Segmentation in Practice: Welfare to Work

Segmentation in Practice: Welfare to Work

          
Article image

Segmentation in Practice: Welfare to Work

  • Add To Interests
  • SAVE CONTENT
  • PRINT
  • PDF

  • Related Articles
    In This Article
    • Effective management of unemployment benefits and programs is a major challenge in the public sector, with over $500 billion spent annually on such efforts within the OECD.
    • Many countries have responded with welfare-to-work programs, which help people with disparate education levels and work histories find the right employment opportunities.
    • Segmentation—tailoring offerings to the needs of individual "customer" groups—is particularly well suited to improving the effectiveness of such programs.
     

    November 2013
    Using Data-Driven Segmentation to Transform Government Services
    Businesses have long used segmentation—assessing customer segments and tailoring their offerings to meet the needs of each—with great success. Now the approach is set to transform governments as well.

    Few issues facing governments are as challenging—or as important—as reducing unemployment. Unemployment is a stubborn issue in many countries, but the challenges differ depending on the specific country’s economy, employment opportunities, and skills base. Many rich countries are confronted by the twin perils of large budget deficits and high unemployment in the wake of the global financial crisis. (See Exhibit 1.) Developing countries, meanwhile, are just starting to shoulder the responsibility of supporting the unemployed and are searching for where to begin.

    exhibit

    Regardless of the differences, all nations face a similar imperative: to find gainful, sustainable employment for as many people as possible in order to drive economic growth, reduce benefit expenditures, increase tax revenues, and maintain social cohesion. Efforts to transform unemployment programs have proven challenging. The U.K. government has implemented no fewer than seven major employment programs over the past 17 years.

    For many countries, one response has been welfare-to-work programs, in which advisors meet with job seekers to help them find the right employment opportunities. With over $600 billion spent annually on unemployment benefits and programs within the OECD, improving the system for compensating the unemployed and moving them back to work can save billions of dollars, as well as boosting economic growth. And much as the private sector has widely adopted segmentation strategies—dividing customers into distinct groups and personalizing products and services to match each one—governments are likewise finding that welfare-to-work programs are well suited to the approach.

    Segmentation in the Private Sector

    To gain insight into the value of applying segmentation to welfare-to-work programs, it can help to examine a private-sector example with some surprising parallels. Although welfare agencies and the collections departments of credit card companies may seem worlds apart, both involve interactions with customers, with the former providing support and interventions to help with job searches and the latter using different forms of communication to encourage the payment of bills. In both cases, customers have specific characteristics that affect how they respond to whatever interventions are used. And in both cases, external providers are often enlisted to assist with critical elements of the process. (See Exhibit 2.)

    exhibit

    A central issue for credit-card-collections departments is how to proceed when cardholders miss a monthly payment. In such cases, several decisions must be made:

    • What type of intervention is most likely to recover the outstanding payment? Is a letter, an e-mail, or a phone call the best way to make contact? The credit card company does not want to alienate the cardholder (or incur extra costs) by being too pushy, since he or she may simply have forgotten to make the payment.

    • When is the best time to intervene? In the case of customers who are having trouble paying their bills, prompt action is critical. But the company does not want to intervene too soon. If the customer is likely to eventually pay in full, waiting before making contact could lead to the avoidance of costly recovery activity and to higher profits.

    • If overdue debt is sold to an external collections agency, what is a fair price? Accounts must be priced at a negotiated level that is acceptable to both the credit card company and the buyer of the debt. This involves assessing the likelihood that the debt will be recovered. The price must be reasonable, such that the external agency is prepared to take on the cost and risks of collecting, while still creating as much value as possible for the credit card company.

    To answer these and other questions, credit card companies use extremely sophisticated segmentation models—in some cases built up over decades—that provide a comprehensive “catalogue” of customer types. These companies are expert at predicting how different groups will behave. The challenge is setting cutoff points that minimize the number of “bad” customers who will default while continuing to provide services to the maximum number of customers. Getting this tradeoff correct is the key to profitability. For example, the best card companies establish acceptance criteria that simultaneously allow them to reject more than 80 percent of applicants who will likely default and provide cards to more than 80 percent of applicants who will not default.

    Tailoring Programs to Boost Employment

    The power of applying the segmentation approach to welfare-to-work programs is already becoming clear in some countries. In the U.S., unemployment insurance claimants are segmented within the Worker Profiling and Reemployment Services system according to how likely they are to exhaust their benefits. High-risk claimants are referred to employment support programs. In Australia, all unemployed workers are segmented according to the Job Seeker Classification Instrument, which factors in information such as age, occupation, and employment history. Companies that provide employment assistance services are paid partly on the basis of the level of difficulty associated with finding employment for different groups.

    For governments that want to use segmentation to improve welfare-to-work programs, there are many types of potentially relevant information. Relevant data might include demographic information on past customers (such as age and education), as well as work history, including wages earned in previous jobs and information about interactions with agency employees during previous periods in which unemployment benefits were collected. Information about where an individual lives could be factored in to shed light on the local job market, commuting options, and access to child care. Governments could also tap external data sources to build fuller profiles of job seekers.

    This information is then analyzed using specialized statistical software and is overlaid with input from frontline agency workers to provide the basis for a breakdown of the current pool of unemployed individuals into distinct segments. Among the possible categories that might emerge: “self starters,” who are highly motivated and require little support; “quick winners,” who can find employment rapidly with appropriate support; the “skills short,” who are likely to have difficulty finding a job because of a lack of critical skills; and the “less motivated,” who display a lack of willingness to find or keep work.

    The key is to match each of these groups with services that are likely to maximize their chance of finding work while keeping costs in check. Among the service levers that can be tailored for each segment: the types of support offered (such as a resumé-writing workshop, interview training, or IT skills training); the frequency of contact with an advisor (daily, weekly, or monthly); and the means of communication (such as face to face, telephone, or webcam) and associated financial arrangements. (See Exhibit 3.) And just as credit card companies use segmentation to price outsourced collections contracts, governments can use segmentation to determine payments and success fees for third-party welfare-to-work providers, based on the effort required to support the different groups.

    exhibit

    Building such advanced models takes time, but in the credit card business the rewards are well worth the effort. By adopting a similar mindset, committed governments may be able to start reaping similar benefits. No doubt models that correctly predict customer behavior more than 80 percent of the time would be a great asset for most agencies.


    To Contact the Authors

    • Alumnus
    • London
    • Alumnus
    • London
    • Alumnus
    • London

  • Add To Interests
  • SAVE CONTENT
  • PRINT
  • PDF