Fleet Data Analytics: What to Track, Which KPIs Matter, and How to Act on It

Why most fleets collect data they never use

The spreadsheet trap: 74% of fleets still rely on manual reporting

According to a [Fleet Maintenance Magazine survey](https://www.fleetmaintenance.com/), 74% of fleet operations still use spreadsheets as their primary reporting tool, even when they have telematics hardware installed in every vehicle. The data exists in the platform. The fleet manager exports a CSV, pastes it into Excel, builds a pivot table, and emails a PDF to the VP of operations. By the time anyone reads it, the numbers are two weeks old.

Spreadsheets are not analytics. They are data storage with manual effort attached. You cannot set alerts in a spreadsheet. You cannot correlate a spike in fuel costs with a change in driver behavior in a spreadsheet — not without hours of manual cross-referencing. And spreadsheets do not catch problems before they become expensive.

What separates a data-rich fleet from a data-driven one

A data-rich fleet has telematics on every truck, fuel cards with transaction data, and maintenance records in a CMMS. A data-driven fleet uses that same data to answer specific questions every week: Which routes are burning the most fuel per delivery? Which drivers need coaching before they cause an accident? Which trucks should be replaced this quarter instead of repaired again?

The difference comes down to three things: defined KPIs that tie to financial outcomes, a regular review cadence (weekly, not quarterly), and someone accountable for acting on what the data shows. I have seen 30-truck fleets outperform 300-truck operations on cost per mile simply because the smaller fleet had one person who reviewed fuel and maintenance data every Monday morning and made decisions by Tuesday.

What fleet data should you actually track?

Fleet data falls into five categories that matter: fuel, maintenance, safety, utilization, and driver behavior. Everything else is noise until you have these five dialed in. Each category has two or three metrics that drive 80% of the value. Here is what to focus on and what to skip.

Fuel data: cost per mile, idle time, and MPG by driver

Fuel is the largest variable cost in most fleet operations. According to [ATRI's 2024 operational cost report](https://truckingresearch.org/2024/11/18/an-analysis-of-the-operational-costs-of-trucking-2024/), fuel accounts for roughly 24% of total per-mile operating costs for trucking fleets. The three fuel metrics that actually drive decisions:

• Cost per mile by vehicle — not fleet average, per truck. A single vehicle running 20% worse than the fleet average costs $3,000-5,000 extra per year in fuel alone.
• Idle time percentage — excessive idling wastes 0.5-1.0 gallons per hour. Fleets with idle management programs cut fuel costs by 5-8%, per [Department of Energy estimates](https://afdc.energy.gov/conserve/idle-reduction.html).
• MPG variance by driver — same truck, same route, different driver. If Driver A gets 6.8 MPG and Driver B gets 5.9 MPG on the same equipment, that is a coaching opportunity worth $4,000+ annually.

Skip fleet-wide average MPG as a standalone metric. It hides the variance that matters. A fleet averaging 6.5 MPG might have trucks ranging from 5.2 to 7.4, and the trucks at the bottom are where the money is.

Maintenance data: PM compliance, cost per repair, and unplanned downtime

Maintenance is where fleet analytics pays for itself fastest. According to [NPTC fleet benchmarking data](https://www.nptc.org/benchmarking), fleets with 90%+ preventive maintenance compliance spend 25-30% less on total maintenance than fleets running below 80% PM compliance. Three maintenance metrics to track weekly:

• PM compliance rate — percentage of scheduled maintenance completed on time. Target: 95%+. Every point below 90% correlates with higher unplanned repair costs.
• Maintenance cost per mile — typical range is $0.15-0.22/mile for Class 8 trucks. If a specific asset is trending above $0.25/mile, it is signaling replacement.
• Unplanned downtime hours — the metric that connects maintenance to revenue. One day of unplanned downtime on a revenue-generating truck costs $800-1,500 in lost productivity and emergency repair premiums.

Safety data: harsh events, CSA scores, and incident frequency

• Harsh event frequency per 1,000 miles — the best leading indicator for accident risk. Fleets that coach drivers on harsh events see 20-30% reduction in preventable accidents, per [FMCSA safety data](https://ai.fmcsa.dot.gov/).
• CSA BASIC scores — monitored monthly. Any BASIC above the 75th percentile triggers increased FMCSA scrutiny and can disqualify you from hauling for safety-conscious shippers.
• Preventable incident rate — accidents per million miles. Benchmark: the [National Safety Council](https://www.nsc.org/) reports that top-performing fleets maintain rates below 1.5 per million miles.

