Where to apply Data Analytics in Water Utilities: Operations
15 August, 2018What’s Industry 4.0 concept? A brief and understanding explanation
16 August, 2018
Theory is quite good sometimes. Previous posts about this subject, Data Analytics applied to Water Utilities, have been possibly too generalist. But they are necessary if we want to understand the recommendations I’m going to provide you here. It’s time to go deep in then.
We know why already. We also know the potential benefits. Let’s talk about how to and where I’m going to apply such a valuable technology. In despite the high detail of the exercise, I don’t pretend to provide a bible or complete list of things to promote into the Water Utilities to become a Data-Driven organization, step by step. This is more like a guide to everybody who wants to begin to draw a consistent strategy with any background about Data Analytics. This would be the right way to start the needed deeper change in the organization to guarantee the success of any Data-oriented project.
So, to make my choice among all the possible cases I’ve balanced two basic topics: the potential value for society and environment of each one, and their “time-to-market” considering the state of the art of the technology behind. This is, try to find what could be more valuable first for the company. Fantasies later.
I’ve spoken in some other posts about how hard is move these data initiatives forward without a change in persons, processes or tools. But reality is even cruel. Water industry doesn’t have the proper data to start to work with. In fact, we don’t have enough sensitivity to understand we’re not gathering, treating and loading the necessary data to response all the questions I’ve prepared for each case. And that’s the problem’s origin we need to stop immediately. So, perhaps instead of investing a huge amount of money in revolutionary and promising data initiatives, it’s time to look at the state of your data and put the money on better collection processes, data enhancement initiatives and data governance strategies.
So, in this post we’ll focus on the Water Utility billing and commercial side, including billing, metering and customer care.
BILLING
Regarding Billing matters, the main efforts should revolve around analyzing, comparing and predicting it, analyzing the debt, predicting fraud and optimizing tariff structure.
Recommended actions for DESCRIPTIVE analysis – What happened?
Account analysis
- Key indicators
- Customer KPIs – 360 vision
- Key questions
- Who's behind the contracts? Who's really my client?
- Can I segment my client based on the information they're providing me? Do we have enough data to do that?
- How are my customers divided by usage? And according the different typologies that I've created?
- Where are the most critical ones? Are they well protected?
Non-payment reduction and analysis
- Key indicators
- Key questions
- Where and when is the bill non-payment concentrated? Who is affecting mainly? Can I correctly characterize the subscription type?
- Is there a relationship between the service condition and non-payment?
- Is non-payment predictable? How will it evolve in the future?
- How should I adapt my customer relationship to reduce debt and helping them at the same time?
Recommended actions for PREDICTIVE analysis – What will happen?
Anomalous consumption and fraud forecasting
- Key indicators
- Water consumption
- Revenue Water
- Number of inspections
- Key questions
- What is the standard consumption pattern of each subscriber? Can I detect anomalies?
- Is the anomaly a technical problem or a fraud?
- Do I need to go and inspect it? How does the inspection volume affect my daily operations? May reduce the number?
- Which is the impact of the "anomalous consumptions" over the entire registered consumption?
Billing forecast
- Key indicators
- Key questions
- What's the expected current period turnover? And for the next one?
- What number of additions and withdrawals may anticipate in the future?
- Can I segment it for special uses or tariffs?
- Is there any obvious trend in that evolution? Should I take any special action?
Recommended actions for PRESCRIPTIVE analysis – What should I do to make it happen?
Billing process optimization
- Key indicators
- Error rate in billing process
- Key questions
- What error rate can I expect for the current billing process? Is it increasing or decreasing over time?
- Where are these errors concentrated? What impact do they have on the billing operation?
- How do they relate to operational aspects? And with the metering process?
- How can I reduce the error rate before launching the billing process?
Tariff simulation
- Key indicators
- Tariff structure
- Collected volume
- Key questions
- How can I structure my tariff to ensure the highest service quality while minimizing the impact for the customer?
- What different alternatives may I check?
- What are the factors that I can handle to build it?
METERING
Regarding Metering matters, the main efforts should revolve around analyzing and comparing water consumptions, analyzing and segmenting my clients, optimizing meter management and optimizing measurement strategy.
Recommended actions for DESCRIPTIVE analysis – What happened?
Water consumption analysis
- Key indicators
- Overall water consumption
- Key questions
- How is the water consumption distributed geographically and temporary along the net? Where's the demand concentrated?
- What are my target clients? How can I characterize their behavior and group them?
- How is water consumption currently distributed among these groups? And how will be in the future?
- What's the average consumption per capita? And by neighborhood? And by type?
Data collection provider analysis and recommendation
- Key indicators
- Valid manual readings rate
- Key questions
- How is the manual metering process distributed? Where are the worst success rates?
