The Passenger Counting Market in 2026

An automatic passenger counting information system, usually shortened to APC-IS, is the combination of three layers: sensors above or beside vehicle doors that detect people boarding and alighting, an onboard gateway that ties each count to a stop and a timestamp using GNSS and door-open signals, and a back-office platform that turns those events into ridership reports, occupancy feeds for passengers and compliance files for funders. A business in this category sells some or all of those layers to transit agencies, private bus and coach operators, rail operators and, increasingly, airports and campuses running shuttle fleets.

Analyst firms disagree about the size of the market because they disagree about what to include. MarketsandMarkets values the combined counting-and-information-system market at $9.70 billion in 2025, rising to $14.80 billion by 2030 at an 8.8% compound annual growth rate (MarketsandMarkets, 2025). A separate release puts the same combined category at $9.35 billion in 2025 and $21.53 billion by 2035, a CAGR of 8.7% (OpenPR market release, 2025). Strip out the passenger-information displays, announcement systems and fleet management software, and the counting hardware and software alone is much smaller: roughly $452 million in 2025, forecast to reach $832 million by 2030 at a 13% CAGR (ResearchAndMarkets, 2025).

That gap matters for your plan. The $9.7 billion figure is the whole intelligent-transport stack, and a startup will never address most of it. The $452 million figure is the pool a sensor-plus-analytics company actually competes for, and it is growing faster than the wider category. Lenders who check your numbers will find both. State which one you mean and why.

What is pulling demand forward

Four buyer pressures account for most of the growth, and each one gives you a different sales argument.

• Federal reporting in the US. Agencies that report to the National Transit Database must report unlinked passenger trips and passenger miles. Where a reliable 100% count is unavailable, they sample to a 10% precision level at 95% confidence. Certified APC data replaces expensive manual sampling.
• Open data rules in England. Operators already have to publish timetable, fares and vehicle-location data through the Bus Open Data Service, and the Bus Services Act 2025, which became law on 27 October 2025, gives the Secretary of State powers to request and publish more operator-level data (GOV.UK consultation, 2025).
• Fleet size. American public transportation systems operated 133,153 railcars, buses, vans and other vehicles at peak in report year 2024 (APTA Public Transportation Fact Book, 2026). Each of those vehicles is a potential sensor install, and most older ones still have none.
• Overcrowding and service-quality contracts. Franchised and subsidised networks increasingly write occupancy and punctuality into contracts, which makes verifiable counts a revenue-protection tool for the operator rather than a reporting chore.

Who buys, and what each buyer needs

The buyer decides the product. A large metropolitan agency issues a tender, wants integration with an existing computer-aided dispatch and automatic vehicle location system, and will ask for an accuracy benchmark before award. A 30-vehicle independent operator wants a kit its own fitters can install in an afternoon, a monthly invoice, and a report it can send to the local authority that pays its subsidy. A rail operator cares about train-length coverage, EN 50155 hardware ratings and per-car loads. A university running a campus shuttle wants a dashboard and a number to put in a sustainability report.

Sales-cycle ranges are Avvale planning assumptions drawn from how public procurement typically proceeds, not statistics from a published survey. Put your own range in the plan once you have spoken to three or four buyers, and show lenders how many months of runway your cash covers while a tender is open. The founders who get funded are the ones who aim the first 18 months at the segments with short cycles and treat the metro tender as a year-three prize.

Concentration is worth a paragraph. Reported industry structure shows five main players holding roughly 30 to 35% of the market, with second-tier specialists such as Clever Devices, Trapeze Group, Syncromatics and Eurotech taking around 28% of revenue (Spherical Insights, 2025). That leaves a long tail of regional integrators and sensor-only makers, which is where a new entrant has room. It also tells you who you will meet in every tender.

What Operators Ask Before They Buy

These are the questions that come up again and again in search results and in procurement meetings. Your plan's product and marketing sections should answer each one with a number you can defend.

How does an automatic passenger counting system work?

A sensor mounted above each door looks down at the entrance steps and detects heads or bodies crossing a virtual line, recording one event for boarding and one for alighting. Early systems used paired infrared beams. Current designs use stereoscopic cameras or time-of-flight depth sensors that build a three-dimensional image, which lets the software separate an adult from a child, a pushchair or a suitcase. The sensor passes the event to an onboard unit that adds the GNSS position, the stop identifier and the time, and uploads the batch over a cellular link when the vehicle is in coverage.

How accurate are automatic passenger counters?

