The anatomy of an IoT solution: Oil, data and the humble washing machine

By and Karen Lewis | 15 minute read | October 17, 2016

The anatomy of an IoT solution: Oil, data and the humble washing machine

A lot of people think about data as the new gold, but a better analogy is data is the new oil. When oil comes out of the ground it is raw, it has intrinsic value but until that oil is refined into petrol and diesel, its true value is not gained. Data from sensors is very similar to oil. The data that comes from the sensor is raw, to gain insight from it, the data needs to be refined. Refining the data is at the heart of a successful Internet of Things project which leads to business growth and transformation.

There are many examples of how data can be used to gain value in different ways. Let’s use the humble washing machine as a use case. At first glance the washing machine does not look like an interesting use case, but when examined more closely it turns out to be a great way to illustrate how IoT works by showing us how value can be gained from the data in a washing machine; and how its data can be used for business purposes. The use cases for gaining value from washing machine data are not specific to the washing machine, they can be applied across multiple “things” in all sectors and industries.

The humble washing machine: the missing link

Consider the washing machine when it is manufactured and goes out the factory door, the manufacturer loses sight of that washing machine. Once the machine goes through the distribution, reseller and sales chain, the manufacturer has lost sight and does not know who ends up owning or using the washing machine.

One of the first benefits of a connected washing machine (which is the definition of an Internet of Things thing) is where data can flow. With a connected machine, a manufacturer now has the ability to communicate with the owner or user of that washing machine. The connection between a manufacturer and the end user of a product is exceedingly valuable for the chief marketing officer of a company. CMOs want to interact with the owner or the user of the thing they produce. The value lies in the fact that the manufacturer has a direct touch point with the consumer – a way to talk to or connect with the owner or the user of the thing. The value of this touch point is used in many of the following ways.

Better customer service

The second benefit becomes evident when a washing machine breaks. Today if a washing machine breaks, typically a consumer will do one of two things. If it is not in warranty, the broken machine might be discarded and replaced with a new model. If it is in warranty, a consumer might call out the service engineer who will then come around to investigate the problem. While on site the service engineer makes a diagnosis, perhaps discovering the pump is at fault. The engineer goes out to the white van only to realize the right pump is not in the van. Returning to the owner they apologize and agree to return once the part is in stock. The engineer leaves, orders the pump, and two weeks later returns to repair the washing machine and gets it going again.

In this scenario, the result is satisfactory as the washing machine ends up working, but the owner or the user is not happy as their machine has been out of action for two weeks. No doubt the absence of a washing machine is a difficult situation for a homeowner; but, for the owner of a commercial washing machine in a hospital or launderette, it’s an even worse situation. And what about the service company? The field engineer had to schedule two trips in order to complete one job, has used two lots of labor, two lots of fuel, plus spent time addressing extra administrative tasks.

In the case of a connected washing machine all these wasted cycles could be avoided. The connected washing machine already has a lot of data in it – temperatures, pressures, cycles, lime scale build up, parts wear and tear, etc. That’s how the washing machine works. It uses the sensor data that is already in the washing machine to control the wash cycles. Until the washing machine is connected the data is siloed, and is only of use for controlling the machine. Once connected, the washing machine’s data can flow and be refined using analytic capability.

Predictive analytics to predictive maintenance

The data from the washing machine can be compared to a computer model of the washing machine. If the washing machine data correlates then life is good, the washing machine is working well. However, if trends or anomalies are detected which deviate from the model, predictions can be made about what might happen in the future. If a problem can be predicted, an action can be taken to mitigate the problem. Take the example given earlier. If the pump breaking is predicted preventative actions can be put taken. First, ensure the right pump is in stock and order it; once in stock, contact the owner or user of the machine to arrange a convenient maintenance slot to undertake the repair. The service engineer then visits the site with the right parts at a convenient time, resolves the problem before it occurs, thereby providing a better level of service which in turn goes a long way to improving customer satisfaction because the machine never breaks, and preventive maintenance takes place at a convenient time.

In a predictive world efficient maintenance leads to a better customer experience

In the new predictive world both the user of the washing machine and the service company benefit. The user’s satisfaction increases as the washing machine is only out of action for a short, convenient maintenance window, rather than days or weeks in the pre-predictive world. The service company benefits as only a single visit is required to fix the problem rather than two visits. The use of a connected washing machine, and the application of the data flowing from that washing machine enables the optimization of multiple processes – call outs, field engineer time, van mileage, inventory management– creating savings in resource, time, money, and asset use. These savings result in more opportunities to grow the service company’s business– using the field engineer and expensive assets to do more jobs with using the same assets, for example, the person, the field engineer and his van.

