Digital TransformationAdvanced8 min read

IoT Strategy

IoT Strategy is the deliberate plan for instrumenting physical things — equipment, vehicles, buildings, products — with sensors and connectivity, then converting the resulting data into decisions that improve operations, enable new revenue, or differentiate products. The strategic theory is compelling: if you can see what's happening to every machine in real time, you can predict failures, optimize utilization, charge for outcomes (uptime, throughput) instead of equipment, and build durable customer relationships through service. The reality is that IoT programs more often produce expensive dashboards that nobody uses than the predictive-maintenance gold rush their business cases promised. IoT done right is operating-model transformation; IoT done wrong is sensor theater.

Also known asInternet of Things StrategyIndustrial IoTIIoTConnected Products StrategySmart Operations

Challenge a friend Browse library

The Trap

The trap is starting with 'connect everything' rather than 'what decision does this enable?' Many IoT programs install thousands of sensors, build a data lake, then look for a use case to justify it. The result: hundreds of TB of telemetry, no clear analytics workflow, no operator changing behavior because of the data. The other trap: underestimating the unsexy parts — connectivity (cellular, LoRaWAN, satellite all have tradeoffs and ongoing cost), device management (firmware updates across 50,000 devices), security (every connected device is an attack surface), and integration with operational workflows. Vendors quote the platform; the operating reality includes all the rest.

What to Do

Apply the 'decision-first' IoT design: (1) name the operational decision the IoT data should change (e.g., 'when should this pump be serviced?'), (2) name the decision-maker (the maintenance technician, the operations manager, the dispatcher), (3) define the integration: how does the IoT signal reach the decision-maker in the workflow they already use (work orders, alerts, dashboards)? (4) THEN scope the sensor, connectivity, and platform required to deliver that signal. Pilots that follow this sequence routinely deliver ROI. Pilots that skip to sensor selection routinely produce dashboards nobody opens.

Formula

IoT ROI = (Decisions Improved × Value per Decision) − (Hardware + Connectivity + Platform + Integration + Device Management + Security Operating Cost)

In Practice

GE's Predix platform is the canonical cautionary tale: a multi-billion-dollar bet to be the 'industrial internet OS,' positioning Predix as the platform for industrial IoT analytics. By the late 2010s, GE had written down the investment and divested significant parts of GE Digital after concluding the platform business was harder to scale than expected and not the right fit for GE's core industrial operations. By contrast, Schneider Electric's EcoStruxure has had a more grounded trajectory by tying IoT to specific operational outcomes — energy management, equipment uptime — for specific customer segments. The lesson: IoT platforms succeed when scoped to specific operational decisions; they fail when sold as horizontal infrastructure.

Pro Tips

01
The single most expensive line item in IoT TCO is connectivity over the device lifetime, not the sensor or the platform. A $40 sensor with $8/month cellular connectivity over 7 years is $670 of connectivity vs $40 of sensor. Connectivity choice (cellular vs LoRaWAN vs WiFi vs satellite) often determines whether the business case works.
02
Device firmware management is the most-underestimated operating burden. Push a buggy firmware update to 30,000 devices and you have 30,000 operational incidents. Build the OTA (over-the-air) update infrastructure, staged rollout capability, and rollback before scaling beyond the pilot — not after.
03
Treat IoT data as a stream, not a lake. Most useful IoT decisions need <60 seconds of latency; storing everything in a data lake for batch analytics misses the point. The architecture for an actionable IoT signal is event stream → rule engine → operational system, not sensor → data lake → BI tool.

Myth vs Reality

Myth

“IoT primarily generates new revenue (outcome-based pricing)”

Reality

Outcome-based business models are the headline pitch but the rare reality. Most IoT value comes from operational cost reduction (predictive maintenance, energy optimization, asset utilization) — much less glamorous but more measurable. Companies that justify IoT on revenue transformation usually disappoint; companies that justify it on cost reduction usually deliver.

Myth

“IoT analytics requires AI and machine learning”

Reality

Most useful IoT decisions come from simple rules: 'temperature exceeds threshold for 5 minutes, alert,' 'vibration pattern matches known failure signature, schedule maintenance.' ML adds value at scale but not at the start. Many IoT programs over-invest in ML platforms before the simple rule engine is producing actionable alerts.

Try it

Run the numbers.

Pressure-test the concept against your own knowledge — answer the challenge or try the live scenario.

🧪

Knowledge Check

A manufacturer installs 4,000 sensors across a plant for 'predictive maintenance' but 18 months in, no actual maintenance schedule has changed based on the data. What's the most likely root cause?

Industry benchmarks

Is your number good?

Calibrate against real-world tiers. Use these ranges as targets — not absolutes.

Connectivity TCO as % of Total IoT Program Cost

Industrial and consumer IoT deployments over 5-7 year device lifetime

Optimized (LoRaWAN/private)

< 15%

Typical (mixed protocols)

15-30%

Cellular-Heavy

30-45%

Connectivity-Dominated

45-60%

Connectivity Eats the Case

> 60%

Source: Patterns from McKinsey IoT and IoT Analytics market reports

Real-world cases

Companies that lived this.

