The Pricing Committee Pack - Pricing Impact Profile

Rika Taute
Aug 13
3 min read

Edition 15 used the Differential Lift Chart to show where the proposed model changes the pricing signal, whether observed experience supports that change, and how much exposure sits in each decile.

This edition takes the next committee step: once the lift chart shows where the model wants to move price, the pack should show who receives the largest implied total increases, what drives those increases, and how much policy and premium volume is affected.

The Decision Question

Which affected profiles receive the largest implied total rate movements, what factors contribute to the movement, and is the action credible, material and transition-safe?

This matters because a lift chart is still primarily a model-performance view. It tells us where the proposed model disagrees with the current pricing signal. This edition looks at the next question: who is affected, how much premium is involved, which factors explain the movement, and what the committee should do about it.

Default Practice

After a lift view, pricing packs often move to a rate dislocation or policyholder impact table. That table usually bands policies by proposed rate change and shows policy count, premium, average movement and perhaps the impact of caps.

This is useful and widely recognisable inside pricing implementation work.

Suggested Upgrade:

Pricing Action Impact Profile

The proposed upgrade is a stacked implied total increase profile. Rank or band policies by their implied total rate movement. For each band, show the total as a stacked bar of contributing factors, then place A/E, policy count, premium and action next to the same row.

The segment profile should describe who is concentrated in the band - for example, male, married, age 45-55 - but it should be treated as an impact and governance description, not as a standalone pricing justification.

Committee lens	Question to answer
Rate movement	What is the implied total increase or decrease for each band?
Drivers	Which components create the movement: loss-cost adequacy, age, SP profile, channel/territory or interaction?
Impact profile	Which SP group is concentrated in the band, and how many policies and premium are affected?
Evidence	Does the A/E or actual experience support the movement?
Action	Does the band need a cap, phase-in, review, monitoring rule or anti-selection check?

How To Read The Chart

Read the rows from the highest increase down. The full length of the stacked bar shows the total proposed pricing movement, while the colored components show what is driving that movement across age, gender or household, vehicle or risk characteristics, territory and claims history.

Then read the segment profile alongside the bar. This shows which customer and risk characteristics are concentrated in each impact band, linking the pricing movement back to the groups affected.

Finally, read across to the policy count. This adds materiality: a large increase affecting a small segment may require a different response from a more moderate increase affecting a much larger part of the portfolio.

What This Immediately Reveals

· Which customer profiles sit behind the highest implied increases.

· Whether the increase is mostly actuarial adequacy, a specific factor effect, or a broader interaction.

· Whether the affected band is material by policy count and premium, not just by percentage movement.

· Which bands need caps, phase-ins, communications, underwriting review or post-implementation monitoring.

Trade-offs & Risks

The chart is more decision-useful than a plain dislocation table, but it is also easier to misuse. The stack can look like a precise causal decomposition even when some components are judgemental allocations or correlated effects.

The segment also needs care. If it contains sensitive or regulator-relevant characteristics, the profile should be used for impact assessment, fairness/proxy review and transition planning. It should not be presented as the sole reason for the rate action.

Decision Boundary

Use this chart when...	Don’t use this chart when...
Showing which profiles receive the largest implied total rate increases or decreases.	Assessing the predictive performance of the model in isolation.
Explaining the main contributors to the rate movement, such as loss-cost adequacy, age band, profile effect, channel, territory or interaction effects.	Setting the overall portfolio rate level or replacing the executive selected-rate-action view.
The committee needs both the size of the rate movement and the amount of policy and premium volume affected.	Justifying a rating factor solely because a profile appears in a high-increase band.
Turning a lift or dislocation view into a transition decision: phase-in, review, monitor, accept or check anti-selection.	Using sensitive or regulator-relevant profile variables without the agreed fairness, proxy or conduct review.
The profile labels are available on a consistent, governance-approved basis and can be used for impact review.	The contributing components are not calculated on a comparable basis, or the split is too judgmental to present as a driver stack.