The price tag is the smallest part of the story
When people first budget for connected devices, they usually start with the hardware: so many sensors at so much each. That number is real, but it is rarely the whole cost, and for deployments that run for years it is often not the biggest part. A useful estimate looks at the whole life of the system, from planning through to retirement. Thinking this way is sometimes called total cost of ownership.
The categories below are a starting checklist. Which ones matter most depends on the project.
One-time costs
- Design and planning: staff time to define the problem, choose parts, and test ideas, including any prototype and pilot work.
- Hardware: sensors, controllers, gateways, enclosures, mounting hardware, cables, and spares. Include accessories that are easy to forget, such as antennas, power supplies and weatherproof housings.
- Software and integration: setting up data storage, dashboards, alerts, and connections to other business systems.
- Installation: labor, travel, access equipment, and any permits or safety measures. Installation can cost a meaningful share of the whole project, especially across many sites or hard-to-reach locations.
- Commissioning and testing: checking that each device works, reports correctly and is in the right place.
- Training: time for the people who will use and maintain the system.
Recurring costs
- Connectivity: cellular plans, network equipment, or the cost of extending existing Wi-Fi coverage. Per-device fees look small but add up over many devices and years. Different options, such as cellular IoT or low-power networks, have very different cost structures.
- Cloud and data services: storage, processing and software subscriptions. Costs often grow with the amount of data and the frequency of readings, which is one reason to avoid collecting more than you need. See also how time series data accumulates over time.
- Power: electricity for mains-powered devices, and batteries for the rest. Battery replacement involves both the cost of the cells and the labor to change them.
- Maintenance: site visits, cleaning, recalibration, repairs, and troubleshooting. Staff time to watch the system and respond to alerts also counts.
- Software upkeep: applying security fixes and delivering updates, and keeping integrations working when other systems change.
- Support and management: keeping track of devices, replacing lost credentials, and handling user requests.
Costs that arrive later
Some expenses appear only after launch:
- Replacement: sensors drift, batteries die, and hardware fails. Plan for a failure rate and a replacement cycle. The expected device lifespan is often shorter than people hope, particularly outdoors.
- Obsolescence: a network technology or a supplier’s service may be retired, forcing early replacement.
- Scaling effects: things that work for a few devices can become expensive or labor-intensive with many.
- Disposal: responsible removal and recycling of devices and batteries carries a cost, covered in the topic of electronic waste.
- Compliance and security incidents: handling data properly, and responding if something goes wrong, can require time and money.
A simple hypothetical example
The numbers below are made up only to show the method.
Suppose a project uses 20 devices over 5 years. Hardware is 50 per device, so 1,000 in total. Installation is 30 per device, so 600. Connectivity and data services are 2 per device per month. Over 60 months that is 2 x 60 = 120 per device, or 2,400 for 20 devices. Maintenance visits cost 400 per year, or 2,000 over 5 years.
The sum is 1,000 + 600 + 2,400 + 2,000 = 6,000. Hardware is only 1,000 of that, about one sixth. Even in this simple, invented case, the purchase price understates the full cost. Real projects have more categories, and actual numbers vary widely.
Making the estimate more reliable
- Estimate in ranges, not single figures, and note your assumptions.
- Add a contingency, since early estimates are usually optimistic.
- Use the pilot to measure real values, such as installation time, data volume and failure rate, rather than guessing.
- Compare options on the same basis, including when you weigh whether to build or buy.
- Revisit the estimate as the project changes.
Cost should also be set against benefit. A deployment that costs more than the value of the decisions it supports is not worth running, however clever the technology. An honest estimate is what makes that comparison possible, and it is much easier to adjust a plan on paper than after devices are installed.