Usage-based insurance vs traditional pricing: what gets measured, what it can change and how to decide

Usage-based insurance prices part of your premium from measured driving: miles, braking, speed, time of day and, in some programs, phone use. Traditional pricing uses group factors and your record. How the two machines differ, which program shapes can raise a price as well as lower it, and a worksheet for deciding from your own driving pattern.

Estimates and published averages, not a quote. Not insurance advice. State rules and insurer rules vary. Confirm requirements with your state insurance regulator or a licensed agent in your state.

The two pricing machines, side by side

Every auto premium on this site is described the same way, because the NAIC describes it the same way: two factors determine what you pay. The first is underwriting, where the insurer assesses your risk and groups you with similar risks. The second is rating, which assigns a price based on what the insurer believes it will cost to carry your potential claims. Traditional pricing runs entirely on information known before or outside the drive itself: the driving records of everyone on the policy, the vehicle, the garaging address, the limits and deductibles you choose, your prior coverage history, and the claim frequency of the group you are sorted into. The NAIC is explicit that claim frequency within the group is the single greatest influence on the rating process, with your own claim history adding an effect on top. Usage-based pricing does not replace that machine. It adds a second, narrower reading on top of it: your measured behaviour in your own car, over a defined window. Both prices are still quoted per policy, per insurer, under a rating plan filed in your state.

What telematics actually measures

The NAIC consumer page on auto insurance lists the elements telematics devices measure: miles driven, time of day, where the vehicle is driven by GPS, rapid acceleration, hard braking, hard cornering and air bag deployment. NAIC consumer materials also describe programs that register whether you use a cell phone while driving. Collection runs through a smartphone app, a device you plug into the car, or equipment the manufacturer built in, and the NAIC notes that the level of data collected reflects both the technology used and the policyholder’s willingness to share personal data. Two consequences follow. First, a mileage-only program and a behaviour program are different products wearing the same label, and the quote conversation has to establish which one you are being offered. Second, the device cannot tell who is driving. On a shared car, the measured layer reads the car, and every driver of that car contributes to the reading.

The three program shapes, and why the shape matters more than the label

NAIC consumer guidance describes pay-as-you-drive and pay-per-mile as the two long-standing structures. In a pay-per-mile design, the insurer charges a base rate plus a per-mile amount, tracked by a device, and NAIC guidance notes this shape can suit drivers who drive little, work from home, use transit or attend college. In a pay-how-you-drive design, the measured behaviours themselves move the price. Across both structures, insurers build three different deals. Some programs apply a participation discount for enrolling and then a measured discount that can shrink or grow at renewal. Some re-price directly from the data, which means the measured layer can raise the renewal price as well as lower it. Some cap the downside in writing, so the tracker can only help. NAIC consumer materials note that an insurer may give a small discount just for agreeing to participate. The only version of the deal that protects you is the one stated in the program’s own terms, so ask two questions before enrolling and get the answers in writing: can this program ever charge me more than the same policy without it, and what happens to my price if I leave mid-term?

Whose driving pattern the measured layer tends to fit

This site does not publish a UBI saving figure, because no official source publishes one we can stand behind. What can be stated is the fit, in worksheet terms. The measured layer has the most room to help a driver whose unmeasured price is held up by group factors but whose own driving is light and regular: low annual miles, trips mostly in daylight, one driver per car, gentle braking because the driving is mostly uncongested. It has the least room to help, and the most room to hurt in a re-pricing program, where the pattern is the reverse: high miles, a commute in heavy braking traffic, regular driving between midnight and the early hours, or a car shared with a driver whose habits you would not choose. None of that is a verdict. It is the list of facts about your own driving to write down before you look at a UBI quote, which the worksheet below does in order.

Read the data terms before the discount terms

A traditional quote asks for facts you already know. A UBI quote asks for a data feed, and the terms for that feed are part of the price. Before enrolling, establish what is collected, in the program’s own list rather than a summary: location history or only trip mileage, phone handling or only phone motion, passengers’ phones or only the enrolled driver’s. Establish how long the data is kept, whether it can be used if you file a claim, whether it is shared with third parties or used for anything besides rating this policy, and what you see yourself, since many programs show a dashboard of the same events the insurer prices. The NAIC framing is the right test: the level of data collected reflects your willingness to share it. If a program cannot state its data list plainly, that is information about the program, and a traditional quote on identical limits is the comparison to fall back on.

