IIoT platform buyer's guide: how to choose the right system for discrete manufacturing

If you run a discrete manufacturing operation, you've probably noticed something: the industrial IoT platform market has exploded with options, and most of them seem built for someone else's factory. The pitch decks all look the same, the buzzwords blur together, and the gap between vendor demos and your actual plant floor feels enormous.

Here's what matters. You need to know which machines are running, which are down, and why, across every line, every shift. You need that answer without ripping out your existing controls or launching an 18-month IT project. This guide cuts through the noise and gives you a practical framework for choosing the right IIoT platform for your operation, whether you're running CNC mills, injection molders, stamping presses, or packaging lines.

Key terms every evaluator should know

Before diving into evaluation criteria, let's get the terminology straight. These distinctions matter when you're comparing vendors.

The critical point: IIoT doesn't replace SCADA. It sits alongside your control layer, adding analytics, multi-site visibility, and predictive capabilities that SCADA was never designed to provide. Think of it as a real-time operating layer on top of what you already have.

The real problem: you're losing money you can't see

Most plants track major breakdowns. What they miss are the micro-stops, speed losses, and idle time that silently drain throughput. Industry analyses, including research from Deloitte and BCG, consistently point to hidden production losses representing 5–20% of total output in typical plants, and most operations have zero visibility into them.

Here's what that looks like in practice: a molding shop thinks uptime is 80%. Detailed machine-level data reveals that nearly half the losses come from micro-stops and speed loss the operators have been compensating for without reporting. Every 8-hour shift loses roughly 1.6 hours to events nobody tracks.

The financial impact compounds fast. If your plant runs 2 shifts, 5 days a week with an average downtime cost of $500/hour (labor plus lost output), a 16-hour weekly loss adds up to roughly $416,000/year. Improving availability by even 5 percentage points recovers roughly $104,000 annually. Your specific cost structure and product mix will influence these numbers.

This is why the right machine monitoring solution for manufacturing pays for itself quickly: it exposes losses that are invisible without granular, machine-level data.

How non-invasive monitoring actually works on legacy equipment

The question every operations leader asks: "Can I monitor my existing equipment without shutting it down or modifying the control system?"

Yes. Here's how the most common approaches work in practice:

Current sensing through clip-on current sensors deserves special attention because it works on virtually any powered machine, regardless of age or manufacturer. Guidewheel's FactoryOps platform uses simple clip-on current sensors paired with proprietary algorithms to translate raw electrical signals into operational states, giving you run/idle/down data plus cycle-time visibility across everything from decades-old hydraulic presses to brand-new servo-driven lines. The system operates over cellular connectivity — no plant-wide Wi-Fi or complex network infrastructure required.

The key insight: the sensor is the easy part. The real value comes from turning raw signals into clear, actionable insight your team can use on the floor.

What "good" uptime actually looks like, by equipment type

When you're evaluating an industrial IoT platform, you need benchmarks to understand whether your plant has room to improve. These reference points come from industry surveys and should be adapted to your specific operational context, recognizing that product mix, material properties, and production scheduling all influence what "good" looks like for your facility.

Data from Guidewheel Performance Analysis across 3,000+ connected machines reveals an important nuance: the gap between median uptime and volume-weighted average uptime can be significant. High-volume machines heavily skew aggregate statistics, which means a plant's "average" number might mask the fact that most of your machines perform well below that line.

This chart illustrates why simple averages mislead. For packaging operations, the median runtime sits at 36% while the volume-weighted average jumps to 65%, a nearly 30-point spread. The takeaway: you need machine-level granularity, not plant-wide averages, to find your real improvement opportunities.

Where your downtime is actually hiding

Once you have machine-level data, the next question is: what's causing the losses? While external demand factors ("No Business/Orders") often dominate statistically, the downtime categories within your direct control represent the most actionable improvement targets.

