Top 3 secrets to enterprise IoT project success from industry titans
One of my favorite aspects of my job is the opportunity to connect with people who are developing novel approaches to address major enterprise challenges. About once a month, I host a discussion panel with technology leaders from around the country who have successfully implemented IoT and analytics systems. The goal of these panels is to extract actionable insights from their experiences through interactive discussions and to build relationships between organizations who are tackling similar challenges. Our guests include executive leaders from companies such as Oracle, Ericsson, Hitachi, AT&T, GE, Verizon and many others.
As these discussions have progressed, one thing is clear: The path to IoT success is not one size fits all, but there are some common elements to the journey. Here are three themes that have come up in every single discussion with titans of the IoT industry.
1. Think big, start small
Most projects start with ambitious top-level goals, but it’s important to start your organization’s IoT efforts with one challenge to address. Focusing on one problem or inefficiency relating to a specific process will enable your organization to craft a realistic IoT strategy and allocate the right resources as you walk through a starter project. Begin by selecting one process that could be improved through IoT and analyze it from several angles:
- Inputs and outputs: What are the major inputs to and outputs from the process?
- Bottlenecks: Where are the bottlenecks in the process? Do the bottlenecks arise from human error, lack of resources, lack of documentation and/or an inefficient process?
- Financial value: Where in the process is money is made and/or lost?
- Time value: How long does it takes to move from one step of the process to another? Is there any part of the process where things seem to slow down? Can you identify alternative approaches to speed up the process?
- Automation: What could be automated?
- Insights value: Where in the process could you collect additional data that would help improve the product development cycle or customer experience?
The answers to these questions should be the focus of everything that follows, especially the determination of what data to collect. Don’t try to boil the ocean. There are many kinds of data that could be useful, but only some kinds that will help advance the vision. With thousands of IoT sensors generating data in real time, it’s easy to get overwhelmed with data volume and complexity. Collecting data on everything can lead to massive data volumes, as well as security and compliance risks. Every piece of information collected should provide value toward your organization’s top-level goals. That value needs to be weighed against the risk of collecting it. Showing early success by addressing a major organizational challenge will increase your organization’s executive leadership’s team confidence in and commitment to the IoT initiative.
2. Reimagine your team
Diverse skills and perspectives are crucial to the success of your IoT project. While many of the skill sets can be addressed with existing resources after internal development and training, it’s likely you’ll need to also bring in new talent. Almost every executive with whom we spoke stressed the need to hire external resources — both permanent hires and consultants — and build new capabilities in the organization. Some key roles include data scientists and architects, database administrators, advanced networking and cloud resources, data security experts, business analysts and customer experience resources.
Investing time at the beginning of the project to outline the roles and skills required for your project can prevent costly delays resulting from resource gaps.
3. Iterate indefinitely
There’s no such thing as a fix-it-and-forget-it IoT platform. It’s a commitment to continuous improvement. Finding and refining the value in your data is like discovering the perfect gem. You have some idea where to start looking, so you deploy sensors with a basic framework of descriptive analytics. That may lead to rocks or diamonds in the rough, but you won’t know until you polish and refine them. Sort through the data using various lenses to discover patterns by intervals of time, geographies, demographics and so forth. Sometimes that means developing new algorithms to test the different patterns. As insights become clearer, you need to refine those algorithms and begin moving them closer to the digital edge where the data is being generated and consumed. As you continue to hone your findings with predictive analytics and machine learning algorithms, the path to value becomes clearer.
For example, a tire manufacturer initially deployed sensors in its tires on how road conditions and braking patterns were affecting tire longevity. But over time, the tire manufacturer used that data to strengthen its relationship with fleet owners by alerting them before tires needed to be replaced. This information was incredibly valuable to fleet owners because it reduce the risk of accidents caused by blowouts and reduced the amount of expensive downtime for the trucks. Through these insights, the tire manufacturer was able to measurably strengthen its relationship with its partners and improve its market position.
The iterative process also needs to include refining the underlying IT architecture for your IoT system. It’s not practical to haul massive amounts of data back to a central location for processing. That’s especially true for latency-sensitive use cases that demand urgency, such as emergency response or healthcare, or ones with high volumes of data exchanged between sensors and systems, such as in aviation or manufacturing. Avoiding “slow is the new down” means moving computing power and analytics to the digital edge — closer to the people, things, clouds and ecosystems that are generating and using the data.
A constantly evolving edge
If done right, your IoT platform should get smarter and yield better insights over time. Staying focused, including the right people and continually refining will help you evolve it from being merely descriptive to being contextually aware.
Who wouldn’t want a smart IoT system that can find the next treasure trove of competitive advantage for your business?
All IoT Agenda network contributors are responsible for the content and accuracy of their posts. Opinions are of the writers and do not necessarily convey the thoughts of IoT Agenda.