- What's the cost associated with inspections of each manual metering supplier? What's the quality services KPIs forecast? Should I take any decision in short-tem? According to my meter strategy, is there space for improvement in the process of assigning metering routes?
Recommended actions for DIAGNOSTIC analysis – Why it happened?
Smart metering adoption analysis
- Key indicators
- Smart meter readings overall rate
- Key questions
- What's the current smart metering deployment?
- Is it working correctly? Is the service acceptable?
- What usage am I doing to available data? Can I easy calculate the smart metering ROI?
- How can I increase it?
Meter data collection methods comparative analysis
- Key indicators
- Usage by data collection method
- Economic cost associated by data collection method
- Key questions
- What types of metering method am I using? What performance and derived service quality do I get from each one currently? And what is the expected evolution?
- Have I optimized the manual metering process by smart task dispatching? Do I have space for improvement to apply it?
- When does it make sense to switch to smart metering?
- What's the business case that could optimize my meter operation using different methods? Which is the right plan to start with?
Recommended actions for PREDICTIVE analysis – What will happen?
IoT sensors strategy analysis and comparison
- Key indicators
- Facilities sensorization cost
- Key questions
- Depending on my loT sensors strategy, which sensors should I buy?
- What tests should I do before making the purchase effective? May I use installation historical data to reinforce the process?
- What factors are most relevant when making the purchase decision? Can I change them?
- What's the estimated investment for the coming years?
Measurement error rate forecasting
- Key indicators
- Error rate by method
- Quality and feasibility rate of obtained measures
- Key questions
- What's the current average error rate in the measurements obtained? Is it increasing or decreasing over time?
- What impact does it have on operation and maintenance? Is the associated performance acceptable? When should I act to address it to acceptable levels?
- According to the deployment and replacement strategy, what error forecasting will have in the future?
- Are they calibrated correctly? How should I do to ensure the right data quality
Recommended actions for PRESCRIPTIVE analysis – What should I do to make it happen?
IoT sensors location strategy optimization
- Key indicators
- Key questions
- In order to maximize ROI and ROA, where should I have to install sensors along my networks? And in my plants and other facilities?
- How should I progressively apply this sensing strategy? How many phases should I execute?
- What type of assets should I monitor first? And subsequently?
- What's the impact on my capital investment planning?
Data collection process optimization
- Key indicators
- Reading error rate
- Volume of inspections due to errors
- Meter reading process overall cost
- Key questions
- How often am I collecting meter data manually? Is this process planned efficiently? May I do better?
- What's the cost of the associated inspections after each data collection campaigns? Is it high compared to entire process?
- What would be the impact if I change the data collection plan? In short-term? And what about midterm?
- What's the cost over customer satisfaction of this disruptive change? How should I modify my tariff structure to minimize it?
Meter inventory optimization
- Key indicators
- Meter aging and condition
- Key questions
- How can I fit meter typology and characteristics to my customer use and typology to maximize its technical performance?
- How often should I change the customer meter?
- How can I optimize the purchase of devices to fulfill my strategy requirements?
- How may I plan a replacement campaign to maximize operational efficiency?
CUSTOMER CARE
Regarding Customer Care matters, the main efforts should revolve around analyzing and predicting customer behavior and optimizing my relation channels based on interaction analysis and segmentation
Recommended actions for DESCRIPTIVE analysis – What happened?
Customer care service analysis
- Key indicators
- Service quality
- Communication volumes by channel
- Key questions
- How am I relating with my customers? And with those who are still not?
- What are the most accepted communication channels? How much do people use each channel? To do what?
- Do we respond well on time and correctly? Am I doing it right or the service is every day worst?
- Are my clients happy? How can I know that?
Behavioral segmentation
- Key indicators
- Customer activity
- Customer consumption
- Key questions
- Regarding water consumption, what's the behavior of each of my customers and how is he connecting with me? Can I group them?
- How should I approach each of these groups in order to increase their satisfaction?
- What actions and channels should I use?
- What kind of recommendations can I make?
Recommended actions for PREDICTIVE analysis – What will happen?
Customer satisfaction forecasting and optimization
- Key indicators
- Customer satisfaction level
- Key questions
- According to the contact historical data, how should my customer perceive us?
- How do affect network condition, service affections and other operational aspects in their mood about us?
- Can I find out their mood during calls using cognitive services (voice)? Or by what it says in social networks?
- What are the recommended actions I can take to make my customer happier?
Recommended actions for PRESCRIPTIVE analysis – What should I do to make it happen?
Multi-Channel customer service model 7 optimization
- Key questions
- How should I change to adapt the way I communicate with my customer expectations?
- What other channels may I use? Which is the best that fit my reality? And customer's reality?
- How can I make this transformation without a negative impact to the service?
- What's the expected economic? How does that fit into my CIP?
