How to Do RFM Analysis for SaaS

To create a profile of a customer using Recency, Frequency, and Monetary values for a SaaS application, the accounting department should be equipped with information on the recent activities of the account. Many SaaS companies engage in the process of identifying churn factors through behavior analysis and subsequently adjusting contract value according to usage patterns. 

This section will take you through the steps of building RFM scores, or more precisely, how to perform a segmentation and conversion process for at-risk accounts, and how to increase the number of high-value customers worldwide.

When giving this consideration, the strength of your team’s engineering and the availability of useful data also need to be factored in. You can get a reflection of your readiness by answering these four basic questions: Do you have clean transaction records from all the payment gateways? Is your product analytics system capable of tracking user events? Is your primary objective just to reduce churn or go to new markets?  Is your team capable of implementing the manual SQL scripts or does it need a trigger?

The availability of technical resources influences the readiness for a real-time data pipeline. It is better to start with simple 3-tier spreadsheet segmentation on core accounts. On the other hand, if the system includes more than 1 million micro-transactions performed in different parts of the world, it is necessary to use scoring based on an API to keep the accuracy of the segments in place. 

Thus, it is important to understand the limitations of the resources available for retention activities and to direct them only towards the retention of the revenue that is at stake in the immediate future. This way, you can avoid investing too much time and resources in activities that might not bring in immediate revenue.

To create a database of customer activities, the information available from payment gateways, CRMs and product analytics tools should be pulled. The presence of incomplete information, duplicate records, test accounts, or unverified trial profiles within SaaS data may affect the precision of associated values.

Thus, it is necessary to clean the data to eliminate such practices and concentrate on the actual customer behavior that is generated by live accounts. However, before running any scoring scripts, the data are to be formatted consistently across all the tools in which they are integrated. Otherwise, it should be set up as a periodic process to avoid slowing down the system and to keep the reports accurate.

Data Collection Checklist

• Obtain product analysis logs (Mixpanel, Heap, or Amplitude) for account logins and feature execution events.  
• Export payment records from the subscription billing process with the help of transaction dates, gross amounts, and currencies. 
• Eliminate internal accounts, staging environments, and free trial accounts using a filter. 
• Add a filter for external user IDs in CRM tools (HubSpot, Salesforce) and payment gateways to create one single account.

Isolating Recency and Frequency Metrics

In order to identify specific customer transaction dates (Recency) and renewal patterns (Frequency): 

• Recency Calculation: Keep an eye on the exact time of the last action with high relevance—for example, a successful billing event, or login, or core process execution—rather than just passive browsing. Take away this date from the total number of days in the analysis and get the exact number of days.

The formula is: R = T(reference) – T(last_purchase)

• Frequency Calculation: Consider high value activities as a product of a single period, for example, 90 days for user activity or 12 to 36 months for enterprise contract renewals in the case of subscription software.  For subscription software, count active seats or expansion conversations as parts of the period as a measure of real growth in the relationship.

To create an accurate “Monetary” score for the accounts based on the total amount of checkouts, including variable regional sales taxes (VAT/GST), currency conversion issues, and payment processing fees requires careful methodology. Comparisons of accounts using gross revenue, without adjusting for regional taxes, may result in differing assessments of value, given that $100 in Europe and $100 in the US can exhibit distinct profitability levels due to VAT. Foreign exchange rates have an impact on the total spending figures observed over the subscription period. Removing the fees generated by the platform gives way to the understanding of what the account is really making.

Consolidating all transactions into one currency for reporting purposes relates to the framework for positioning. 

1. Taxes and Fees: Locate any taxes, goods and services tax (GST), or payment processor fees attached to the gross transaction amount. 
2. Convert to Base Currency: Apply the foreign exchange (FX) rate taken from the currency converter at the time of the transaction to convert the local payment into the reporting currency (USD in this case).
3. Calculate Net Contribution: Find the net revenue derived from an account using the method of taking the average rate from all the billing periods.

Revenue Normalization Comparison

In order to convert the raw measurements of Recency, Frequency and Monetary values into a score, you should allocate scores between 1 and 5 for each of the components. Dividing the customer base into quintiles gives an opportunity to put 20% of the customers in bucket 5 and 20% in bucket 1 according to the performance rating. Obtaining such an aggregate result gives an RFM code that contains information on the account’s current stage in the lifecycle. With these codes, marketing campaigns can be formulated for specific recipient groups, differentiating from non-specific email communications.

Every 6 months, it is necessary to prepare such boundaries again to align the segments with the changing patterns of product usage.

RFM Segment Matrix:

In order to perform RFM analysis in SaaS applications or traditional databases, you have to prepare SQL queries to divide accounts into these predefined percentile ranks.

If a platform uses RFM segmentation for accounts and integrates it with billing policies, it may influence the management of dunning periods and the application of proactive discounts, which could relate to account retention. Accounts where Recency scores are falling can be associated with payment challenges, with expired cards or modified card management practices as potential influences. If retention strategies include changing the payment retry schedule, users will have more time to resolve the issue and avoid cutting off the service immediately. 

Locating an option to enable payment methods specific to the region where the user is located or changing the subscription period temporarily can help in keeping accounts that are going through a process of redefining the budget. The direct connection of these triggers to the billing engine affects a platform’s revenue considerations and may influence the extent of manual selling required.

When an account reaches the “At-Risk” stage (e.g., score 155 or 244):

1. Modify Dunning Sequences: Adjusting the dunning sequence to extend the retention attempt period from 3 days to 7 days provides account owners with a longer timeframe for updating expired card details before service termination. 
2. Apply In-App Credits or Discounts: Activate an administrative process using payment infrastructure tools to offer a 15% discount off the subscription or credit card payment with no notice that a payment has failed or reduced login activity.  
3. Route High-Value Alerts: Notify selected customers and success managers of an account with a Monetary score of 5 entering a low Recency tier within less than 10 minutes. 

The process of identifying users with the highest RFM score (“Champions“) and utilizing existing features to facilitate expansion into new territories or an increase in license numbers requires that you:

1. Filter for Top Scores: Get your CRM or analytics tool and find accounts that have a 555 or 554 score in all the territories. 
2. Enable Localized Upgrades: Set the checkout processes to show local currencies on payment pages, preferred regional payment methods (for example, SEPA in Europe, iDEAL in the Netherlands, Alipay in APAC), and remove the taxes automatically if these accounts see up-selling opportunities. 
3. Prompt Expansion: Show case to individual accounts offers for more seats or enterprise editions based on the location and the usage patterns.