Ideas, concepts, solutions, news
Becoming a Data Driven Customer Success Organization
Google, Facebook and Amazon are all running on data. That’s their fuel. Any plan is translated into a set of metrics and targets, which are continuously measured. Clearly data driven is the approach that every modern company should adopt.
Is it the case in your B2B company?
A recent Harvard Business Review study finds that virtually all respondents say their firms are trying to move in that direction, but only about one-third have succeeded at this objective.
Most companies in the Customer Success space are already collecting large number of data signals and surfacing them to the right people. But not that many are taking it to the next level. Identifying which data signals are significant, setting measurable targets and ultimately driving their team behavior.
Becoming data driven org does not need to be a long and costly journey. You can start by investing a couple of hours (yes hours not days or weeks) and in a couple of weeks you’ll start seeing results.
There are three simple steps:
- Define measurable metrics
- Identify the most significant data signals and set targets
- Get your team to use it
What is success for you?
The foundation of any data driven organization are measurable metrics. It also allows machine learning algorithms to correlate data signals with the desired results.
Customer Success organizations should be measured by % retention and $ expansion. Additional measurable success metric, which is a little harder to attribute, is sales referrals (e.g. NPS).
Measure your Customer Success by % retention and $ expansion
Pacific Crest published info about retention and renewal metrics used by public SaaS companies. Download PDF. Most of them bring retention and expansion into one metric based on $ amounts. Some are using # of customers as the basis for the metric. As an example, one of our customers AppFolio is using the following metric “…dividing (i) revenue generated from the sale of our solutions in the given fiscal year from our base customers by (ii) revenue generated from the sale of our core solutions in the base year from our base customers”
Most companies are already collecting data signals so they can better monitor their customers’ behavior (digital body language). For customer success most SaaS companies are collecting product usage data that represents users engagement with their service, customer service cases (how many, time to close, etc) as well as different types of surveys including NPS (Net Promoter Survey).
Most companies do not know the significance of each data signal and what is the desired target.
A data driven solution will use the success metric (% retention and $ expansion) to figure which data signals have high impact and what is the right target for each one. It is based on a Machine Learning model that is trained using historic data (renewal and expansion opportunities and their final won/lost stage, relevant data signals, etc).
Advanced companies are using their own data science team or hire consultants to get this information. The only known out-of-the-cloud solution that addresses this is Komiko.
Here is an example of the way that Komiko can rank each data signal and set the target. It is based on win/loss history as well as account segmentation.
Listening is important. It allows you to identify risks and opportunities and make them transparent to your leadership team.
Just listening to data signals is not much different than checking the weather and dressing accordingly. You want to do more!
As a customer success leader you want to influence retention and expansion. The way to do it is by engaging with the right people, at the right time and the right content. That’s were the Customer Success team can make a difference.
Engagement plan should address three aspects:
- who (champions, sponsors, end users)
- how often and which media (email, text, phone)
- and what content
The engagement plan should focus on addressing significant data signals. Each one of them may be negatively out of target indicating renewal risk or positively outside indicating expansion opportunity.
SURPRISE! Engaging with your customers creates a new set of data signals that are mostly overlooked.
You can apply the same approach used to identify the most significant data signals and identify the most important engagement types. Here is an example.
It is pretty hard to influence product usage signals, but it is relatively easy to change people’s engagement pattern. These changes can show results in a matter of days.
When applying people engagement data signals you can reach a pretty high accuracy. Komiko managed to reach 95% accuracy.
Here are some more details. The precision/recall curve shows that when using only product usage data the precision achieved was ~80% and when adding people engagement (emails, meetings, people seniority, response time) it went up to 95%.
The last step. So you now have a process that can help you identify pretty accurately the likelihood to churn and expand accounts. And it also indicating which signals you need to pay attention to. You just need to be able to communicate them effectively to your team so they can quickly take action. Here are few examples that should be surfaced in your team’s working environment.
PS Earlier this week, I was asked by Boaz Arbel VP Customer Success at logz.io and Irit Eizips CEO at CSM Practice to share some AI concepts with a forum of Israeli VP CS & CCOs. This blog follows the presentation.