Category Archives: FICO
New Challenges for Strong Customer Authentication

Last month I presented at the PSD2 Strong Customer Authentication Summit in London on the challenges of making sure that every customer has a compliant route to authentication. Two days later, the EBA published their latest opinion on SCA and while it seems to offer a little flexibility in terms of the timescales for implementation, in my view it also added some practical challenges in terms of how it can be implemented.
While it seems that there are several viable routes to establish possession, including the use of one-time passcode via both app and SMS, the choice of a second factor may not be easy — at least not for every customer, on every occasion.
For many customers, biometric authentication is not practical — both because of limited consumer adoption but also because availability through 3D Secure is not wide. This makes the ability to use inherence as a factor problematic.
Furthermore, as many financial institutions are opting for server-side biometrics opposed to on device, this is going to require customer enrollment, which will further delay inherence adoption and use.
It is likely that many PSPs will be forced to revert to knowledge as the second factor and that is also challenging:
- Many organizations have made a strategic decision to move away from passwords, and haven’t collected them from customers for several years. Re-establishing the use of passwords will be a significant project.
- The use of static card details such as PAN and CVV has been ruled out of use as both a knowledge-based factor and as a possession factor — somewhat contrary to an earlier opinion from the FCA.
- Other forms of knowledge-based authentication have already been ruled out by an EBA opinion that determines that the knowledge-based factor must use information that is only known by the user. Asking questions based on known information about a customer — such as mother’s maiden name, date of birth or first pet’s name — will not do.
The routes to authentication will be different for every PSP and change will continue for the foreseeable future as consumers adapt to new methods and regulation evolves. As I said at the conference flexibility in deployment is key.
To learn more, see our pages on both managing SCA and on limiting the need to use SCA by using permitted exemptions that secure payments by transaction risk analysis.
Account Takeover Fraud in Telecom – 4 Things to Watch

My last post introduced the topic of subscription fraud in telecommunications. Here, I will broaden the focus to include account takeover fraud.
As CSPs have looked to introduce additional controls and checks at the front-end onboarding process, fraudsters have moved to account takeover. As with onboarding, the customer experience and convenience are king for upgrade and re-sign processes. CSPs have taken steps to simplify account access for existing customers, but this also gives fraudsters a chance to systematically test for weaknesses to exploit.
In the case of account takeover, there are a number of factors that CSPs should look out for when existing customers place new orders. These include:
- Recent changes to the account, such as home and delivery addresses, email, password or other credentials
- Age of customer, as older account holders are more likely to be victims of fraud and may be more susceptible to social engineering
- Exceptional spend — devices with significantly higher value than the customer’s previous
- Problems in particular delivery areas — effective liaison with delivery contractors and couriers can help here
Of course, there are multiple other factors which come into play and, whilst business rules can be used to add controls, these will add barriers to the genuine customer that wants to transact. They’ll also become obsolete pretty quickly as fraudsters change tactics after hitting the barrier. Effective analytics should be used to properly understand the relationships within the data and their likelihood of being indicative of fraud.
In my next post I’ll look at analytics that can help CSPs manage subscription fraud and account takeover fraud.
SPDB Credit Card Centre Automates Early Collections

Shanghai Pudong Development Bank (SPDB) Credit Card Centre (CCC), one of the leading credit card issuers in China with more than 40 million cards issued, has used FICO® Customer Communication Services (CCS) to boost collections performance while driving down business costs.
Recent years have seen a rapid development of SPDB’s credit card market, an expanding scale of assets and stricter regulatory requirements on its collection business. In addition, the lender was also experiencing challenges to collection work and risk control due to human resource limitations. So, to improve the risk management of its business and break through the limitations of traditional human collection agents, SPDB looked to intelligent automation and introduced FICO’s CCS system in 2016.
CCS uses intelligent two-way communications such as phone calls, SMS and email to connect with customers with the right message at the right time.
CCS allowed SPDB CCC to tailor different treatment strategies to different groups to increase its effectiveness in debt collection, especially with early delinquency customers. The solution was also used to apply champion/challenger tests to explore more scientific and effective collection strategies and optimize them. Each collection approach was continually improved through analyzing the results and tweaking the strategy using machine learning. This was imperative to make sure the portfolios were being effectively managed during SPDB’s growth period.