Technology decides the answer. Infrared beam arrays are commonly reported at 85 to 90% accuracy in normal operation. Stereoscopic and time-of-flight sensors are reported at 98% and above, and one vendor guide cites 99.9% for AI-assisted 3D units (Smart Urban Sensing, fleet ROI guide). Manual clicker counts are reported at 10 to 15% error because of fatigue and crowding. Vendor accuracy claims are measured under their own test conditions, so a buyer will usually run its own benchmark against hand counts on a sample of trips. Budget for that exercise in your pilot phase.

How much does an APC system cost per bus?

Published buying guides put sensor hardware at about $400 to $800 per vehicle for basic 2D units, $850 to $1,500 for thermal and $1,100 to $2,000 for stereo vision, with installation labour of $220 to $380 per vehicle, software at $18 to $42 per vehicle per month, and calibration and support at $120 to $200 per vehicle per year (Alibaba Electronics buying guide). An Oregon DOT white paper from 2021 quoted an infrared matrix system at about $4,000 per bus to acquire and install, with roughly $1,000 a year for software services (ODOT / Trillium white paper, 2021). The spread between those two sources, which is a factor of two or more, is your pricing opportunity and also your risk.

Do passenger counting systems comply with GDPR?

Time-of-flight, 3D stereo and similar depth-sensing designs output anonymous counts and store no faces, so they sit comfortably inside privacy rules. Counters built on ordinary RGB video can raise biometric-data questions under GDPR and, in the EU, the AI Act, which is why several procurement teams now ask for camera-free or depth-only designs (Terabee, GDPR and people counting). Decide early which family your product belongs to, because it shapes your compliance file, your data processing agreements and the sort of buyer who will sign.

Startup Costs & Capital Needs

Starting an APC-IS company typically needs $145,000 to $520,000 (£115,000 to £410,000) before the first paid contract, depending on one decision: do you design your own sensor, or do you integrate a third-party sensor and compete on software, installation and service? The low end assumes the second route. The high end assumes you fund embedded firmware, a custom enclosure and your own electromagnetic testing. These are Avvale planning estimates built from component pricing, typical engineering day rates and certification quotes, not audited figures.

Line-by-line budget

Notice what is missing: premises. A sensor company can start from a rented bench in a shared electronics space, one test bay and a van. That is why the founders who run this business well spend their money on engineering time, validation and inventory, not property.

The hidden cost nobody lists

Payment terms. A public agency may pay 45 to 90 days after acceptance, and acceptance can wait for the accuracy benchmark. If you buy $78,000 of kit for a 100-vehicle order in month one and are paid in month five, you have financed your customer. Your cash-flow forecast should show hardware cost leaving in the month of purchase and cash arriving 120 days later, and your funding ask should cover the largest gap in that forecast. Many first-time APC founders raise enough to build the product and not enough to ship a second large order.

Sensor Hardware, Vendors & Supply Chain

Your bill of materials and your supplier relationships decide gross margin. Here is how the technology options compare, followed by the named companies you will meet as suppliers and competitors.

Sensor technology comparison

Accuracy and price ranges are compiled from vendor guides and a transit-industry review; see the sources in the previous sections. Use them as planning inputs, then replace them with quotes.

Who you will compete with, and sometimes buy from

• iris-GmbH infrared & intelligent sensors: sensor specialist known for infrared and stereoscopic counters; launched the IRMA 6 R2 sensor in September 2023, with AI-based passenger and object counting (MarketsandMarkets company list). Often sells sensors to integrators who then own the customer.
• DILAX Intelcom GmbH: German counting specialist with both onboard and station counting; frequently appears in European tenders.
• HELLA Aglaia Mobile Vision GmbH: camera-based counting with a vehicle-industry parent, strong in bus and rail original equipment.
• Clever Devices and Trapeze Group: North American and global transit software groups that bundle counting into wider fleet platforms; they win on integration with dispatch and scheduling.
• Eurotech and Siemens Mobility: Eurotech supplies rugged edge computers with AI modules for buses, trains and trams, and Siemens Mobility sells counting as part of rail and traffic systems.
• Syncromatics and INIT: passenger information and fleet suppliers that regularly pair counting with real-time arrival displays.

Your position against this list is the main strategic choice in the plan. Competing head to head with a tier-one vendor on a metro tender is a poor use of a small team. A defensible positioning is retrofit and service for independent operators and regional agencies, where a $145,000 tender is too small for the big names to chase and where install speed and support responsiveness decide the sale.

