Decision models¶
The Decision screen turns CRM data into decision support: what to do next, scenario simulation, who your best customers are and how much you will bill. It covers UFCD 10869 - CRM, decision analysis models.
Next Best Action¶
A list of the highest-impact actions, per account, generated by rules from the real state:
- Resolve overdue case - the company has a case with a breached SLA.
- Contact hot lead - high lead score (≥70) still in Lead/MQL.
- Create proposal - qualified opportunity (SQL) with no open proposal.
- Follow up - proposal sent over 14 days ago without closing.
- Reactivate customer - no new purchases in over 90 days.
Click Open to go to the company record and act.
What-If simulator¶
Drag the sliders to see, in real time, the impact on the business:
- Price change (± on the average ticket)
- Conversion-rate change
- Number-of-leads change
The simulator shows projected deals, the resulting conversion and the projected revenue, compared with the current base scenario. It answers questions like "what if I cut the price by 10% but bring in 20% more leads?".
Sales forecast¶
The open pipeline is weighted by the typical closing probability of each stage:
| Stage | Probability |
|---|---|
| Created | 20% |
| Sent | 40% |
| Negotiation | 70% |
The expected revenue is the weighted sum - a more realistic forecast than adding up the whole pipeline.
RFM segmentation¶
Classifies customers along three axes:
- Recency - how long ago they bought.
- Frequency - how many times they bought.
- Monetary - how much they spent.
Customers fall into segments - Champions, Loyal, Developing, At risk, Dormant - that guide action: keep the Champions loyal, reactivate those At risk.
Pedagogical models
The probabilities, thresholds and rules are simplified for training. In a real CRM they would be calibrated with history and, often, machine learning.