From predicative analytics to predictive engineering

In the scenario to predict problems, the class of data used – temperatures, pressure, and etcetera– is telemetry data. But there’s another interesting class of data that can be collected once a thing is connected, and that is usage data – how is somebody using something? For example, the average washing machine has ten or more wash cycles. Are all of these wash programs used? Through a simple polling of a number of people visiting the IBM IoT showcase lab around 99.5% of the visitors use one, two or three programs. These findings suggest many of the programs are not used or useful.

In the instance of a connected washing machine, not only is that machine sending telemetry data, but it can also send information about which programs are actually being used. Once that data is received and analyzed (and we don’t know what the result will be), hypothetically if the manufacturer was to find out that 99.9% of users use the same three wash programs, that knowledge is exceedingly valuable. Now when they build the next version of the washing machine, they can get rid of a number of redundant programs, dramatically reducing the design, engineering and test costs. For the consumer, the washing machine is made easier to use – with fewer choices, and a simplified set of settings.

Telemetry data is what most users think about when looking at IoT, but usage data is just as important. Both telemetry and usage data combined can make a difference to continuous engineering cycles when building the next version of a product. It’s interesting to note that it applies just as much to software as ides does hardware.

Expanding the value chain with data as a service

The next step involves how individuals and organizations interact with the things in different ways. So far the owner or the user of the washing machine, the manufacturer, and the service company have benefited from the connected washing machine; but, there are many other people and companies that might like to interact with the washing machine, and use the data from the washing machine.

The first example is an insurance company. Insurance companies are all about pricing things based on risk. The connected washing machine lowers the risk. With predictive capability the chances of a washing machine breaking and flooding the kitchen floor are much lower.

If the insurance company knows it is a highly reliable washing machine that will not flood the kitchen floor, they may be amenable to lowering the price of home insurance. The more reliable white goods and connected security systems are, the more amenable the insurance company might be to lowering the price of the premium. It’s very similar to pay-as-you-drive and pay-for-how-you-drive insurance, but applied to home insurance.

Another interested party might be the energy suppliers. Energy suppliers would like to be able to interact with the washing machine to say please don’t come on at peak periods when energy is at a premium, but you can come on at any other point in time. This helps spread the load across the day.

There are washing powder companies who would love to have something a little bit like the Amazon dash button but actually go the next step. The washing machine knows how much powder it is using and can request the powder be replenished before the current supply runs out. The user of the washing machine does not need to think about getting the washing powder and the supplier gains stickiness with their customers. Again, this is not a new idea, it’s a business model which has been used successfully with printer ink for some time. With a connected washing machine, the same model can be applied.

Another example involves fraud or more specifically the prevention of fraud. Take the example where an individual raises a claim regarding a washing machine that has flooded the kitchen floor. The claim is submitted requesting a sum of money that replaces the machine, and damage to the floor. But, what if it wasn’t really the washing machine that caused the flood? What if it really was the fault of the owner who left the sink or bath tap turned on, with the plug in, and it was that which flooded the kitchen floor. In this instance, the claimant who does not have accidental damage would not be covered if they claimed for leaving the tap turned on. To ensure coverage, the claimant submits a fraudulent claim.

A connected washing machine can help the insurance company detect fraudulent claims by providing data to indicate that the washing machine really is broken. The insurance company benefits as it helps reduce fraud, minimizing false claims and reducing the costs associated with insurance fraud. In turn, the consumer benefits as less fraud means less risk which enables the insurance premium to be lowered.

Data diffusion made possible by API management

Suddenly, a lot of people would like to get to interact with an intelligent washing machine, in particular the data from the washing machine. But how can the data be made available to different companies? This is where the API economy and API management comes into its own providing controlled access to the data. Controlled access determines who can access the data, when and how often access is allowed, and just as importantly, how much it will cost to access the data.

As the data has so much value providing data-as-a-service is one approach to help a business expand. The API economy enables data to be shared between departments, with partners and to other businesses. Whether access to the data is directly monetized is a business decision.