The trial window is the real quote

Most behaviour programs measure for a defined period before the measured price applies, and that window is the closest thing UBI has to an honest quote. Treat it as one. Drive the car the way it is actually driven, by everyone who drives it, on the real schedule, including the late trip and the congested commute. A careful month that nobody in the household can sustain produces a discount built on a pattern that will not survive the next measurement, and the renewal corrects it. If the program reports events as they happen, read them during the window rather than after: a hard-braking count concentrated on one junction, or phone events recorded while a passenger held the phone, are facts you may be able to raise with the insurer while the window is still open, under the program’s own correction process if it has one.

A five-line worksheet for the decision

Work these five lines from your own records before comparing quotes. Line one: your annual miles, from two odometer readings or your last two inspection or service records, not a guess. Line two: your late-night share, estimated honestly from your schedule: trips between midnight and 4 a.m. are the window behaviour programs most commonly count against a driver. Line three: the driver list, meaning everyone who will actually drive the measured car, since the device reads the car. Line four: the program shape, discount-only or re-pricing, taken from the written terms. Line five: the exit, meaning the cost and process of leaving the program mid-term and at renewal. A discount-only program, a low line one and a daylight line two is the combination where enrolling is close to a free option. A re-pricing program with a high line one and a night-shift line two is where the traditional quote deserves the benefit of the doubt.

Compare the two quotes on an identical specification

A UBI quote and a traditional quote are only comparable on the same coverage specification: same liability limits, same deductibles, same drivers, same garaging address, same policy term. Price the traditional version first, in writing, then the UBI version, and record what the UBI figure assumes: an enrolment discount, a projected measured discount, or a measured result from a completed window. A projected discount is a forecast, not a price, and the gap between the two quotes should be read that way. To place either quote in context, use your state index page on this site: the combined average premium for your state, liability plus collision plus comprehensive for one insured vehicle, was published in the NAIC 2024 Auto Insurance Database Average Premium Supplement, supplement Table 5, with a countrywide combined average premium of $1,613 for 2024. A published average describes a group in a past period. It does not predict what a tracker will do to your renewal, which is why the written program terms outrank every comparison, including this one.

What the measured layer cannot see

A telematics reading is a count of events, not an account of them. A hard brake to avoid a child who ran into the road and a hard brake because you were following too closely on your phone register the same way unless the program records more context, and most record less than drivers assume. Congestion does the same flattening: a commute through stop-and-go traffic generates braking and acceleration events that an uncongested rural drive never produces, at the same level of care. Weather, roadworks and other drivers’ behaviour all enter your reading without entering your conduct. This is not an argument that the measurement is useless; insurers adopt it because, across large groups, the counts do separate risk. It is an argument for reading your own event log during the trial window and raising specific, dated errors through the program’s correction process where one exists, and for weighting the discount-only structure more heavily the less your driving environment resembles the pattern the score was built around. One example of a vehicle whose mileage differs from the norm is covered in electric vehicle car insurance.

The tracker is one line of the renewal, not the whole renewal

A measured discount and a higher total premium can arrive in the same envelope, because the tracker governs only its own layer. The traditional layer underneath it still moves with repair costs, injury claim costs, your state’s loss experience and every other rating factor, and a general increase can outweigh a genuinely earned telematics discount. Read the renewal the way this site reads every renewal: line by line, against the previous declarations page, on an unchanged specification. If the UBI line improved and the total still rose, the tracker did its part and the base rate did the rest, and the correct next step is the identical-specification comparison in our quote guide, not abandoning the program in frustration. The reverse reading matters too: a lower total after enrolling does not prove the tracker paid, if the base rate fell for unrelated reasons. Keep the program’s own reported discount figure, in writing, separate from the premium total. It is the only number that describes what the measured layer itself did.

Common questions