Equipment and tooling checklist

• Door sensors, two per vehicle for dual-door buses: $200 to $1,000 each at volume depending on technology.
• Onboard gateway: an industrial computer with GNSS, CAN or J1939 input, and a cellular modem; $150 to $450 at volume.
• Door-state and odometer interface: harness and relay boards so the gateway knows when doors open and close; $20 to $60.
• Mounting hardware and enclosures: vandal-resistant brackets and sealed housings; $15 to $50 per door.
• Calibration rig: a reusable door frame with a height-adjustable gantry and a set of mannequins or recorded test walks; $3,000 to $9,000 one-off.
• Test vehicle access: a donated or leased bus for six months of pilot work; often negotiated with a friendly operator in exchange for free hardware.
• Diagnostic tools: laptop with CAN analyser, cellular signal meter, thermal camera for enclosure checks; $2,500 to $6,000.
• Spares pool: 5 to 8% of installed units held back for warranty swaps, which becomes a line in your inventory budget.

Ask three distributors for quotes at 50, 200 and 1,000 units before you write the margin assumptions. Component price breaks at 200 units are often large enough to move your hardware gross margin by five to eight points.

Revenue Model & Unit Economics

An APC-IS company makes money in four ways, and the proportions change as the business matures. In year one, hardware and installation dominate. By year three, a healthy company earns most of its gross profit from software subscriptions, support contracts and data services.

• Hardware sales or leases: sensor kits sold per vehicle, or leased on a monthly basis for operators who cannot fund capital outlay.
• Installation and commissioning: fixed price per vehicle, or a day rate for large fleets.
• Software subscription: per-vehicle monthly fee for the dashboard, reports and API; published buying guides cite $18 to $42 per vehicle per month.
• Services: annual calibration and support at $120 to $200 per vehicle, certification support packs for agencies preparing an NTD benchmarking plan, and custom data work.

Worked example: a 100-vehicle retrofit contract

A regional operator signs for 100 two-door vehicles. Your numbers for the first year look like this.

Hardware cost of $780 per kit assumes two sensors at $310 each plus $160 of gateway share, harness and enclosure. Software direct cost of $5.80 per vehicle per month covers cellular data, hosting and support tooling. In year two the same 100 vehicles produce $34,800 of subscription revenue at an 80% gross margin and $12,000 to $20,000 of calibration and support work, so a mature contract contributes about $40,000 to $48,000 of annual gross profit with very little selling cost.

Now test the model against your overheads. If a six-person team, rented bench space, insurance and travel cost $520,000 a year, you need roughly five 100-vehicle contracts in year one to cover fixed cost from first-year gross profit alone ($520,000 divided by $100,840 is 5.2). Recurring revenue from earlier contracts lowers that hurdle each year, which is why subscription attach rate is the single assumption lenders will push on. A plan that shows 90% of customers on subscriptions at month 12 is more credible than one that assumes 100%.

Margin ranges to defend

• Hardware gross margin: 35 to 50%, rising with volume purchasing and falling when you discount to win tenders.
• Software gross margin: 70 to 80% after cellular, hosting and support costs.
• Blended gross margin: 45 to 60% in years one and two, drifting up as recurring revenue grows.
• Net margin once stable: 8 to 18%, a planning range for small transit-technology firms; your own figure depends on engineering headcount.

The payback argument your customer will test

Operators do not buy sensors; they buy decisions. One vendor guide reports that reallocating a single underused 40-foot bus to a high-demand corridor can save more than $85,000 a year in fuel, maintenance and driver hours, and that smaller operators typically break even on an APC investment in 14 to 22 months, with the return becoming measurable at around 15 vehicles (Smart Urban Sensing, 2026). Treat those as vendor claims and build a case from your pilot customer's data. A one-page return-on-investment sheet, populated with the operator's own fuel cost, driver wage and fleet size, is the document that moves a purchase through a finance committee.

Funding: SBA 7(a), Start Up Loans & Grants

Capital for this business divides into three buckets: debt secured against equipment and receivables, equity for engineering, and public money tied to transit modernisation. Each lender reads the business plan differently.

US: SBA 7(a) data for technology firms

In 2025, SBA 7(a) lenders approved about $205.5 million of loans across 488 software and IT businesses, with an average loan of roughly $421,000. Only 23 of those loans, around 5%, went to startups (GoSBALoans, software and IT lender ranking, 2025). Across all industries the fiscal 2025 average was about $477,571. The message is plain: 7(a) money exists for technology companies, but lenders rarely write it for pre-revenue ones.