At the same time, the SPDB Credit Card Center has saved significant labor costs using automated outbound collections. The collection business now runs its operations using 210 less staff per month, a 30 percent reduction, which has significantly cut collection operating costs, management costs and risk costs. The success of this project has allowed SPDB to reshape its business using Big Data and machine learning to improve efficiency and reduce risk.
During the past three years of continuous model and strategy optimization and iteration, CCS has helped SPDB to fully realize automatic intelligent collection for early delinquency customers and to effectively identify and prevent the risk of non-performing loans at an early stage.
“SPDB continues to use big data, machine-learning and AI with confidence,” said Sandy Wang, general manager for FICO in China. “The bank had already embraced these technologies for scoring, so it was a sensible extension to deploy them for customer collections as well. For their innovation with CCS in modernizing and digitize their banking services SPDB won our 2018 FICO® Decisions Award for Debt Collection.”
Cybersecurity: How to Protect Yourself Online (Video)

Want to protect yourself online – and protect your family – from data compromise and identity theft?
FICO’s Doug Clare, vice president for cybersecurity solutions, offers some advice in this interview with NBC King 5 News in Seattle. He was interviewed in conjunction with his talk at the US Chamber of Commerce Cybersecurity Series, where he spoke about cyber risk and third-party risk management.
“It’s important for consumers to realize that when they do business with somebody, they may be doing business with more than one party,” Clare said. “Everybody’s got a supply chain.”
For consumers, protection means paying close attention. “It’s important to stay vigilant, particularly about email,” Clare said. “If you get an email with a link, check it out, and don’t click it until you’re sure. Make sure the email is coming from who you think it’s coming from, that the domain name on the email address is correct. Email is a big challenge, be careful.”
Clare also urged viewers, “Have that conversation with your kids and your parents.”
Third-party risk management is a hot issue in the world of cybersecurity, since vulnerabilities in a firm’s supply chain, partner or vendor networks can expose sensitive data. It’s estimated that half of all data breaches occur through third parties.
According to Ponemon Institute’s 2018 Data Risk in the Third-Party Ecosystem, more than 60 percent of US CISOs have indicated being the victim of a third-party breach incident.
More tips on how to protect yourself online are discussed in this blog post and video.
Got Enough Fraud… Models That Is?

When designing a strategy for detecting and preventing fraud, everyone always comes to the same conclusion—there is no silver bullet. There are simply too many variables, and too much change in technology, customer behavior and fraudsters’ tactics for any one solution to work effectively and sustainably for every organization, no matter how sophisticated.
Consequently, experienced fraud management executives are constantly experimenting and evaluating new data sources, scores, models, algorithms and technologies for that competitive edge. They observe customers’ behavior, survey their preferences and maintain a working knowledge of fraudsters’ evolving tactics.
The goal is the same for everyone—minimize fraud losses while effectively balancing customers’ experiences and operational expenses. But the exact recipe each organization lands on—the mix of processes, people and products—varies widely and changes constantly.
Fraud Models – Five keys to finding the right fraud score
Many different providers—whether associations, processors, switches or analytic firms—have begun to offer unique fraud scoring models, targeting different products, channels and customer segments. There are also many fraud platforms that allow organizations to build and deploy their own internal models.
So, which fraud scores will be most effective for your organization? There really isn’t a secret formula, but there are some basic principles, gleaned from years of experience working with industry leaders, to incorporate into your organization.
First, you should know that leveraging multiple fraud scores is a perfectly fine practice. Each vendor has different techniques and algorithms to produce their fraud score. Every technique has its own advantages and disadvantages. While it is important to understand the underlying technology driving the different fraud scores, what is more important is the performance and effectiveness of each fraud score, and whether it solves your business challenges.
Second, you should not underestimate the power of consortium data. The sources, quality and quantity of data is a critical component in developing robust models. Be mindful of startup vendors with a minimal client base touting consortium models. A good consortium should be representative of the industry it is representing.
