Delta-EE Research Blog

Keep up to date with our latest news and research by subscribing to the Delta-EE blog - just click on the button below.

Subscribe to our blog

Why electrical demand disaggregation could help personalise the energy customer experience

In my last article I looked at the wide range of insights being deployed by energy companies as they seek to capture the benefits of improved customer engagement.

Many businesses already use basic intelligence to help their customers better understand their consumption at a high level. However in recent years we have seen a rise in the use of demand disaggregation techniques (DD), which take insights to a more granular level.

DD methods aim to provide a greater depth of insight into the energy profiles and ‘energy lifestyles’ of customers, by inferring the demand associated with individual appliances.

When Delta-ee first started investigating Home Energy Management (HEM) back in 2010, the energy sector had only just started considering the benefits of offering energy monitoring devices to customers. These devices displayed instantaneous and cumulative energy use for the day, but did little more. While some customers found these high level insights interesting, these early innovations lacked the functionality to be truly compelling. A classic case of ‘two weeks to kitchen drawer.’

Smart kitchen

Few utilities pursued DD at this time, simply because of the low consumer interest in in-home displays and the fact most available solutions suffered from low accuracy when it came to appliance detection.

In 2017 electricity demand disaggregation is coming of age

We recently revisited the DD market and can report that the game has changed. Our new report on Electrical Demand Disaggregation demonstrates just how much the market for this particular type of energy insights has grown.

 

Here are 5 takeaways from our report:

  1. The market has grown beyond a tiny niche: there are now millions of users with disaggregation solutions. Solution providers such as Bidgely and ONZO have been at the forefront of this development.
  2. There is rising competition amongst solution providers: there are now many competing solution providers in this space and we are currently benchmarking the top 20 companies. Although there are significant differences between solution providers. From a hardware perspective, some are reliant on smart meters, some use clamps, while others are using their own innovations. Data analysis approaches also differ. While most utilise machine learning to better identify the ‘demand fingerprint’ of appliances, some players look to improve accuracy by blending in manual user-defined information from the very start. Interestingly, a few solutions encourage user interaction through gamification techniques.
  3. Accuracy is improving: as with any modelling exercise, it is difficult if not impossible to achieve 100% accuracy. Instead some providers are using careful customer messaging to cope with the inevitable level of inaccuracy. For example, instead of using a chart that shows that a customer used their washing machine at 8am last Tuesday (when with a 10% error rate some customers may know they didn’t use it), telling the customer that the machine was used 20 times last month (say, compared to 16 times the previous month) could still be a meaningful insight even with a 10% margin of error.
  4. Costs are falling fast: although price models do vary significantly, we are seeing prices fall, especially for use cases that benefit from sub-second data (where a special device in the premises is required). 
  5. Several key customer use cases are emerging: assisted living and preventative maintenance are two of the most compelling uses for DD. However, DD may have a wide number of applications.  For example, predicting whether a coffee machine might need new pods based on how many times it has been used, offers a compelling e-commerce use case. Fault diagnostics might be sold as an appliance insurance wrapper, or to encourage the purchase of repair services, or even a new appliance.

It is clear that electrical demand disaggregation has now entered the mainstream of customer engagement tactics in the energy sector. Most energy suppliers are known to have trialled a solution or have it firmly on their innovation radar. Indeed some, such as Eneco, have even developed in-house solutions.  

Is DD a way to personalise energy experience?

 

At Delta-ee we believe demand disaggregation technology has now evolved to the point where it should be taken seriously by energy companies. Indeed, as techniques develop and models increase in accuracy, it may have the potential to personalise the energy experience in a more compelling way than we’ve seen to date. 

The trick will be finding ways to use demand disaggregation to help people care more about what their consumption looks like and act on this insight. People might act to save money, or even to rectify potential problems with appliances before they become defective. Arguably, the technology will need to become more accurate, but if suppliers can encourage people to submit data then the demand disaggregation estimates will undoubtedly become more powerful.

Increased accuracy will improve personalisation, but this depends on suppliers scaling demand disaggregation to a point where they have enough ‘customer observations’ to decrease their margin of error. Ultimately, those that roll-out the technology to the largest consumer bases are likely to develop the most compelling solutions in the long run.

 

If you would like to learn more about Delta-ee’s Electrical Demand Disaggregation report, please contact us on +44 131 285 1773

James Miller is Principal Customer Strategy & Data Analytics at Delta Energy & Environmentand leads Delta-ee’s Customer Data Analytics Service.

Vertical unlimited: 6 success factors for vertical...
All change in channels to market for heat?

Related Posts

By accepting you will be accessing a service provided by a third-party external to https://www.delta-ee.com/

Search Downloads

Delta-ee Bloggers

Andy Bradley
18 posts
Jan Hughes
98 posts
Stephen Harkin
7 posts
Jennifer Arran
13 posts
Jon Slowe
28 posts
Lindsay Sugden
22 posts
Steven Ashurst
15 posts
John Murray
17 posts
Nigel Timperley
2 posts
Arthur Jouannic
20 posts