Everyone is talking about how IoT will revolutionise the way we do business, but before you get too caught up in the hype make sure you are able to implement IoT projects with measurable benefits.
This is the view of JP Smith, pre-sales director: Sub-Sahara Africa at Hitachi Vantara.
“Whether it’s a smart scale in our home which advises us how to adapt our eating and exercise regimes, or a smart shelf within a retail store that is able to ‘restock’ itself using sensors which automatically trigger stock orders as shelves empty, it’s easy to get excited about the possibilities the IoT revolution brings to both our homes and the workplace,” Smith says.
“It’s important though not to get too carried away by the current hype around IoT and make sure that the projects you want to implement are going to yield proper returns on investment. You might think you have an amazing idea but what are its actual business benefits?”
Smith says that to ensure your IoT project is a success, there are a few simple steps businesses can follow:
Identify your challenges
Before getting started make sure you understand the challenges you face both inside and outside your organisation. Generally, there are three areas you need to take into consideration, including people, technology and process.
It’s important that all your business stakeholders are aligned with what you are trying to achieve.
Another consideration would be your legal environment. Are there any pieces of legislation which you need to adhere to or which may prevent you from proceeding?
Before you start, make sure all your data sources are available and that you can use them. You also need to think about how you will anonymise that data.
Set realistic goals
As with any project, you will need to set goals and make sure you can realise them in a reasonable amount of time — this would be roughly one to six months for a starter project. When starting out your focus should be on solutions which are easy to implement but also offer high value.
Good examples of these types of projects might include offloading your data from a data warehouse and then enhancing that data with data from social media to gain a better understanding of your customer environment.
Avoid solutions that are complicated and offer low value as these will only get you in trouble.
Ultimately your aim will be to implement solutions which are more complex but also offer significant value, for example, predictive maintenance. You want to arrive in a space where you can not only predict what will happen but actually influence it.
Select relevant data sources
Once you have set your goals you can go about selecting the right data sources. These would include IT infrastructure or data centre infrastructure, but will also likely involve a lot of industrial information from devices such as cameras, heavy machinery, cars or the drivers of these cars.
Build the ecosystem
Your next step is to determine the ecosystem necessary to help you work with different data types and handle different speeds of data ingestion. Does your environment require you to handle data streams in milliseconds or are you able to process that data hours, or perhaps even a few days later?
Ultimately, your goal is to have these environments automated as much as possible such that they manage themselves, freeing you to focus on the business outcome of your solution. You’ll also want to try and reduce the time it takes to set up your IoT environment – so when introducing appliances think about easy management, integration and scalability.
To simplify things, you should implement data-in-place processing. Think about a lake containing petabytes of data — ideally you don’t want to have to move that data to an application and should rather move the application to your data. This not only saves you time, but will also make your life easier with regards to network load, infrastructure load and so on.
Integrate data sources
Once your ecosystem is in place, you need to start integrating various sources of data, such as the data from your warehouse and data from unstructured sources like social media. This will enable you to start building a 360-degree profile of your clients, and then combine that information with web analytics to understand your clients’ buying behaviour.
For example, we do this at Hitachi by using a platform called Lumada which can process data coming from a range of different devices and machines directly from where it is generated. Whether it’s from the data warehouse, a video camera or a data sensor, we are able to integrate that data into our Lumada platform. We then have different tools to analyse the data.
“Once you have these processes in place, you can begin to transform your business models, drive greater efficiencies, and ultimately, help promote an enhanced quality of life in whichever sector you operate,” says Smith.