Verified narratives with the numbers that prove (or break) the concept.

⚡

GE Predix

2013-2019

failure

GE positioned Predix as the operating system for the industrial internet, investing billions in the platform and standing up GE Digital as a major business unit. The bet was that GE could become a software company by selling Predix to industrial customers as the platform for IoT analytics. The reality proved harder: building a horizontal platform required scale and ecosystem dynamics GE didn't fully control, customers preferred buying outcomes (predictive maintenance for specific equipment) over a horizontal IoT platform, and the platform business demanded different capabilities than GE's core industrial operations. By the late 2010s, GE wrote down significant portions of the Predix investment and divested or restructured GE Digital.

Investment Scale

Billions of dollars

Strategic Bet

Industrial IoT platform

Outcome

Major writedown and restructuring

Root Cause

Horizontal platform without scale or fit

IoT platforms succeed when scoped to specific operational decisions for specific customer segments. Selling a horizontal IoT 'OS' is a much harder business than selling vertical IoT outcomes. Most enterprises evaluating IoT should be customers of someone else's vertical platform, not builders of their own horizontal one.

🔌

Schneider Electric (EcoStruxure)

2016-2024

success

Schneider Electric built EcoStruxure as an IoT-enabled architecture for energy management and equipment monitoring across buildings, data centers, industrial plants, and grid infrastructure. Unlike Predix's horizontal positioning, EcoStruxure was tied tightly to specific operational outcomes Schneider already sold (energy efficiency, equipment uptime, sustainability reporting) for specific customer segments where Schneider had deep domain expertise. The IoT layer was infrastructure for a value proposition customers already understood and paid for. Reported customer outcomes include measurable energy savings and operational efficiency gains in deployed sites.

Approach

Vertical IoT tied to existing value props

Domain

Energy and equipment management

Customer Segments

Buildings, industrial, data centers, grid

Differentiator

Domain expertise + IoT, not IoT alone

IoT works as a layer on top of existing domain expertise and existing customer value propositions. It rarely works as a standalone proposition. Schneider's EcoStruxure succeeded where Predix struggled because it amplified an existing customer relationship rather than trying to create a new platform business from scratch.

Source ↗

Decision scenario

The Connected Products Bet

You're the COO of a $1.5B industrial equipment manufacturer. Competitors are launching 'connected' product lines. The board wants a strategy. Vendors are pitching a $20M IoT platform. Your equipment is sold through distributors and serviced by customers' own maintenance teams. Margins are under pressure from low-cost competitors.

Annual Revenue

$1.5B

Equipment Sold per Year

~14,000 units

Average Equipment Life

12 years

Service Revenue (Current)

<5% of revenue

Vendor IoT Proposal

$20M Year 1

Decision 1

You can pursue an ambitious horizontal IoT platform strategy ($20M+ to build a connected product line broadly), a focused vertical bet ($3M Phase 1 for one customer segment with one use case), or a fast-follower strategy (license a third-party IoT platform vs building). What's the call?

Approve the $20M horizontal IoT platform — go big to capture the strategic transition before competitors lock inReveal

18 months in: $16M spent, beautiful platform exists, ~6,000 connected units in field. But customers haven't paid extra for the connected tier, distributors don't sell it well (different sales motion), and operational outcomes inside customers' plants are unclear because integration with customer maintenance systems is incomplete. Year 2 board review: program restructured at much smaller scope. ~$12M effectively written off. Sound similar to a known industrial IoT cautionary tale.

Spend by Month 18: $16MConnected Units: 6,000 (target was 25,000)Service Revenue Lift: NegligibleStrategic Outcome: Restructure + writedown

Focused Phase 1: $3M to instrument the top 200 high-uptime customers with predictive maintenance for ONE equipment line. Pair with explicit service tier pricing.Reveal

Months 1-9: instrumented 180 of 200 target customers. Built integration with the 3 most-used customer maintenance systems (SAP PM, Maximo, IBM Maximo Cloud). By month 12, 45 customers signed up for paid 'connected service' tier at $X0K/year average. Reduced their unplanned downtime measurably. By month 18, you have 90 paying customers, $4M in new annual service revenue, and a clear playbook. Phase 2 ($8M) scales to next 1,000 customers and another equipment line in months 19-30. Total program spend by month 30: $11M, with measurable revenue lift, real customer references, and a defensible competitive position.

Phase 1 Spend: $3MPaying Customers (Month 18): 90Service Revenue Lift (Annual): $4M and growingStrategic Outcome: Replicable scaling model

Fast-follower: license a third-party connected-equipment platform, white-label it, focus internal investment on integration and go-to-market.Reveal

Months 1-12: licensed platform live, modest customization. Faster time-to-market than building. But: data ownership questions, vendor lock-in concerns, less differentiation vs other manufacturers using the same platform. By month 24, you have 4,000 connected units and $1.5M in service revenue. Reasonable middle-ground outcome but doesn't establish strong competitive moat. Risk: if the platform vendor gets acquired or pivots, you're exposed.

Time to Market: Faster than buildDifferentiation: ModestVendor Risk: Real

Related concepts