In our engagements with valuation teams using Prophet, we’ve observed that most users are confident with in-force projections—such as those required for monthly valuation or embedded value reporting. However, when it comes to new business (“NB”) processing, many find the setup and logic less intuitive.
In this article, we aim to clarify how new business processing works in Prophet, highlight the differences between the cross multiplication and project methods, and explain why the project method is often the preferred approach—particularly for Family Takaful operators.
What Is New Business Processing?
New business processing in Prophet is the mechanism for projecting future sales. It uses monthly sales volumes provided in a Sales File to generate projected cash flows. While Prophet supports three NB processing methods, in practice, the most widely used are:
- Cross Multiplication Method
- Project Method
To understand the mechanics, it’s helpful to break the process down into two core components: the NB Profile and the Sales File.
A. Overall Concept
At its core, new business projection follows this logic:
NB Profile / NB Premium × Sales File = Projected Cash FlowsHere, the NB profile refers to a per-dollar projection of cash flows. Prophet multiplies this by the monthly sales volumes to estimate aggregate cash flows. Let’s walk through a simplified example:
- The NB profile estimates premium income of $100,000 in policy month 1 and $95,000 in month 2.
- The profile is based on a model point with an annualized premium of $1,200,000, defined using the PREM_AMOUNT variable (a Prophet-specified field that must be spelled exactly).
- If expected new business sales in January 2025 is $600,000, then projected premium for February 2025 is:
- If February 2025 sales is $900,000, premium income in the same month (for new policies) is:
To determine total expected premium for February 2025, we combine:
- $47,500 from February (month 2 of January business)
- $75,000 from February (month 1 of February business)
- Total = $122,500
B. The Role of NB Profile
The NB profile is derived from NB model points, which Prophet identifies based on their SPCODE values. Any SPCODE greater than the Last Sub-product Code for Existing Business—as defined in the Structures—is considered a new business point. Additionally, ensure that DUR_M = 0 for these NB points.
For example, if the last SPCODE for existing business is set at 50, then all points with SPCODEs 51 and above are treated as NB.
A common question we hear is: “Should we use more NB model points for better results?”
In reality, NB model points represent the distribution of projected new business. The most common practice is to use actual recent NB policies to reflect this distribution. However, you can also construct it manually using fewer, representative points.
In principle, one well-constructed model point can yield the same NB profile as 100,000 homogenous points—what matters is how accurately the distribution represents expected business.
C. Understanding the Sales File
The Sales File provides the monthly new business volumes. It is structured by product code and SPCODE. For example, if the NB model points use SPCODEs 51 and 52, your Sales File should include columns labeled:
- PRODUCT#51
- PRODUCT#52
Each column captures the sales amount for that product/SPCODE combination, by month.
Cross Multiplication vs. Project Method
The core distinction between the two methods lies in how the NB profile is generated and applied:
Cross Multiplication Method
- The NB profile is created once using a fixed start date.
- This single profile is applied across all months in the projection using the monthly sales values.
Project Method
- Prophet regenerates the NB profile each month, based on that month’s starting date.
- The system performs a new projection for every month’s sales entry.
- Think of the project method as running a series of cross multiplications, each tied to a different start month.
Why Is This Important?
The need to regenerate the NB profile monthly depends on whether your model assumptions are calendar-specific. If your assumptions are purely policy-based (e.g., benefits occur one year after issue), then the cross multiplication method is likely sufficient.
However, in Family Takaful, timing often matters. Surplus distribution and profit sharing, for instance, are commonly aligned with calendar milestones—such as year-end on 31 December. If you rely on the cross multiplication method, all policies share the same profile regardless of their issue date. Prophet would inaccurately apply surplus sharing as if it happens monthly.
To reflect calendar-specific events accurately, the project method is essential. It ensures that each policy has a properly aligned projection timeline, based on when it was sold.
Addressing Performance Concerns
There’s no denying it: the project method is slower. For a 3-year projection with 36 monthly sales records, Prophet must generate and project 36 unique NB profiles. As a rough estimate, the process can take 36 times longer than the cross multiplication method.
That said, performance can be managed effectively. We recommend using grouped model points, which aggregate similar profiles. This reduces data volume while preserving accuracy in the projection patterns.
Conclusion
Choosing between the cross multiplication and project methods is not merely a question of performance—it’s about choosing the right approach to reflect your product’s underlying business logic.
For Family Takaful models, where profit and surplus sharing are typically based on fixed calendar events, the project method offers superior accuracy. While it comes with a runtime trade-off, the impact can be mitigated through efficient model point grouping.
Ultimately, adopting the appropriate NB processing method ensures your Prophet projections are both technically sound and commercially relevant—a key priority in any actuarial modeling exercise.