Moving from marketing obsolescence to the ‘as-a-service’ model

Today, a consumer typically buys a washing machine, uses it until it breaks, then buys another washing machine. This is a familiar model of business across many industries from technology, automotive, electronics. The world is changing, moving to a pay for what is used model. Going forward, a new potential model for the washing machine is one where the consumer is given a washing machine as a-pay-as-you-use machine, supported by a pay-per-wash plan. Every time the machine detects a new wash load, the consumer is charged a small amount – a small number of pence, cents or euros per wash cycle. While it sounds radical, the model is not entirely new. In the UK, there was a company called Radio Rentals that rented washing machines and charged the user on a monthly basis. It was a common model before everybody could afford to have their own washing machine. This is just a re-spin on that old model but changing the way the amortization occurs – rather than monthly for the apparatus, it’s now pay per wash, based on usage.

The change to pay per use may have interesting knock on impacts. Typically, a washing machine is designed to last approximately five to seven years. With the prospect of the service company giving you the washing machine with a pay per use plan, the service company now wants the machine to last for as long as possible in order to ensure a healthy return on investment. Looking to the future this type of model can help push a design ethos change which encourages the manufacture of things which will last for a longer time. By improving the longevity of the washing machine, or any things, the amount of accumulating waste could be reduced, resulting in immense and positive impact on the planet.

Data flowing in both directions

For the use cases discussed, most of the data is flowing from the washing machine up to something like a Watson IoT Platform, into the cloud where it can be analyzed. However, the beauty of IoT is it can flow in both directions – from the washing machine to the cloud, and from the cloud to the washing machine. But why send data to the washing machine?

The data could take the form of a software update, or a security fix. If a problem is detected with the washing machine, a software update can be created and pushed to the washing machine. More interestingly, as long as there is spare capacity within the machine new features such as a new wash program can be developed and pushed out. In order to recoup the development cost of the new feature the manufacturer or service company can charge more for the new wash program.

The ability to deploy fixes and new features over the air to existing things changes the whole built in obsolescence business model where devices are designed to only last a short time before they break and you are forced to buy a new one. If new features can be pushed out there, there may no longer be a need to buy a new thing in order to take advantage of the latest feature. In the new model, the latest features can be pushed out to the existing thing. A great example of this model in action is the way Tesla cars are pushing out new features such as battery optimization and self-drive features to their cars. In addition to over-the-air software updates, there is an emerging trend for modular hardware that enables parts of the hardware to be upgraded.

What did we learn from the humble washing machine?

  • A connected washing machine gives the manufacturer a touch point to communicate to the end user.
  • Telemetry data can be used to determine whether something’s going to break before it actually breaks.
  • Preventive maintenance can now be used to take preventive actions that can stop problems occurring, or to enable a service department to fix a problem in a convenient maintenance window with limited disruption to the thing’s service.
  • Usage data, a new class of data, can now be accessed and used to understand user experiences and interactions.
  • Continuous engineering can be used to optimize future designs, creating improvements based on real user experiences.
  • New updates and features can be added based on interests and usage patterns thereby increasing loyalty and satisfaction.
  • Data as a service opportunities can be identified and used to better enable a wider set of stake holders across the value – different companies, partners, and even different departments within an organization. There are several organizations that stand to benefit from interacting with the data from that washing machine, ranging from utilities, insurance, washing powder–all of which have a vested interest in the connected washing machine and its user.
  • New business transformation opportunities can be explored whereby the washing machine itself can be turned into a service where the thing is not purchased, but given to the user, with the user adopting a pay per wash experience.
  • A new model for releases and updates can be established allowing the user to receive new features, security updates, and fixes as unobtrusive downloads into the machine’s programming interface, not only giving the machine a longer shelf life without compromising relevance or modernity, but paving the way for more sustainable practices in the future.
  • Exploration in this direction can lead to greener opportunities with wider ramifications for waste and redundancy, where suddenly machine obsolesce is not the end game, but longevity emerges as a better.

The humble washing machine as depicted here is just one example. The use cases explored in the article are applicable across multiple industries, across different types of things. Going forward, the challenge for developers, engineers and designers will be to explore how and where these features and ideas can be used in new devices and solutions, or, perhaps retrofitted in existing things.

The IBM IoT technology and cloud portfolio provides the capability to implement all the uses cases described. To help get started here are a few links:

Learn more