Utilization data: asset usage rates and right-sizing signals

Utilization is the metric most fleets ignore until someone in finance asks why the company owns 120 trucks but only 95 move on any given day. Vehicle utilization rate measures the percentage of your fleet that generates revenue on a given day. A healthy benchmark for most operations is 85-92%.

Below 80% utilization, you are paying insurance, depreciation, and registration on trucks that sit. Each idle asset costs $12,000-18,000 per year in carrying costs before it turns a wheel, according to [Fleet Advantage lifecycle data](https://www.fleetadvantage.com/). Utilization data tells you when to right-size your fleet, when to add capacity, and which assets are candidates for disposal or redeployment.

Driver behavior data: speeding, HOS compliance, and coaching triggers

• Speeding events per 100 miles — every mph above 55 reduces fuel economy by approximately 0.1 MPG, per the [Department of Energy](https://fueleconomy.gov/feg/driveHabits.jsp). Track by driver, not by fleet.
• HOS compliance rate — percentage of logs submitted on time with zero violations. Target: 98%+. Below 95%, expect FMCSA attention.
• Coaching completion rate — how many flagged events result in actual driver conversations. An analytics platform that flags 200 events per week means nothing if nobody reviews them.

Fleet analytics maturity: where does your operation fall?

Not every fleet needs predictive AI models and real-time optimization engines. Analytics maturity is a spectrum, and jumping to Level 4 when you have not mastered Level 1 wastes money and creates frustration. Here is how to assess where your fleet sits today and what the next step looks like.

Level 1 — Reactive: fixing problems after they cost you money

At Level 1, data exists but nobody looks at it until something breaks. Telematics hardware might be installed, but the login credentials for the dashboard are in a Post-it on someone's monitor, and the only person who used it left six months ago.

Level 2 — Descriptive: dashboards that tell you what happened

The limitation of Level 2 is that it only tells you what already happened. The dashboard shows that fuel costs spiked 12% last month, but it does not tell you why. Was it a price increase? A routing change? A driver who started idling an extra 45 minutes per day? You need to dig manually, and most fleet managers do not have the time.

Level 3 — Predictive: forecasting breakdowns and cost trends

Predictive analytics uses historical patterns to forecast what will happen next. A truck that has been trending upward in engine fault codes over the past 90 days is statistically likely to fail within 30 days. A driver whose harsh braking rate has doubled over six weeks is more likely to be involved in an incident. Geotab's MyGeotab platform and Samsara's AI engine both offer predictive maintenance alerts as of 2026.

According to a [McKinsey report on predictive maintenance](https://www.mckinsey.com/capabilities/operations/our-insights/the-future-of-maintenance-for-distributed-fixed-assets), predictive approaches reduce maintenance costs by 10-40% and cut unplanned downtime by 50% compared to reactive maintenance. For a 100-truck fleet spending $500,000/year on maintenance, that is $50,000-200,000 in savings.

Level 4 — Prescriptive: automated decisions and closed-loop optimization

Prescriptive analytics goes beyond prediction to recommendation and automation. The system does not just flag that Truck 47 will likely need a turbo actuator replacement in three weeks — it schedules the repair at the closest dealer, orders the part, and adjusts the dispatch schedule to cover the gap. Very few fleets operate here today. This requires tight integration between telematics, maintenance, dispatch, and parts procurement systems.

Fleet KPIs that actually move the bottom line

KPIs are only useful if someone acts on them. A dashboard with 40 metrics is a screensaver, not a management tool. The four KPIs below cover the financial levers that matter most. If you are building a fleet analytics program from scratch, start with these four and add complexity later.

Cost per mile: the single number that tells you everything

Cost per mile is the master KPI for fleet operations. It rolls fuel, maintenance, insurance, depreciation, and driver costs into one number. According to [ATRI's 2024 data](https://truckingresearch.org/2024/11/18/an-analysis-of-the-operational-costs-of-trucking-2024/), the industry average for for-hire carriers is $2.27/mile. Private fleets typically run $1.80-2.50/mile depending on operation type.

Track cost per mile at three levels: fleet-wide (for executive reporting), by vehicle (to identify assets that need replacement), and by route or customer (to find unprofitable lanes). The vehicle-level view is where most savings hide. I have seen fleets where 10% of the trucks account for 30% of the total maintenance spend. Without per-vehicle cost per mile, those trucks stay in service burning cash.