Guidewheel Performance Analysis of over 10,700 downtime events reveals five controllable loss categories worth equal attention:

• Mechanical breakdowns are the most frequent actionable events at 3,316 events sampled, averaging 72 minutes per occurrence. These are prime candidates for predictive maintenance features in any IIoT platform you evaluate.

• Other operational losses represent 28% of total downtime at roughly 81 minutes per event, capturing the miscellaneous stops that add up invisibly.

• Staffing issues are less frequent (1,132 events) but cause significantly longer disruptions averaging 197 minutes each. Remote alerting and operator notification tools directly address notification delays.

• Maintenance and cleaning events average 85 minutes and account for about 11% of total downtime. Standardizing these routines based on data cuts this meaningfully.

• Material and supply disruptions average 119 minutes per event across 965 occurrences, representing opportunities for better upstream coordination.

The pattern is clear: mechanical breakdowns happen often but resolve faster, while staffing and material issues happen less often but cause far longer individual disruptions. A good industrial IoT solution helps you tackle both ends of that spectrum.

The five-criteria framework for choosing your platform

After working through the data, here's a practical scorecard for evaluating any IIoT platform or machine monitoring solution for your factory:

When estimating payback across a larger fleet, say 100 machines, the math scales predictably. Sensor and gateway costs typically run $1,000–$5,000 per machine depending on complexity, with software running $300–$700/month for the platform. Against even a conservative 3–5% availability improvement, most plants see payback within 3–6 months from reduced unplanned stoppages alone. Your specific numbers will depend on your hourly production value and current loss profile.

Why pilots stall, and how to prevent it

If you've ever asked why your IIoT pilot stalls after the first few machines, you're not alone. The most common reasons have nothing to do with the technology:

• Over-scoping from the start. Trying to solve uptime, predictive maintenance, ERP integration, and energy monitoring simultaneously. Start with machine status and downtime tracking. Add layers later.

• No operator involvement. IT-driven projects that bypass the people closest to the work fail. Include operators in vendor evaluation and pilot design.

• Perfectionism on data models. Waiting to define every possible downtime code before launching. Start with 70% of your categories defined and iterate.

• Deploying during peak production. Schedule your pilot during a maintenance window or slower demand period, not during a production crunch.

• No executive sponsor. Without a plant manager or operations director explicitly backing the project, it gets deprioritized at the first obstacle.

The organizations that scale successfully follow a consistent pattern: Pilot 5–10 machines, prove value in weeks, then expand systematically. Budget 4–5 months total for a full 40-machine site rollout, including integration and training, not the 2–3 years a SCADA replacement would demand.

Start seeing what your machines are telling you

Guidewheel's FactoryOps platform turns machine connectivity into operational clarity fast — without overhauling your controls, your network, or your team's workflow. That's the point.

If your plant is still relying on whiteboards, shift reports, or gut-feel estimates to understand production performance, the gap between what you think is happening and what's actually happening is costing you real money every week. The data consistently shows that the plants gaining ground aren't the ones with the biggest IT budgets. They're the ones that started measuring, started learning, and started improving — one machine at a time.

We had our best month of the year, increasing production from 26k–35k cases/month to 46k cases in March. I attribute this to Guidewheel. Being able to see downtime data and address downtime reasons directly correlates to higher production.

Michael Palmer, Direct Pack.

Guidewheel's FactoryOps platform is built for this exact approach: clip-on current sensors on any machine, cellular connectivity, and real-time visibility flowing within days, not months. The hidden capacity in your plant is real. The question is whether you're measuring it. If you're ready to see what your machines are actually telling you, Book a Demo and find out how much hidden capacity is sitting in your current lines — and how fast you can start recovering it.

About the author

Lauren Dunford is the CEO and Co-Founder of Guidewheel, a FactoryOps platform that empowers factories to reach a sustainable peak of performance. A graduate of Stanford, she is a JOURNEY Fellow and World Economic Forum Tech Pioneer. Watch her TED Talk—the future isn't just coded, it's built.