RFM analysis

RFM analysis is a one-off analytical project (can be made regular): based on purchase history (CRM, billing, orders) we compute Recency/Frequency/Monetary per customer, split the base into RFM segments (e.g. champions, loyal, at risk of churn, dormant, newcomers) and deliver the segments with descriptions and recommendations on how to work with each. Honestly about the nature, this is key: RFM is DESCRIPTIVE segmentation by past behavior (what the customer did), NOT a forecast of the future and NOT a proven cause of behavior; it is a snapshot at the calculation date — over time customers move between segments, so the analysis is worth refreshing. Honestly about data: correct purchase history tied to the customer is needed — without it or with duplicates the segments will be wrong (garbage in, garbage out); segment thresholds are a choice we agree on, there are no 'single correct' boundaries. Honestly about the essence: this is segmentation and recommendations, NOT mailings, not CRM setup and NOT sales growth by itself — what to send to whom and how to work with the segments is decided and implemented by you or your marketer (implementation — separately); the fact of segmentation does not raise revenue. Requirements: access to purchase history/CRM. An important boundary: this is RFM analysis, not predictive models (churn/LTV — separate), not full segmentation across all data and not campaign implementation. If there are few customers or one-off purchases, RFM is uninformative — we say so honestly. Picture this: instead of 'we send everyone the same mailing' you see who is a champion (protect), who is at risk (win back), who is dormant (reactivate) — and work in a targeted way. The base price starts from 15,000 ₽ per analysis; it depends on the base size and data state.

• You work with the whole base the same way, without segments.
• You do not know who your best, dormant and churning customers are.
• Communications and offers are not targeted.
• There is no basis for prioritizing customer work.

• Preparing purchase history (CRM, billing, orders)
• Computing Recency, Frequency, Monetary per customer
• Agreeing on thresholds and the number of segments
• Splitting the base into RFM segments with descriptions
• Recommendations on working with each segment
• Flagging problems/duplicates in data
• A report and review with you
• Exporting segments for further work

• You understand who the best, dormant and churning customers are
• Targeted segments instead of 'the same for everyone'
• Priorities for working with the base
• A base for campaigns (implementation and growth — separately)

• We clarify data, thresholds, needed segments; collect access to purchase history
• We compute RFM, form segments, prepare recommendations
• We compile a report and export, review with you

• Will RFM analysis raise sales?

  No by itself. It gives targeted segments and recommendations, but mailings, offers and segment work are implemented by you or your marketer (separately). Segmentation itself does not raise revenue — it is a base for targeted work.

• Does RFM predict who will leave or how much they will bring?

  No. RFM is descriptive segmentation by past behavior, a snapshot at a date, not a forecast. For predictions you need predictive models (churn, LTV) — separate services. Over time customers change segments, the analysis is worth refreshing.

• My base is small — is it worth it?

  With few customers or one-off purchases RFM is uninformative, and duplicates/dirty data will distort the segments (garbage in, garbage out). If data is scarce, we honestly say it is too early and what to prepare.