Third, some fraud scores are now “mandatory.” What this means is that a provider (scheme/association or processor/switch) may be requiring the use of their fraud score, but don’t be afraid to question and quantify the effectiveness of the fraud score.
Fourth, measure the effectiveness of fraud scores. You can measure model performance effectiveness in a dozen different ways. What is important is that you are using the same approach and methodology across all fraud scores. Never apply performance metrics you have received from one vendor across all other vendors, as they are all likely using different ways to measure performance. A simple metric like value detection rate can be measured in several different ways. Find a common suite of performance metrics you can measure against all fraud scoring models.
Lastly, don’t forget about cost and benefit. Understanding the cost should always be part of your evaluation of fraud scores. The benefit is equally important, as it is in any performance comparison. Do your fraud scores overlap in some areas? Can one model be utilized for part of your portfolio and another model for the other portfolios?
As fraud continues to evolve, so should fraud technology and scoring models. There are now a number of different providers out in the market that provide effective scoring models. As such, you shouldn’t rely solely on one provider; rather, leverage as many as possible. And lastly, ensure a fair model comparison has been completed utilizing a common suite of key performance.
+++
Drew Manuel is a senior director within the Fraud, Security and Compliance unit of FICO Advisors. He has over 24 years’ experience in the fraud industry and is regularly called upon to do fraud model/score reviews by clients around the world.
Enjoyed this blog? Why not read this one too.
Cyber Scores: What Do They Mean?

It’s a great question, and needs to be asked.
Cyber scores and ratings have been around for some time now, gaining steady momentum over the last five years. That said, the market for security risk assessment scores and ratings remains nascent, with a double-digit CAGR that will likely continue into the foreseeable future.
With new data protection and privacy regulations coming online — such as the EU’s General Data Protection Regulation (GDPR) and the California Consumer Protection Act (CCPA) –– interest in understanding and managing cyber risk is at an all-time high. A drumbeat of high-profile breaches underscores the risk, and the new regulations demand more diligence in managing first-party (your company) and third-party (supply chain) risk.
Third-Party Risk Is Top of Mind
The latter topic (third-party risk) is increasingly important. While organizations can readily gain some insight into their own security posture (and a second opinion from one of the commercial cyber rating firms, if they want it), the appeal of an independent, non-intrusive assessment of supply chain partners’ cyber risk is clear and compelling.
The same is true for cyber insurance brokers, underwriters and reinsurers. As more carriers enter the market (intensifying competition) and coverage moves down-market (requiring carriers to underwrite policies with lower premiums and less information), there is an acute need for an efficient, accurate way to assess cyber risk.
What Do Cyber Scores Mean?
As organizations that have done proof-of-concept pilots with cyber scores or ratings consider exactly how to leverage them in supply chain decision workflows, the question naturally comes up: Just what does this score actually mean? Insurance carriers using these scores and ratings to underwrite and price cyber risk policies are asking the same thing.
At FICO, we encourage you to ask. If you’re using one of these scores/ratings, or are considering doing so, you deserve an up-front answer.
The reality is that some of the providers in this space can’t answer the question. The scores or ratings they produce are generated by judgmental scorecards that apply “informed but arbitrary” weighting to myriad risk signals they collect. Certainly there are experts in these companies who can render a directionally correct opinion on any given input –– but the weights assigned to these signals have no statistical basis or mathematical foundation. Their relationship to actual security outcomes was never established.
And for that matter, what specific security outcome are they attempting to measure? When you compile a score based on multiple signals that are evaluated in this way, without a well-defined objective outcome, you really don’t know what you are measuring.
A Score Built on Real Data and Sound Methods
At FICO, we take a different approach. And we have the experience, tools, methods and data to back it up. FICO’s Cyber Risk Score is empirically derived, with a transparent and documented objective outcome. Our model is built to forecast the likelihood of a material breach event in the next 12 months. It’s not an opinion, a current-state assessment, or an arbitrary grade attached to a long list of potential security vulnerabilities.