Vehicle utilization rate: are you paying for trucks that sit?

Utilization rate is calculated as the number of revenue-generating days divided by total available days. A truck in the shop is not available. A truck without a driver assignment is available but not utilized. The distinction matters because your analytics platform should separate the two causes.

Target 85-92% for most commercial fleets. Construction and seasonal operations run lower (70-80%) by nature of the business. If your utilization is below 75% for non-seasonal reasons, you are either over-fleeted or your maintenance program is creating too much downtime. Either way, the data should tell you which before you make a decision.

PM compliance rate: the leading indicator for unplanned downtime

PM compliance measures the percentage of preventive maintenance tasks completed within their scheduled window. A PM that is due at 25,000 miles and gets done at 27,500 is late. A PM due in March that happens in May is late. Late PMs are how you end up with roadside breakdowns, emergency tow bills, and a truck sitting at a dealer for three days waiting on parts.

Fleets running above 95% PM compliance experience 40-60% fewer unplanned breakdowns than fleets below 80%, per [TMC/ATA maintenance benchmarking data](https://www.trucking.org/technology-maintenance-council). Fleetio and other maintenance-focused platforms make this metric easy to track with automated service reminders based on mileage, engine hours, or calendar intervals.

Total cost of ownership by asset

Total cost of ownership (TCO) per asset combines acquisition cost, financing, fuel, maintenance, insurance, and depreciation into one lifecycle number. This is the metric that tells you when to replace a truck. Most fleets run assets too long because they look at monthly maintenance costs in isolation. A truck costing $800/month in maintenance looks manageable until you add the $400/month in lost fuel efficiency versus a newer model and the $200/month opportunity cost of higher downtime.

According to [Fleet Advantage](https://www.fleetadvantage.com/), the optimal Class 8 replacement cycle is 4-5 years or 400,000-500,000 miles for most operations. Beyond that point, maintenance costs escalate nonlinearly and resale value drops below the cost of continued operation. Your analytics platform should surface TCO trending by asset so you can time replacement decisions proactively.

Fleet KPI benchmark table by fleet size

These benchmarks reflect 2024-2025 industry data. Your actual numbers will vary by region, freight type, and vehicle class. The value is not in hitting every benchmark — it is in knowing where you stand relative to peers so you can prioritize the biggest gaps.

How Samsara, Geotab, Motive, and Fleetio handle fleet analytics

Every major fleet platform claims to offer analytics. The reality is that they offer very different approaches to data, and the best fit depends on your maturity level, technical resources, and what you actually need to measure. Here is how the four leading platforms compare on analytics capabilities as of 2026.

Samsara: real-time operational dashboards and AI alerts

Samsara's analytics strength is real-time visibility. Their [Connected Operations platform](https://www.samsara.com/products/analytics) provides live dashboards for fuel efficiency, driver safety scores, vehicle location, and equipment utilization. Samsara's AI engine processes data from GPS, cameras, and engine diagnostics to generate automated alerts — like flagging a vehicle whose fuel efficiency dropped 15% over two weeks.

Where Samsara excels: operations with 50+ vehicles that need a single dashboard view across GPS, ELD, cameras, and sensors. Their pre-built report library covers most standard fleet KPIs without requiring custom configuration. Where Samsara falls short: customization. If you need to build highly specific reports or integrate with homegrown systems, Samsara's API is more limited than Geotab's.

Geotab: MyGeotab analytics and open API for custom reporting

[Geotab's MyGeotab platform](https://www.geotab.com/fleet-management-solutions/fleet-tracking-software/) is the most data-flexible option in the fleet telematics space. Their open API and SDK allow fleet teams (or their IT departments) to build custom dashboards, pull raw data into business intelligence tools like Power BI or Tableau, and create automated workflows triggered by specific data conditions.

Geotab processes over 75 billion data points daily across their global fleet, per their [corporate data page](https://www.geotab.com/fleet-management-solutions/data-analytics/). Their Marketplace offers pre-built analytics add-ons from third-party developers — everything from advanced fuel analytics to predictive tire management. For enterprise fleets with dedicated data teams, Geotab is the platform that gets out of your way and lets you build exactly what you need. For a 30-truck operation without IT support, that flexibility becomes complexity.