• What helps: signed pilot agreements or letters of intent from two or three operators, a founder with transit or embedded-systems experience, and equipment or inventory that serves as collateral.
• What hurts: a plan that depends on one metro tender, no accuracy test data, and software value that cannot be pledged.
• Classification: a hardware-heavy company may be classified under NAICS 334519 (other measuring and controlling devices) and a software-heavy one under 541511 (custom software development). Ask your lender which fits your revenue mix before you apply.
• Equity injection: startups should expect to contribute roughly 10% of the project cost in cash, so a $310,000 project needs about $31,000 from the founders.

UK: Start Up Loans and public money

The British Business Bank's Start Up Loans scheme lends from £500 to £25,000 per person over one to five years, with no fees and no guarantor. For applications from 6 April 2026 the fixed rate is 7.5% a year, up from the 6% that older guides still quote (Start Up Loans service update, 2026). Two co-founders can borrow £50,000 between them, which covers a pilot fleet and a certification run but not the full launch. The rest typically comes from angel investment, SEIS or EIS-eligible equity, Innovate UK programmes and a commercial equipment lender. The government has said it is investing over £1 billion a year in bus services between 2025/26 and 2028/29 (GOV.UK, vision for buses), and local authorities that spend it need evidence of ridership. That is demand for your product, and the plan should say so.

How to read this for your own plan

A sensible stack for a $310,000 raise looks like this: $60,000 to $100,000 from founders and friends, $100,000 to $150,000 of angel or SEIS equity tied to the pilot results, and a $75,000 to $125,000 term loan or equipment facility once you hold signed purchase orders. Show the lender the order in which each tranche arrives, because a loan that closes before you have customers is harder to approve than one that follows a signed pilot.

Where the Demand Sits: US, UK & EU

Demand for passenger counting is regional because the buyers, the rules and the money are regional. Decide your launch geography with the same discipline you apply to product choice.

For a UK founder, the cleanest launch sequence is England first, because the data publication rules create a built-in need for your outputs, then the US through a reseller or a state-level pilot. For a US founder, start with two or three agencies in one state so you can walk each site and visit within a day.

Certification, Privacy & Compliance

No single licence lets you sell passenger counting systems. Instead, your product passes through three kinds of checkpoint: your customer's reporting rules, product safety and electromagnetic testing, and privacy law. Each has a cost in the plan.

United States

• FTA approval for NTD use: agencies that use APC data for NTD reporting need FTA approval, a benchmarking plan and a maintenance plan, must meet 95% confidence and 10% precision on counts, and recertify roughly every three years. The FTA certifies the data methodology, not your hardware, so your product should make that process easier (TransTrack Systems, APC certification; Federal Register, NTD notice).
• Buy America: FTA-funded purchases have domestic-content rules that apply to rolling-stock components and must be checked per contract.
• FCC equipment authorization: radio-emitting devices, including a cellular modem, need FCC compliance, usually through a pre-certified module.
• State business registration, sales-tax and workers' compensation: standard for a company with fitters on site.

United Kingdom

• UKCA or CE marking and EMC testing: required for electronic equipment placed on the Great Britain market; confirm current recognition rules with your test lab.
• UK GDPR and Data Protection Act 2018: register with the Information Commissioner's Office and keep a data protection impact assessment if your sensor processes video.
• Bus Open Data Service formats: your outputs should map to the data standards operators must publish, so customers can reuse your feeds.
• Health and safety on depot sites: risk assessments and method statements for fitters; many operators ask for these before granting depot access.

European Union and elsewhere

• EU GDPR and the AI Act: camera-based counters that process biometric features attract extra scrutiny; depth-only or time-of-flight designs are easier to defend.
• CE marking, vehicle EMC approval and, for rail, EN 50155: the standard for electronic equipment on rolling stock.
• Data processing agreements: operators who hand you ridership data expect a written agreement on storage location, retention and deletion.

Your plan should name each of these, assign a cost and a month, and show the order. A founder who can say that the accuracy benchmark ends in month eight and the EMC report arrives in month ten reads as someone who has done this before.

Five Mistakes That Sink APC Startups

1. Promising a single accuracy number

"99% accurate" means nothing without a test condition. Accuracy at a quiet suburban stop with one passenger at a time differs sharply from accuracy at a university stop with 40 students boarding in 20 seconds. Agencies preparing an NTD benchmarking plan will measure against hand counts on sampled trips. Publish accuracy by crowding level, and show the sample size behind each figure. Founders who quote one number to a procurement panel get asked to prove it, and the first failed benchmark can cost the account.

