The 3-Step Revenue Growth Framework

CUREᵀᴹ(Conversion, Up-Sell, Retention and Engagement) Analytics

As a Revenue Growth leader of a subscription business, you know that to grow your customer base and scale your revenue, you need to learn from your data and put those insights to action faster.

Below is a framework on how to think about the process you want to put in place for a holistic strategy.

1. What mitigation activities do you have at your disposal?
2. What data is required?
3. How can you continuously learn from the activities?

Let us look at each of these elements in some more detail

MITIGATION STRATEGIES

The first thing to consider is the type of activity that will be most useful. Typically, these fall into the four types — Promotion, Incentives, Notification & Product. (Courtesy: REFORGE)

The second thing to consider is the channel of communication with your customer which can be — Email, In-App, Text Message, Phone Call, Website (or Social Media Property).

It is important to note that the mitigation activities carried out for converting, up-selling and retaining and engaging the users may be carried out by different departments within the organization. For example:

• Phone calls to retain key B2B accounts may lie with Customer Success Teams
• Email Promotions and Incentives may lie with Growth Marketing
• Up-Sell based email / text notifications may lie with Sales
• Website may lie with Marketing

These different activities may also use different tools such as Salesforce Marketing Cloud, GainSight for Customer Success.

The one thing that is common between these different mitigation activities is identifying the right users to target with these activities. In turn, identifying the right users is really about utilizing ALL the right data and NOT using the noisy data.

We will talk about the data later but you need to first think about different campaigns and categorize them by cost of execution & value given away (think discounts).

Below is an example of a food delivery service (by extension this can be applied to almost any subscription or online company) and should help articulating the point.

The five categories of mitigation activities are:

• Very Light Grey — Lowest Cost — Website
• Light Grey — Low cost — Email delivery or Social Media
• Medium Grey — Medium cost — In-App messages
• Dark Grey — Expensive — Phone calls
• Very Dark Grey — Most Expensive — Product

So the above matrix can be used, as described below:

• For website you want the highest coverage with the least amount of precision
• For email you want high coverage of users (also called high recall) and a lower precision can be okay
• For In-app messages you want a moderate precision and moderate coverage
• For more expensive methods such as phone calls you want highly precise list of users who are going to churn (also called precision)
• For the most expensive methods such as product changes you want to be very precise to enable it for folks that would find it useful

The next step is to target these customer on multiple channels. Research shows that users that are targeted at least 3 times are twice as likely to take notice and act on the promotion. These omni-channel flows can have multiple permutations esp. if you the include the timing of delivering the activities. Fortunately, multiple marketing automation tools exist that help you in orchestrating the journeys (will not be describing there here).

DATA

With a subscription or online business the strategies that you can utilize are very interesting because you have lots of product usage data (trial and paid) at your disposal.

There are five categories of data that you need to look at (in this order):

• Product Usage
• Demographic
• Transaction
• Marketing Outreach and corresponding user actions
• Third party data

Product usage data is a critical component as usage data is a leading indicator of churn. Oftentimes, this is the hardest to retrieve as there is no standard way to track all the events of interest. Key elements of a product usage data collection strategies are:

• Hypothesis free data collection: Ensuring all the right events are collected not only those that are required for a particular outcome. You don’t want to miss out on collecting events that you aren’t aware are impacting the outcome.
• Avoid Messy Data: Create a consistent taxonomy and single source of truth for event definitions.
• Prevent Data Quality Issues: Invest in tools/scripts to catch analytics bugs before they happen with automated reports and alerts.
• Ensure Data Privacy: Complete visibility and control on tracked data so as not to send PII data from your application.

The Transaction, Demographic and Marketing Action data is relatively easier to ingest because it resides in internal Data Warehouses. One note of caution is to not be overwhelmed with a deluge of data and tables. Having a clear focus on why you need the data will prevent the data deluge and will get your ALL the important data and NOT the noisy data.

There are many third party companies that provide the data and are easy to integrate and implement.

BUILDING A LEARNING SYSTEM

Building a learning system requires one to create a full loop integration.

Getting entangled with stacks of analytical tools that don’t play well together, stops you from understanding which users are most likely to respond to the right mitigation strategies, that we have defined above.

Elements of a Full-loop Integrated System are:

Data Ingest

It is essential to integrate data from several data pipelines into a single canonical format before you run analytics on it as shown below.

Data Processing

Data Processing pipelines need to incorporate several ML models that can be executed in parallel or otherwise and then be able to create an ensemble for extracting the most out of the data. Most ML Data Science teams should be able to create the right one for your business.

Integrate Results with a Marketing Automation System

Tools such as Salesforce are used by Sales and Marketing teams and allow you to run omni-channel sales and marketing actions. CS teams use tools such as GainSight that allow you to work on outbound messaging workflows.

User actions for the mitigation activity needs to be integrated into the Data Warehouse which flows back into the Data ingest pipeline, above.

CONCLUSION

Innovative companies pay attention to the entire customer journey to identify areas of growth. They focus their attention across the continuum from Conversion, Up-sell, Retention to Engagement, what we at GrowthSimple call CUREᵀᴹ Analytics.

In summary, there are two main take-aways from working with several customers.

• First, it is essential to identify and start with ALL the right data, (without it the ML model predictions will result in underwhelming growth outcomes) and remove NOT the noisy data.
• Second, for compounding growth the closer you are to having a full loop integration in place (that allows for continuous learning) the better.

Once these things are in place, the following questions then become easy to answer.

• Who is converting and why?
• Who is the right customer to up-sell?
• Who is most likely to retain/churn?
• Who is most engaged and why?

GrowthSimple, is a Smart Scoring startup, based in the San Francisco Bay Area and helps subscription companies identify the right users to target based on product usage, transaction and other key data sets. We also provide full loop integration services.

NEXT STEPS

We work with 4 SaaS companies each month to provide CUREᵀᴹ (Conversion, Up-Sell, Retention and Engagement) data diagnosis at a nominal cost.

Our Data Science team will analyze the data for each of these companies based on our comprehensive proprietary framework to identify the most important challenge that the company has:

• Conversion
• First 30/60/90 day Churn or Ongoing 30 day Churn
• Up-Sell &
• Engagement levels

Please get in touch if this is of interest by sending me an email at saksena@growthsimple.ai