Combining data and domain to transform collections and generate value


The utilities industry is experiencing a period of unprecedented change. Demand for ultra-customer centricity, the rise of prosumers demanding electricity flows from and to the grid, technology advancements, digital disruption, the entry of new sets of players from within and outside the industry, and increasing regulatory mandates are causing immense pressure. Energy and utility organisations are seeing an adverse impact on revenue growth of roughly -4.4%, as well a slowdown in profit growth of 1.2%1.


One of the key issues affecting poor performance for utilities is increasing consumer debt:

  • Nearly £400m is owed to energy companies, with 2.93 million UK households with energy debt owing an average of £134 each2
  • Bad debt charge as a percentage of revenue is 3.2% for water companies and 1.5% for energy utilities3
  • Days sale outstanding (DSO) levels have gone up from 27 days to 32 days, impacting working capital4
  • It is estimated that bad debt currently adds approximately £21 annually to each customer’s water bill5

While all major utilities have invested heavily in credit and collections activities, the current difficult background of regulatory restrictions means that all the easy wins have been taken. Utilities are now struggling to balance between recovering debts while ensuring customer experience does not take a massive hit. A new report from Citizens Advice5 suggests a number of common problems experienced by customers who are behind on their energy bills:

  • The energy suppliers’ approach to debt collection is often seen as aggressive, which can make people’s problems worse
  • People in vulnerable circumstances face specific barriers to engaging with their suppliers
  • Vulnerable customers are unlikely to engage with support unless it is clearly framed as a way out of their problems

The above signs are good indicator that utilities need to reimagine their collections processes and put the customer at the center of their strategy. Integrating domain expertise with advanced data models and digital technologies can reduce bad debt charges and the cost of debt collections while improving customer experience.

Smart Collections Framework for Utilities

Debt should not be treated as a standalone event. Instead, debt should be integrated across the customer journey, from acquisition to leave, to prevent debt or optimise debt collections.

This would include taking preventive measures to avoid customers getting into debt rather than solely focusing on debt collection optimisation, all while focusing on providing a consistent and fair customer experience and regulatory compliance.

EXL’s Smart Collections framework for utilities focusses on debt prevention and optimisation by putting the customer at the heart of the strategy.




1.A Prevention – Harnessing the Power of Data

While traditional debt collection processes focus on the cost of collection and collection efficiency, a smaller upfront investment to prevent debt will have a significantly higher ROI. Utilities can focus on the following areas to prevent customers going into debt:

  • Target the right new customers: Using advanced modelling techniques, utilities can identify the right customers to go after based on variables including lifetime value, demographics, potential for delinquency, and others. This analysis can inform the marketing teams driving the acquisition efforts.
  • Tailor credit decisioning strategies: Onboarding new consumer and business clients quickly, securely, with accurate identification and fraud checks, and a greater knowledge of each applicant’s circumstances can positively tailor the relationship to minimise risk using appropriate credit decisioning strategies. Fraud checks and accurate identification can be automated using cognitive BOTs to reduce cost and improve turnaround time. The resulting faster processing, coupled with machine learning algorithms, can improve decisions in terms of offering pre-pay meters, appropriate pricing and deposits, and offering direct debit as the only payment option, based on the risk profile of the customer
  • Upfront comprehensive data collection: In an increasingly competitive market, it is tempting for sales teams to sign up as many customers as possible using only limited customer and financial data, ignoring any barriers that may deter new customers. Unfortunately, this short-term thinking often causes significant repercussions down the line. It is therefore crucial that utilities obtain the correct information about new customers, both from the customer themselves as well as from multiple external data sources. This information should include at least the following:
    • Double checking the correct spelling of the customer’s or the company's name
    • Getting landline and mobile numbers along with the correct email address
    • For business customers, identifying whether you are talking to the director of the company, who is liable as a corporate entity, identifying whether the customer is an individual trading in some form and therefore personally liable to pay the invoices, and other information
    • Clarifying whether the customer is a landlord, and if so, make it clear from the outset that the landlord is responsible for paying the bills
    • Identifying whether it is a student let or similar situation, which means that the occupier is likely to change regularly
    • Identifying potential cases of vulnerability
  • Start dealing with potential delinquency ahead of time through predictive modelling: Utilities should track a customer’s changing circumstances on an ongoing basis to adjust strategies based on any significant changes in the customer profile. By developing deep insights into a customer’s payment performance, preferences, and circumstances, companies can create customised treatment paths to work with high-risk customers, vulnerable customers or those with a change in financial circumstances.
    These models can determine the payment patterns that indicate that a customer is struggling. For example, it could start with being just a few days late or changing the payment mode from direct debit to cash or cheque. Any variation from the usual payment schedule should be a red flag for the system. A mechanism can be put in place that self-triggers when an unwanted pattern emerges. For example, this could look like identifying a customer moving from low-risk to high-risk, and incetivising them to move to a direct debit payment plan before debt starts to build up.


1.B Prevention – Building synergies between the front office and back office

To effectively use data across the customer journey, utilities should look at processes holistically instead of taking a traditional siloed approach and ensure all interactions improve customer experience. This will reduce failure demand, proactively address customer issues that might lead to debt, and reduce goodwill amounts given each month. Identifying issues proactively and fixing them before customers feel the impact of broken processes can reduce the bad debt charge and cost of collections:


2. Debt Collection – Customer Segmentation

By using advanced analytics and applying machine learning algorithms, utilities can move to a deeper, more nuanced understanding of their at-risk customers. To do this, utilities should first bring all data together to create a holistic view of the customer, this includes internal, external and engineered data.

With this more complex picture, customers can be classified into microsegments, and more effective interventions can be designed for them.