The FICO Cyber Risk Score translates directly to the “event odds” of a material breach occurring in a specified time period (12 months from the score date). It is built using the measured correlations between signals and the objective outcome. Subscribers are provided with a detailed model report that describes the objective outcome, outlines the score-to-odds relationship, and exposes the population distribution across the score range.
FICO’s users know exactly what the score means.
The veracity of our approach and the transparency behind the meaning of FICO’s Cyber Risk Score are key reasons why Chartis Research recently named FICO a category leader in Cyber Risk Quantification solutions. You can read their analysis of FICO here.
We’re proud of the recognition, but even more proud that we’re able to answer the question, “What does the FICO Cyber Risk Score mean?” If you’re using a competing score, we encourage you to ask that question of your provider. If you don’t like the answer, give us a call or visit https://cyberscore.fico.com.
Follow me on Twitter @dougoclare.
Got Enough Fraud… Models That Is?

When designing a strategy for detecting and preventing fraud, everyone always comes to the same conclusion—there is no silver bullet. There are simply too many variables, and too much change in technology, customer behavior and fraudsters’ tactics for any one solution to work effectively and sustainably for every organization, no matter how sophisticated.
Consequently, experienced fraud management executives are constantly experimenting and evaluating new data sources, scores, models, algorithms and technologies for that competitive edge. They observe customers’ behavior, survey their preferences and maintain a working knowledge of fraudsters’ evolving tactics.
The goal is the same for everyone—minimize fraud losses while effectively balancing customers’ experiences and operational expenses. But the exact recipe each organization lands on—the mix of processes, people and products—varies widely and changes constantly.
Fraud Models – Five keys to finding the right fraud score
Many different providers—whether associations, processors, switches or analytic firms—have begun to offer unique fraud scoring models, targeting different products, channels and customer segments. There are also many fraud platforms that allow organizations to build and deploy their own internal models.
So, which fraud scores will be most effective for your organization? There really isn’t a secret formula, but there are some basic principles, gleaned from years of experience working with industry leaders, to incorporate into your organization.
First, you should know that leveraging multiple fraud scores is a perfectly fine practice. Each vendor has different techniques and algorithms to produce their fraud score. Every technique has its own advantages and disadvantages. While it is important to understand the underlying technology driving the different fraud scores, what is more important is the performance and effectiveness of each fraud score, and whether it solves your business challenges.
Second, you should not underestimate the power of consortium data. The sources, quality and quantity of data is a critical component in developing robust models. Be mindful of startup vendors with a minimal client base touting consortium models. A good consortium should be representative of the industry it is representing.
Third, some fraud scores are now “mandatory.” What this means is that a provider (scheme/association or processor/switch) may be requiring the use of their fraud score, but don’t be afraid to question and quantify the effectiveness of the fraud score.
Fourth, measure the effectiveness of fraud scores. You can measure model performance effectiveness in a dozen different ways. What is important is that you are using the same approach and methodology across all fraud scores. Never apply performance metrics you have received from one vendor across all other vendors, as they are all likely using different ways to measure performance. A simple metric like value detection rate can be measured in several different ways. Find a common suite of performance metrics you can measure against all fraud scoring models.
Lastly, don’t forget about cost and benefit. Understanding the cost should always be part of your evaluation of fraud scores. The benefit is equally important, as it is in any performance comparison. Do your fraud scores overlap in some areas? Can one model be utilized for part of your portfolio and another model for the other portfolios?
As fraud continues to evolve, so should fraud technology and scoring models. There are now a number of different providers out in the market that provide effective scoring models. As such, you shouldn’t rely solely on one provider; rather, leverage as many as possible. And lastly, ensure a fair model comparison has been completed utilizing a common suite of key performance.
+++
Drew Manuel is a senior director within the Fraud, Security and Compliance unit of FICO Advisors. He has over 24 years’ experience in the fraud industry and is regularly called upon to do fraud model/score reviews by clients around the world.
Enjoyed this blog? Why not read this one too.
Telecoms Leverage Omni-channel Collections to Reduce Churn and Improve the Customer Experience

Telecoms today face many challenges, including high roll rates and expensive high-touch contact strategies. A key challenge is to more effectively communicate with customers. A growing number of consumers prefer to conduct business on their mobile devices, and many have a preference for the type of communication they receive, whether it’s via IVR, email, or SMS. Telecoms are struggling to establish best practices around how to optimize these interactions.
