This article provides an overview of the Data Science Scores in Renewal Center.
Gainsight calculates the Data Science scores in Renewal Center using a proprietary framework called Random Forest-Based Model. This is both flexible and extendable, and captures disparate aspects of the customer’s health.
The major advantage of the framework is that it evaluates a larger set of variables such as surveys sent, open and closed CTAs, surveys response rate, and so on, to capture the 360-degree aspects of customer data. This data is then used to determine factors that could lead to churn.
The Random Forest-Based Model offers the following advantages:
- More Accurate Scores: This model captures complex interactions within your data to generate a more accurate DS Likelihood score.
Note: It requires more than 50 opportunity records.
- Automatic Variable Selection: From the numerous variables available, this model automatically selects the variables that are important for predicting the probability of closing opportunities successfully.
- Automatically Tuned: The model is adjusted automatically based on cross-validated datasets to improve the accuracy of the model.
- Enhanced Feature Generation Framework: Gainsight uses this model to leverage the significance of individual variables, and their connection to other variables to highlight the most relevant and actionable reasons.
This framework creates multiple models for the tenant to calculate the renewal likelihood of every opportunity. The number of models created by the framework depends on the tenant configuration, as follows:
- Unique models are created for all scorecards in the tenant. The framework evaluates all scorecards associated with the company to which the opportunity belongs.
- Unique models are created for each opportunity type, namely upsell, downsell, and renewal opportunities.
- Unique models are created for a renewal opportunity based on its prediction period. This framework considers the prediction period of four months or less and more than four months, while creating models.
Data Science (DS) Scores
The following two scores are calculated by Data Science in Renewal Center:
DS Likelihood Score
This field generates a score based on the success or failure of past renewals (including renewals with upsell and downsell). The score is updated on a monthly basis by learning the pattern of past closed renewals. Gainsight’s proprietary framework is used to perform the calculations, using the following elements:
- Health: Captures the health of a customer as determined by the measures configured in your Scorecard.
- CS Engagement: Captures the effect of customer success activities you implemented for the account. The variables involved in the calculation are:
- Number of surveys sent to the customers in the last six months.
- Number of Call to Actions (CTAs) closed in the last six months.
- Number of open CTAs.
- Number of Success Plans closed in the last one year.
- Number of Timeline activities posted in the last six months.
- Number of calls or meetings with the customer in the last six months.
- Days since the last timeline activity was posted with the customer.
- Days since the last call or meeting happened with the customer.
- Average NPS in the last six months.
- NPS survey response rate.
- All surveys response rate.
- Number of risk CTAs closed in the last six months.
- Number of risk CTAs opened in the last six months.
- Average of overall health score in the last one month.
- Number of outreaches sent in the last six months.
View Factors Contributing to the DS Likelihood Score
Renewal Center displays the various factors (Gainsight features) which contribute to the DS Likelihood Score. You get insights about what exactly is driving the score and also understand how much the Renewal Manager should rely on the score by viewing the data source that supports the score. You get information about the aspects of customer health, engagement, and other areas which are impacting the score.
To view the factors which contributed to the renewal likelihood score of an opportunity, hover your mouse on the required score and view the following:
- Gainsight features contributing to the score.
- Level of impact the feature has on the score. (high, medium, low)
- Nature of impact the feature has on the score. (positive, negative)
- The features contributing to the score and the level of impact of a feature are auto calculated by Data Science. You cannot modify these areas.
- The factors are positioned based on the score. If the score is green, then more than half of the factors are green and listed first. If the score is yellow, then factors from each score color are displayed.
DS Forecast Amount
This amount is calculated based on the historical average Net Retention Revenue (NRR) in each DS likelihood bin. Data Science takes into consideration the overall final amount or target amount (NRR) in each of the red, yellow, and green bins historically. The amount in each bin is then multiplied with the same ratio in a quarter.
|DS Forecast = (Red Target Amount * Historical Red NRR) + (Yellow Target Amount * Historical Yellow NRR) + (Green Target Amount * Historical Green NRR)|
The reasons framework shows the most relevant and comprehensive set of reasons for the scores. The following logic helps in achieving the results:
- A maximum of five contributing factors are shown.
- The factors identified by the framework are given priority in the following order:
Indicator Variables (such as NPS, CSAT) > Scorecard Measures > Engagement Variables (such as Surveys, CTAs)
- Not more than three scorecard measures or measure groups are shown as contributing factors.