Using micro-segments will allow utilities to move away from making decisions based on static classifications. This can enable early identification of self-cure customers and a tailored approach based on value at risk, rather than blanket decisions based on standardised criteria. The aspiration is to have every customer as a “segment of one” with customized treatments.


3. Debt Collection – Contact Strategy

Customers contact preferences and responses are guided by personal considerations with little relation to the risk categories and contact protocols worked out by utilities. In this digital era, most customers prefer to be contacted through digital channels, while a smaller segment remains more responsive to traditional contact methods.

For example, while higher-balance customers are more likely to engage digital channels such as mobile and online banking, most utilities are contacting only low-risk customers in this way. Another example is where utilities contact their digital customers through letters instead of their preferred contact method like email or app notifications.

Using advance data modelling to better understand customers’ diverse preferences and deploying an always-on experimentation approach to contact strategy can improve connection rates and speed up debt collection. This requires investing in omni-channel capabilities to ensure consistent experience for customers across channels.

At a high level, this contact strategy consists of a four-step approach:

1. Identify which customers to contact first.Utilities should maximize their ROI by focusing on the most promising accounts. Using predictive analytics, companies can develop a model that allocates a likelihood score to each potential customer, and rank them in order for contact.

2. Identify best channel to contact. Understand customers channel preference instead of focusing the contact strategy based only on the risk score.


3. Content optimisation.
Use experimentation techniques like A/B testing to test effectiveness of content. The content of the messaging should be dependent on the riskiness of the customer. Effectiveness can also be increased through behavioral nudging, such as presenting high-risk customers with a late-fee waiver that they would lose by not making a payment. Framing the offer in this way, as a loss for a foregone payment, can be twice as effective as offering it as a reward for making a payment.

4. Identify the right time to contact. Using machine learning techniques, utilities can identify the best time to contact the customer. This can be based on historical connections for similar types of customers, demographics, and other variables. This can be further improved by using smart meter consumption data. The graphic depicts the utility trying to contact the business customer with debt and unable to reach the customer – using Smart consumption data, utility should change their contact strategy and reach the customer between 18:00 to 5:30)


4. Debt Collection – Identifying vulnerable customers and refine hardship programmes to mitigate loss
The majority of credit risk issues are focused on larger low income families, low income pensioners, households with unpredictable incomes and younger single-person households. In 2017, 18.2% of the population was aged 65 and over. It is projected that 20.7% will be 65 and over by 2027. The Office of National Statistics (ONS) also found that 18.9% of the population is under 16, and 12.2% of families have at least one dependent child. There are reportedly 13.9 million people living in the UK with a limiting mental or physical disability. It is also estimated that one million people will have dementia in the UK by 2025.7

This presents an opportunity for utilities to use data to better identify and improve the experience of customers in vulnerable circumstances. The drivers of customer vulnerability are many and varied. To tackle this issue, utilities need to focus on:

  • Understanding demographics: Utilities can use their own or publically available data sets to understand customer demographics. This can help identify geographical areas where there is a high density of vulnerability, target additional support accordingly, or inform emergency planning during an incident.
  • Data matching: In some circumstances, utilities can use third parties to augment their customer data. The third party can cross-reference their own data sets to verify and better identify customers who may be eligible for additional support.
  • Signposting: Using the various channels of customer engagement, energy companies can raise a customer’s awareness of additional support services available to them from their utility company, and vice versa. The customer is then in control of contacting the company to access their support services.

This proactive identification of vulnerable customers will help utilities to refine their hardship programmes and enroll more such customers to reduce bad debt charges while supporting customers in need.


5. Debt Collection – Recovery

There are some customers who are most likely to not going to pay anything despite collection efforts. Usually, these customers are deep in debt and unresponsive to any collection outreach. There may be variety of reason for such behavior, like the sudden loss of job, the intention of fraud, divorces, or other factors. Once the customer crosses point of no return, they simply cannot or will not pay.

Utilities must leverage machine learning and analytics to assess the cost associated with each default and determine the right recovery strategy. This could include an in-house recovery process, delegating recovery to collection agencies, litigation, or sale to third parties at a discounted price. Early identification can help companies to limit their credit exposure to minimise the losses.


6. Debt Collection – Reporting

One of the key issues challenging utilities is the availability of real-time data to make key credit and collections strategy decisions. Many organisations still collate information at a weekly or monthly level for performance reviews. Utilities should instead focus on a set of metrics that allow objective assessments for billing, payments, and collections on a near real-time basis using data visualisation tools. Some key metrics that stakeholders should review on a daily basis are DSO, bad debt charge as a percentage of revenue, unbilled debtor days, cash collection rate, doubtful debt as a percentage of net debtors, and percentage of customers whose data has been validated in the last 12 months.

Conclusion

The next generation of credit and collection strategies will be built around advanced analytics and machine learning. The transformation of collections has already begun, as industries assemble the data and develop algorithms to improve their existing collections within a few months. Utilities should embrace this change and combine the power of data, domain, and digital to meet the new delinquency challenges and potentially transform collections.

Sources:

  1. HfS Industry Blueprint: Utility Operations2018 - Revenue and profit for the top 50 publicly traded Utilities Company worldwide
  2. Guardian Article: Number of UK households in energy debt rises by 300,0000 (November 2018)
  3. PwC Report for Ofwat: Retail Services Efficiency benchmarking (September 2017)
  4. PwC Article: Sharp rise in bad debt dictates a need for utilities to re-think collection strategies (May 2016)
  5. Ofwat Article: Water companies must do more to address customer bad debt (September 2017)
  6. Citizens Advice: Ofgem urged to better support vulnerable energy customers in debt (May 2019)
  7. Ofgem: Draft Consumer Vulnerability Strategy 2025 (June 2019)


Author:

Rahul Arora
Energy & Utilities Practice Leader

Contact US