets of a SaaS business is the detailed usage data of your customers. For years, product marketing and product management groups in license software businesses have attempted to guess at the behavior of their customers. Good PM’s would interview, survey, and even watch their customers as they used the product, but it was very hard to truly know how they were using the product on a detailed level and incorporate this feedback in the major annual releases. SaaS businesses should instead learn from their consumer internet peers, and the user interaction methodologies of businesses like Google, Apple, Yahoo, LinkedIn, Yelp, Facebook, and others. These businesses are constantly taking advantage of their web application architecture to analyze detailed customer usage data, a/b test variations, iterate on small details of a page or a feature, and evolve the product each and every day. There are three levels of being Service Savvy, and SaaS businesses should strive for all three:

1) Basic proactive monitoring for likely churn or up-sell opportunities – this is too simple not to do, and all SaaS businesses should do this well. You already know who logs into your product, how often, what they do inside the product, and what results they achieved. So now you need to track the key usage metrics and measures, and create internal dashboards to know which customers are getting the most value (potential up-sell candidates!) and which are likely to churn (time to proactively intervene!). Work with your marketing team to automate “low usage” reports internally, and low usage escalation emails to your customers to ask them for an explanation of the behavioral changes.

2) Rapid testing and development cycles of your application based on customer feedback. Lookat what your customers are doing in the product, not what they tell you they’re doing. Look at which pages, features, and sub-components are getting the most active usage and why. Look at the outliers and understand whether there exist some interesting additional opportunities.

3) Analyze the aggregated data to determine best practices and benchmarks across your
customers. You can actually add value to your customers businesses by making them better in your product, and your product better by watching your best customers.
The other subtlety of a service offering is that customers have to be able to use the offering in production to recognize its value, and for your business to convert CMRR to MRR and GAAP revenue.

This implementation time is measured as the time from the day you acquire a new order until the day it goes into production. You want this lag to be as short as possible, and growth in this number is a strong leading indicator of possible product and/or implementation problems. Prudent SaaS companies have made wise investments in process and technology to make this time as short as possible. These investments span the range of simple items like streamlining workflows and improving user training to fully automated self-provisioning models that allow turn ups to happen without human intervention. As a general rule, the higher your price point the lower the need for automated provisioning and the lower your price point the higher the need. A fairly typical implementation range for SMB focused SaaS providers is 0-60 days, and enterprise focused offerings typically range from 2-4 months. Although “average implementation time” is typically not a key executive metric, it should be tracked as a second level metric and promoted as a primary metric if you are out of range or the trends are inappropriate.

Full Index

No comments: