4 Levers That Drive Fleet Insurance Pricing Models in South Africa

September 2026

Fleet insurers set premiums using two broad approaches: traditional fleet rating, split into composite and schedule methods, and usage-based models such as pay-as-you-drive, pay-how-you-drive and manage-how-you-drive. Whichever model your insurer applies, four levers move the number most: total mileage, driver behaviour, claims history and the locations where your vehicles operate.

TL;DR:

  • Usage-based models like pay-as-you-drive and pay-how-you-drive are more accurate for fleets with variable or seasonal operations, but less effective for new or small fleets with limited data.
  • Telematics can improve premium predictions when enough driving data is available but has limited impact on thinly populated fleets or those with infrequent usage.
  • Insurers prefer composite rating for small, uniform fleets and schedule rating for mixed fleets with diverse vehicle types and risk profiles, while usage-based models suit variable operational patterns.
  • When comparing quotes, fleet managers should clarify whether premiums are based on composite, schedule, or usage-based models, and how telematics data influences pricing adjustments.
  • Regulatory transparency and vehicle valuation updates are essential to avoid underinsurance and ensure accurate, fair premiums across different market conditions.
Truckplant
Align Cover With Fleet Operations
Truck & Plant On-Demand™ lets commercial vehicle and plant operators choose what to insure, when to insure it, and how.

Explore on-demand insurance

Table of Contents

How insurers calculate fleet premiums: the actuarial building blocks

Every premium starts with expected loss: how often claims happen (frequency) multiplied by how much they cost when they do (severity). Insurers estimate both figures for your fleet profile, add administration and margin, then arrive at a base premium.

To get there, actuaries measure exposure variables that predict frequency and severity. These include annual mileage, vehicle class, the type of load carried, the mix of drivers behind the wheel, and the territory where vehicles are based and driven. A refrigerated truck running long-haul routes through flood-prone regions carries different exposure than a light delivery van working a single city.

Fleets are then rated using one of two structures:

  • Composite rating treats the fleet as one unit, applying a single rate per value band across all vehicles, which suits fleets with similar vehicle types and use.
  • Schedule rating prices each vehicle individually based on its own risk factors, which suits larger or mixed fleets where vehicles differ significantly in type, age or use.

Past claims then adjust future pricing through credibility mechanisms, sometimes called bonus-malus systems, where a clean claims history earns a discount and a poor one adds a loading. Multivariate credibility models let actuaries blend historical claims data with newer signals, including telematics inputs, to produce posterior premium adjustments rather than relying on claims history alone. Vehicle valuations sit underneath all of this, and the SAIA Code of Conduct requires insurers to reassess base values regularly so that insured values keep pace with market conditions.

Operational risk factors, like the type of goods hauled or the routes driven, also feed into this exposure picture, which is covered in more depth for fleet operators here.

Common pricing models: composite, schedule and usage-based approaches

Fleet insurers rarely use one pricing method for every client. The model chosen usually reflects fleet size, uniformity and how much data is available on individual vehicles and drivers.

  1. Composite rating applies one blended rate across the fleet, which keeps administration simple and works well for smaller, uniform fleets where individual vehicle tracking would add cost without much pricing benefit.
  2. Schedule rating prices each vehicle on its own risk characteristics, which matters when a fleet mixes heavy trucks, light vans and specialized equipment, since a single blended rate would overcharge low-risk vehicles and undercharge high-risk ones.
  3. Pay-as-you-drive (PAYD) prices exposure by measured mileage rather than an estimated annual figure, which benefits fleets with seasonal or variable usage that would otherwise pay for road time they never use.
  4. Pay-how-you-drive (PHYD) and manage-how-you-drive (MHYD) go a step further, scoring actual driving behaviour, harsh braking, speeding, time of day, and feeding that into coaching programs and premium adjustments that reward improvement over time.

Fleets running consistent, predictable routes tend to do well under composite or schedule rating. Operations with variable hours, seasonal contracts or a mix of owned and hired vehicles often see more accurate pricing, and sometimes lower premiums, under usage-based models.

Telematics, machine learning and data-driven pricing: evidence and limits

Telematics devices capture signals that map directly onto the risk factors underwriters already price for: distance travelled, speed patterns, harsh acceleration and braking, time-of-day exposure and route type. Feeding these into a rating model turns generic assumptions about a driver or vehicle into observed behaviour.

Research applying machine learning to South African motor insurance data found that XGBoost models produced lower prediction error than traditional generalised linear models (GLMs) when forecasting premiums. GLMs, however, remain preferable where insurers need to explain pricing decisions to regulators or clients, since their structure is transparent in a way that machine-learning models are not.