Motive: AI-powered analytics and automated driver coaching

[Motive](https://gomotive.com/products/analytics/) differentiates on AI-driven automation. Their platform uses machine learning to auto-detect coaching events, flag anomalies in fuel consumption, and predict maintenance needs. The AI Omnicam system processes dash cam footage to identify distracted driving, tailgating, and traffic violations without requiring a human reviewer to watch every clip.

Fleetio: maintenance reporting and lifecycle cost tracking

For fleets that already have a telematics platform handling GPS and driver safety but need better maintenance analytics, Fleetio fills the gap. Their cost-per-mile and lifecycle reporting helps fleet managers make data-backed replacement decisions. Fleetio pricing starts at $5/vehicle/month for their Starter plan, making it accessible for fleets that do not need a full telematics overhaul.

Fleet analytics platform comparison table

| Feature | [Samsara](https://www.samsara.com/) | [Geotab](https://www.geotab.com/) | [Motive](https://gomotive.com/) | [Fleetio](https://www.fleetio.com/) | |---------|---------|-------|--------|--------| | Real-time dashboards | Yes — operational focus | Yes — highly customizable | Yes — AI-driven alerts | Limited — maintenance focus | | Custom report builder | Basic | Advanced (SDK + API) | Moderate | Moderate | | API access | REST API | Open API + SDK + Marketplace | REST API | REST API | | Predictive analytics | AI-based alerts | Engine fault trending | AI anomaly detection | PM interval optimization | | Fuel analytics | Cost/mile, idle, MPG | Deep fuel analysis, IFTA | Fuel efficiency, IFTA | Fuel cost tracking via integrations | | Safety analytics | Safety scores, AI dash cam | Harsh event tracking | AI Omnicam coaching | Not primary focus | | Maintenance analytics | Basic work orders | Via Marketplace partners | PM tracking | Core strength — TCO, PM, work orders | | Best for | Mid-large fleets, all-in-one ops | Enterprise, data-heavy, custom | Mid-size, AI automation | Maintenance-focused, any size | | Pricing | ~$30-45/vehicle/month | ~$25-40/vehicle/month via resellers | ~$25-35/vehicle/month | $5-15/vehicle/month |

Pricing reflects 2025-2026 estimates from vendor websites and industry sources. Actual pricing varies by fleet size, contract length, and bundled features. Contact vendors directly for current quotes.

The ROI of data-driven fleet decisions

Fuel savings: 10-15% reduction from idle time and route data

Fuel is the largest controllable operating cost for most fleets. According to the [U.S. Department of Energy](https://afdc.energy.gov/conserve/idle-reduction.html), a heavy-duty truck burns approximately 0.8 gallons of diesel per hour while idling. For a fleet of 50 trucks averaging 2 hours of idle time per day, that is 80 gallons burned daily doing nothing — roughly $320/day or $83,000/year at $4.00/gallon diesel.

Fleets that implement idle reduction programs based on telematics data typically see 10-15% overall fuel savings within the first year. Samsara and Motive both report that customers using their fuel analytics tools reduce fuel spend by 10-12% on average. At $500,000/year in fuel costs for a 50-truck fleet, that is $50,000-75,000 back in operating margin.

Maintenance savings: predictive vs reactive cost difference

Reactive maintenance — waiting for something to break — costs 3-5x more than preventive maintenance per repair event, according to data from the [Technology & Maintenance Council (TMC)](https://www.trucking.org/technology-maintenance-council). A planned brake replacement at a fleet shop costs $400-600 per axle. An emergency brake failure on the road involves a tow ($500-1,500), a roadside repair premium (2x shop rate), a missed delivery ($1,000+ in penalties), and a potential out-of-service violation.

Insurance and safety: how analytics data lowers premiums

Fleet insurance premiums have climbed 10-20% annually since 2020, per [ATRI's cost analysis](https://truckingresearch.org/2024/11/18/an-analysis-of-the-operational-costs-of-trucking-2024/). Insurance is the cost category most fleet managers feel powerless to control. But insurers increasingly offer premium discounts for fleets that can demonstrate data-driven safety programs.

Common fleet analytics mistakes that waste time and money

Investing in analytics platforms without a clear strategy is almost as expensive as ignoring data entirely. These are the three mistakes I see fleet managers make repeatedly, and they all share the same root cause: buying technology before defining the problem it needs to solve.