Telecoms have to do more with less. By deploying advanced analytics, including AI and Machine Learning, they can gain a greater understanding of customer expectations and experiences. By automating processes and improving the customer experience, we have actually seen delinquency rates drop by 40%, collection costs drop by 15%, and a reduction in the number of days it takes to collect.
Customer experience and satisfaction has been ranked a number-one business priority by a large majority of the top telecoms, globally. By incorporating automation into their business, they can deliver the value of the company’s best-performing agents, while flexibly scaling them to handle whatever level of capacity is required on any particular day. These new tools help telecoms deliver the right message for a consumer, and convey it via their preferred channel at the right time – thus maximizing the likelihood of a successful outcome.
FICO and Cox Communications recently presented a webinar on “Telecom Trends and Best Practices in Omni-Channel Collections.”A recording is available at: https://content.fico.com/l/517101/2019-04-04/7xfg5
Using Analytics to Prevent Auto Finance Fraud

Recently I had the pleasure of presenting in Toronto, at the Canadian Auto Remarketing Conference. While the audience was comprised of a diverse group of stakeholders involved throughout the automotive lending lifecycle, pretty much everyone had one thing in common; they had all the unfortunate experience or had knowledge of some sort of fraud in their line of work.
Like most things, auto finance fraud has evolved over time. As technology becomes more advanced, fraudsters are becoming more innovative, meaning the measures organizations are having to take to protect themselves require innovation as well. You can’t solely rely on checking a driver’s license anymore to verify someone’s identity; it’s simply not enough. Auto lenders are always asking themselves, is this individual who they say they are, or do they intend to pay for the loan.
Fraudsters and legitimate consumers have a number of avenues in which they can now initiate a purchase and each one comes with different challenges for a dealership or lender to navigate—which requires quick decision making.
No matter the industry, the challenges remain consistent. There are various forms of auto finance fraud that can make it incredibly difficult to flag suspicious transactions.
- True name application fraud is straight forward as it’s a real person and real information. This means it’s simply a bad person, with bad intentions.
- Manipulated fraud is challenging as the information can seem legitimate but has some small inaccuracies—these could be small misspellings in a name or slight changes to an address. The information seems authentic but makes tracking delinquent payments nearly impossible.
- Synthetic fraud takes parts of a legitimate identity and attaches fake information to it. This can mean using legitimate personal identification and assigning them to fraudulent people. These cases are also difficult to identify because they look like and act like a true and good individual.
- Stolen identity fraud, while most often talked about, is actually the least common type. This is when a fraudster fully uses the identity of a stranger, resulting in a damaged credit score and future.
In the days when consumers were required to come into a dealership or bank to obtain financing, judging the validity of an identification made ruling out some of these fraud types a bit easier. An 18-year old man can’t claim to be a 60-year old woman without raising some serious red flags in person—but it’s easier to slip through the cracks online.
No matter the industry, fraud causes real issues. However, the stakes are seriously high when it comes to auto finance fraud. Once that vehicle is driven off of the lot, it’s gone, and lenders have begun to be held increasingly responsible for payments that default.
What this means is it’s becoming increasingly important for any agency responsible for making credit decisions to become more vigilant in evaluating credit worthiness, which means putting analytics to work. A low credit score is no longer the biggest indicator of a risky candidate, but analytic tools can alert lenders to the crumbs of information that are easily overlooked by simple qualifying questions alone.
Auto Lending or Dating?
If you ask me, evaluating application and originations fraud risk is a lot like dating. This might sound crazy but bear with me.
While many people go through phases of their lives where they are looking for something “casual,” for the most part, the goal of dating is to identify a potential long-term partner whom you can build a lasting relationship with. A lending relationship is no different. Lenders and merchants are looking for trustworthy, dependable candidates. They are lending money to individuals who are committed to the relationship as much as they are to pay back those car loans.