Combining GLM and machine-learning approaches can improve prediction accuracy while preserving the interpretability regulators expect, according to comparative modelling research from South Africa. That balance matters more than raw accuracy alone.

Telematics pricing has real limits. Claims severity has a heavy tail, a handful of large losses can outweigh months of clean data, driver behaviour changes once people know they are being monitored, and thinly-populated fleets take longer to build statistically credible profiles. Academic credibility frameworks address this by blending telematics signals with historical claims rather than replacing one with the other, updating estimates as more data accumulates.

  • Telematics can sharpen pricing for fleets with enough driving data.
  • It cannot substitute for a track record on newer or smaller fleets.
  • It works best layered onto, not instead of, traditional rating factors.

More detail on telematics signals and Pay-As-You-Drive mechanics is available for fleets considering a pilot.

Practical steps fleet managers can take to reduce premiums

Lowering fleet premiums comes down to reducing expected loss and proving it to your insurer with evidence, not assumptions.

  • Deploy telematics with a coaching layer, starting with harsh braking, speeding and idle time, since these are the behaviors most linked to claims frequency.
  • Keep vehicle valuations current and maintain accurate driver records and loss runs, since underwriters price off the data you supply.
  • Strengthen physical security, tracking devices, immobilisers, secured yards, and review deductibles and cover limits to match your actual risk appetite.
  • Use dashcam and telematics evidence to support claims quickly, which shortens disputes and improves your recorded loss experience over time.

Pro Tip: Start telematics monitoring at least six months before renewal so you have enough data to negotiate from, not just promise from.

Fleets that document these changes systematically tend to get further with underwriters than those that simply ask for a lower rate. Analytics-led negotiation tactics can help build that case.

Checklist: what to ask insurers when comparing quotes

Comparing fleet quotes only works when you are comparing the same assumptions. Before signing anything, ask each insurer:

  1. Whether pricing is based on composite or schedule rating, and what mileage assumptions sit behind the quote.
  2. How telematics adjustments are calculated and how often premiums are reviewed against driving data.
  3. What vehicle valuation method is used, what the claims service level agreement looks like, and whether named-driver rules apply.
  4. Which exclusions and sub-limits exist, and how goods-in-transit exposure is handled separately from vehicle cover.

Insurers that cannot answer these clearly in writing are worth questioning further before renewal. A structured approach to comparing renewal offers makes the exercise faster.

Predictive analytics is moving from a pricing add-on to a core part of how fleet risk gets assessed. Dynamic pricing frameworks that combine telematics data with behavioral response models can outperform static, once-a-year pricing in simulation, adjusting exposure estimates as driving patterns shift rather than waiting for the next renewal cycle, according to recent optimization research.

These frameworks require careful budget controls, since discounting too aggressively to reward safe driving can erode margin faster than claims experience justifies. The direction of travel, though, is toward pricing that responds to what a fleet is doing now rather than what it did last year.

Machine-learning ratemaking is following a similar path. Rather than replacing GLMs outright, insurers are increasingly layering models like XGBoost on top of traditional rating structures to sharpen accuracy while keeping the transparency regulators and clients expect. Recommended rollout guidance suggests insurers should treat telematics initially as a correction to classical pricing, then shift toward fuller integration once enough driving data accumulates to support it.

For fleet managers, the practical effect is a market where premiums increasingly reward demonstrated behaviour, safe driving, accurate reporting, proactive maintenance, over static assumptions about vehicle type or fleet size alone.

Market trends and innovations in fleet insurance pricing — overview diagram

Comparing fleet insurance pricing models across different insurers and markets

No two insurers price fleets identically, even when working from the same underlying risk. Some lean heavily on composite rating for administrative simplicity, others default to schedule rating the moment a fleet mixes vehicle types, and a growing number now offer usage-based options alongside traditional structures.

The practical difference shows up in what each insurer asks for. A composite-rating insurer wants fleet-level totals: vehicle count, value bands, general use. A schedule-rating insurer wants a vehicle-by-vehicle breakdown. A usage-based insurer wants telematics access or detailed mileage logs before it will quote competitively.

Markets also differ in how much weight they place on regional exposure. Territory loadings vary significantly depending on local risk patterns, malicious damage, weather exposure, theft rates, and insurers operating in the same market can price the same postal code differently depending on their own claims data. That is one reason fleet managers benefit from requesting the rating basis explicitly, as covered in the checklist above, rather than assuming all quotes reflect the same underlying assumptions.

Fleets operating across multiple regions face an added layer: exposure that looks manageable in one area can look very different in another, which is why understanding regional exposure factors matters before comparing quotes side by side.