Tracking 50 metrics and acting on zero

More metrics do not mean better decisions. I have seen fleet managers with dashboards tracking tire pressure variance, ambient temperature correlations, and regeneration cycle frequency — none of which they had the time, staff, or context to act on. Meanwhile, their PM compliance rate was 72% and nobody noticed because it was buried on page three of a 40-metric report.

Start with five metrics. Seriously. Cost per mile, utilization rate, PM compliance, fuel cost per mile, and one safety metric. Get those five dialed in, review them weekly, and act on the outliers. You can add complexity once the basics are running. Trying to boil the ocean on day one guarantees that nothing gets cooked.

Buying a platform before defining what you need to measure

A 40-truck delivery fleet does not need Geotab's open API and SDK. A 300-truck enterprise does not need Fleetio's basic reporting. But I regularly see the mismatch: small fleets buying enterprise platforms because the sales rep demo looked impressive, and large fleets stuck on tools they outgrew three years ago.

Before evaluating any platform, answer three questions: What are the three operational problems costing you the most money? What data would you need to see weekly to address those problems? Who on your team will actually review and act on that data? The answers determine your maturity level and narrow the vendor field to two or three realistic options.

Ignoring data quality: garbage in, garbage out

Analytics platforms are only as good as the data feeding them. If drivers are not logging fuel purchases correctly, your cost-per-mile numbers are wrong. If maintenance techs are closing work orders without entering parts costs, your TCO calculations are fiction. If GPS devices are not pinging because cellular coverage is spotty on rural routes, your utilization data has blind spots.

Data quality is not a technology problem. It is a process problem. Set expectations for data entry. Audit data completeness monthly. If your fuel cost reports are missing 15% of transactions, fix the fuel card integration before building dashboards on bad numbers.

How to build a fleet analytics program from scratch

You do not need a data science team or a six-figure analytics platform to start making data-driven fleet decisions. You need a clear process, the right starting metrics, and 30 minutes a week. Here is a four-step approach that works for fleets of any size.

Step 1 — Identify your three highest-cost problems

Pull your last 12 months of operating expenses and categorize them: fuel, maintenance, insurance, driver turnover, compliance violations, vehicle depreciation. Rank the categories by total spend. Your analytics program should target the top three cost categories first because that is where the data will generate the largest financial return.

For most fleets, fuel and maintenance dominate. But some operations bleed money on driver turnover ($8,000-12,000 per driver replacement, per the [ATA](https://www.trucking.org/)), insurance claims, or compliance fines. Your ranking might surprise you. That is the point.

Step 2 — Map data sources to those problems

For each of your three cost problems, identify where the relevant data lives. Fuel costs? Fuel card transaction data plus telematics idle reports. Maintenance costs? CMMS work orders plus parts invoices. Driver turnover? HR records plus telematics-based driver scorecards. Some data may already exist in platforms you are paying for but not using.

The goal is not to buy new technology at this step. The goal is to find out what data you already have, what is missing, and where the gaps are. Many fleets discover they are sitting on 80% of the data they need — it is just spread across four systems that do not talk to each other.

Step 3 — Pick a platform that fits your maturity level

If you are at Level 1 (reactive) with under 50 vehicles, a maintenance-focused platform like [Fleetio](https://www.fleetio.com/) ($5-15/vehicle/month) paired with a basic telematics device gives you the foundation. If you are at Level 2 with 50-250 vehicles and need a consolidated view, [Samsara](https://www.samsara.com/) or [Motive](https://gomotive.com/) provides pre-built dashboards that cover most KPIs. If you are at Level 3 with 250+ vehicles and a data team, [Geotab](https://www.geotab.com/) gives you the API access and raw data flexibility for custom analytics.

Do not over-buy. A platform you actually use at 60% of its capability beats a premium platform you use at 15%. The most common waste in fleet technology is paying for enterprise features that nobody configures.

Step 4 — Set review cadence and accountability

Analytics without a review cadence is decoration. Set a weekly 30-minute review where someone — fleet manager, operations director, whoever owns fleet costs — looks at the five core KPIs and identifies the two or three biggest outliers. That person then owns the follow-up: investigating the cause, making a decision, and tracking whether the change worked.

Monthly, run a deeper review: cost per mile trending, asset-level TCO, driver coaching completion rates, PM compliance. Quarterly, assess whether your analytics program is moving the numbers that matter. If cost per mile has not improved after two quarters of active data use, something in the process is broken — usually the accountability step.

Frequently asked questions about fleet data analytics