2. Ignoring door-state integration

A sensor that counts continuously, instead of only when doors are open, records people walking past the stairwell and inflates the totals. Reading door status from the vehicle's CAN bus, or from a relay on the door actuator, is the dull wiring job that separates a working install from a failed pilot. Plan fitter training around it and count wiring time in your installation cost.

3. Underpricing support

Sensors drift. Doors get replaced, brackets get knocked, and a cleaning crew sprays water onto a lens. If your subscription price ignores the technician hours that calibration requires, margin leaks away in year two. The $120 to $200 per vehicle per year published for calibration and support is a floor for planning, not a ceiling.

4. Building for the biggest tender first

Metro tenders reward references, scale and balance-sheet strength, and a new company has none of the three. Founders who spend the first year writing a 400-page bid usually have nothing installed when the money runs out. Win five independent operators, publish the benchmark data, and arrive at the large tender with proof.

5. Treating privacy as a legal footnote

Procurement teams now ask whether your sensors store images, where counts are hosted and how long they are kept. Answering "we are working on it" ends the conversation. Put a one-page data protection summary in the sales pack, and choose a hosting region that matches your first market.

Glossary of Transit-Data Terms

Lenders and transit buyers use specialist vocabulary. Using it correctly in the plan signals that you belong in the room.

• APC: automatic passenger counter, the sensor and software that count boardings and alightings.
• AVL: automatic vehicle location, the GPS-based tracking that tells the system where each count happened.
• NTD: the US National Transit Database, where agencies report ridership and finance data to the Federal Transit Administration.
• UPT and PMT: unlinked passenger trips and passenger miles travelled, the two ridership measures the NTD requires.
• GTFS and GTFS-Realtime: open formats for scheduled and live transit data; occupancy information can be published inside a GTFS-Realtime feed.
• SIRI: the Service Interface for Real Time Information, a European standard for exchanging vehicle and stop data.
• Boarding and alighting: passengers getting on and off at a stop; the difference gives onboard load.
• Load factor: passengers onboard divided by seats or design capacity, the figure used to spot overcrowding.
• Ghost bus: a scheduled trip that runs nearly empty, or a vehicle recorded as running that never operated; counts help expose both.
• Benchmarking plan: the document an agency submits to the FTA showing how APC counts will be compared with manual counts.

24-Month Launch Timeline

• Months 1 to 3: incorporate, secure funding, choose the sensor route (integrate or design), sign a pilot agreement with one operator.
• Months 3 to 6: build the gateway firmware and ingestion platform, fit the first 5 to 20 pilot vehicles, collect paired hand counts.
• Months 6 to 9: publish the accuracy benchmark, fix door-state and calibration issues, start EMC testing.
• Months 9 to 12: convert the pilot to a paid contract, win two to three more independents, hire a field engineer.
• Months 12 to 18: reach 300 to 500 installed vehicles, add GTFS-Realtime occupancy and an API, begin regional agency conversations.
• Months 18 to 24: respond to the first public tender with references, reach positive monthly gross profit from subscriptions and services, open a second market through a reseller.

Related reading on adjacent plans: our GPS vehicle tracking business plan template, the automotive telematics business plan template and the 3D sensors business plan template cover overlapping ground on fleet data, in-vehicle hardware and sensing.

Sample Business Plan Preview

This preview shows how the executive summary and financial view read for a retrofit-focused APC company. The figures are illustrative and tie back to the worked example above.

What's in the Template

Every Avvale business plan template includes these sections, pre-structured for your industry. For an APC-IS company we add prompts on accuracy benchmarking, sensor sourcing and per-vehicle subscription pricing.

• Executive Summary: your business at a glance, written to hook investors in 60 seconds
• Company Overview: legal structure, ownership, location and founding story
• Industry Analysis: market size, growth trends and regulatory rules for transit data
• Customer Analysis: operator segments, buying triggers and procurement routes
• Competitor Analysis: mapping integrators, sensor makers and fleet-software groups, and your differentiation
• Marketing Plan: channels, messaging and customer acquisition strategy for a long sales cycle
• Operations Plan: installation workflow, staffing, supplier management and key milestones
• Management Team: founder bios, advisory board and key hires planned

The optional Financial Forecast add-on (included in our $300/£250 and $1,000/£800 packages) provides a 5-year Excel model with income statement, cash flow, balance sheet, break-even analysis, and startup capital requirements. For this sector the model includes a hardware-and-subscription revenue split, a payment-term cash gap and a fleet-installation ramp.

If you want help beyond the template, our business plan writer service pairs you with a consultant, and our client case studies show finished plans in related technology and transport sectors.

Frequently Asked Questions