Impact of regulatory changes on fleet insurance pricing

Regulatory oversight shapes fleet pricing in ways that are easy to overlook when comparing quotes on premium alone. The SAIA Code of Conduct requires insurers to reassess vehicle base values on a regular basis, which directly affects sums insured and, in turn, the premium calculated against them. A fleet insured against outdated valuations risks being underinsured at claim time, regardless of how competitive the quoted rate looked at renewal.

Regulatory expectations also influence how far insurers can lean on machine-learning pricing models. GLMs remain the preferred structure for regulatory transparency because their logic can be explained line by line, while models like XGBoost, despite stronger predictive accuracy in South African testing, raise interpretability questions that insurers must address before deploying them at scale.

For fleet managers, the practical takeaway is that pricing transparency is not optional. Any insurer using data-driven models should be able to explain, in plain terms, which factors moved your premium and by how much. Valuation reassessment schedules and rating methodology disclosures are worth confirming directly rather than assuming they match industry norms.

Examples or case studies illustrating how different pricing models affect premiums

Consider a mid-sized delivery fleet with fifteen light commercial vehicles operating similar routes and carrying comparable loads. Composite rating suits this fleet well: the vehicles are similar enough that one blended rate captures the risk accurately without the administrative overhead of scheduling each one individually.

Now consider a construction fleet mixing five heavy trucks, ten pieces of plant machinery and a handful of light vehicles for site supervisors. Schedule rating fits better here, since a single blended rate would overcharge the light vehicles to cover the heavier equipment’s higher severity potential.

A third case: a logistics operator whose vehicles sit idle for weeks between seasonal contracts. Under traditional annual rating, this fleet pays for road exposure it never uses. Under a usage-based model like PAYD, premiums track actual mileage, which can better reflect genuine exposure across a variable operating pattern.

Each case reflects the same underlying principle: the pricing model should match how the fleet actually operates. A model that fits a uniform, high-mileage delivery fleet can misprice a mixed, low-utilisation construction fleet, and vice versa. That mismatch is often where fleet managers find the most room to negotiate a better structure, not necessarily a lower rate on the same structure.

Comparison of three fleet pricing models

Publisher perspective: on-demand insurance and when it makes sense

Traditional fleet rating assumes fairly constant usage. On-demand cover breaks that assumption by aligning premium to actual use, which suits hire and rental operations, construction fleets with seasonal peaks, and mixed-use assets that sit idle between contracts. It is less suited to fleets running consistent, year-round mileage, where a traditional composite or schedule policy is often simpler and more cost-effective.

— Coert

If you want an on-demand option: how Truck & Plant On-Demand™ helps

Fleets with variable use rarely fit neatly into a fixed annual premium, and that mismatch is where cover often goes to waste. On-demand insurance lets you insure commercial vehicles and plant machinery only while they are working, adjusting cover as your operations change rather than locking you into a fixed structure for months you do not need it.

Truckplant

  • Cover commercial vehicles, trailers and plant equipment on flexible terms.
  • Suited to variable-usage fleets, tool and plant hire businesses, and contractors managing seasonal peaks.
  • Adjust or pause cover as work changes, without renegotiating a full annual policy.

To get started, work through the checklist above, then request a tailored fleet insurance quote and ask how telematics data can integrate into your on-demand cover.

Sources

Readers who want the underlying detail can consult the SAIA Code of Conduct on vehicle valuations and the comparative South African pricing study on machine-learning ratemaking.

FAQ

How much is fleet insurance?

Fleet insurance premiums depend on fleet size, vehicle type, usage patterns, claims history and location, so there is no single market rate. Insurers calculate a base premium from expected frequency and severity of claims, then adjust it using rating models such as composite, schedule or usage-based pricing.

What are the four types of insurance?

In a fleet context, this typically refers to core commercial cover categories: vehicle damage cover, third-party liability, goods-in-transit cover and specialised equipment cover such as plant or tool insurance. Exact categories vary by insurer and policy structure, so it is worth confirming which are bundled versus sold separately.

What factors do insurance companies use to determine premiums?

Insurers commonly weigh vehicle type, annual mileage, driver records, claims history, fleet size, location and the type of goods or equipment carried. These exposure variables feed into frequency and severity estimates that form the base premium before any credibility adjustments are applied.

What is considered fleet insurance?

Fleet insurance covers multiple vehicles owned or operated by a single business under one policy, rather than insuring each vehicle individually. It typically extends to commercial vehicles, trailers and, in some cases, plant machinery, with pricing based on the fleet’s overall risk profile rather than vehicle-by-vehicle underwriting.