In both scenarios you can expect the truth to be stretched a little bit during the initial stages of the relationship as you’re getting to know one another. Perhaps your potential partner exaggerates their cooking skills on your date, or perhaps your financing candidate exaggerates their income a bit—there’s usually a little white lie somewhere in the equation. In both scenarios, you have to decide what fibs you can live with and which are total deal-breakers.
There are different levels of risk associated with the strategies you choose to vet your potential partner no matter the type of relationship you are pursuing.
- Referrals = friend of a friend Arguably the least risky method when dating is meeting a potential partner through a mutual friend. In these cases, while you don’t know the individual personally, if the friend setting you up is a trustworthy person, you can assume that they will have used their own judgement to suggest someone who will be compatible with you. The same applies when it comes to lending. If a referral comes in from a strong client relationship or has pre-qualified for an offer, you can assume that the candidate is likely legitimate.
- In-dealership = speed dating Having a candidate walk from the street into a lending institution is the equivalent of meeting a potential partner through speed dating. You don’t know anything about them entering into the interaction (aside from what you can see physically) and you only have a limited amount of time to make your judgements about them. In this sort of interaction there are some things that cannot be falsified; gender and age (within reason) for example, but there is little time to evaluate more than that before deciding whether the person is a good candidate.
- Applying online = online dating Online relationships are hard—everyone lies online. According to a recent survey about online dating, 53 percent of people using sites to find partners admit to lying on their profiles and applying online for financing is rarely different. Both scenarios can leave you feeling deceived and disappointed.
When it comes to evaluating risk for an organization though, online applications do have benefits. While it is easy for fraudsters and scammers to lie online, they also leave many breadcrumbs, which if properly monitored, can alert an organization to red flags. Many of these breadcrumbs—like addresses and IP addresses not lining up or registering using brand new email addresses—might not be alarming on their own but can alert an organization to some big issues if analytics are properly being utilized.
In dating or finding reliable customers, there is some effort involved. You can’t expect Mrs. or Mr. Right to come strolling up to you at the grocery store without taking measures to meet the right people with the right qualifications.
Tax Compliance; Enhancements with Analytics

Tax agencies have a limited amount of resources to pursue tax compliance activities (collections and audit). Because of this resource limitation, they are required to build criteria to determine which individuals and businesses to select for audit and which collection cases to focus their efforts on. Typically, this selection is based on experience that informs agency leaders on the types of businesses that generate productive audit leads and collection cases. While the selection process typically utilizes available data, cases are primarily selected based on experience, intuition and business rules, rather than using predictive, mathematically generated analytical models.
Tax Compliance – Why Use Analytical Models at Tax Agencies?
Predictive, analytical models produce better outputs over experienced based “expert” models by increasing the accuracy of the model. The IRS, among other tax agencies are investing in predictive analytics. This results in more productive workloads and outcomes. Analytic models can:
· Increase customer service by minimizing interventions on cases that don’t need tax agency staff to achieve collections;
· Increase staff productivity by assigning them to better cases where their expertise and effort is more likely to result in additional dollars collected;
· Reduce wasted efforts by reducing the no-change audit rate, and reducing the time a collector spends on cases where their intervention does not affect the collection rate; and
· Increase long-term voluntary compliance by maximizing current collections and affecting long-term behaviors that increase future collections.
Tax Compliance – Predictive Analytical Models. Three Types
There are many different types of analytical models which a tax agency could use, including:
· Similarity Models. These look for taxpayers that are statistically similar to cases that had a specific behavior in the past (compliant or non-compliant).
· Outlier Models. These models look for taxpayers that are statistically anomalous. While this doesn’t mean they are non-compliant, they can merit a review to determine why they are different from their peers.
· Prescriptive Models. Prescriptive models are different from the prior Predictive models, because prescriptive models look more holistically for the best overall set of decisions, rather than identifying individual cases which may be productive. A prescriptive model also helps allocate resources between workloads, taking into account your constraints to identify the best set of strategies to maximize your overall result.
The remainder of this document will talk to the specific areas of a tax department that can benefit from analytical models.
Tax Compliance – Tax Analytics Opportunities for Audit
Auditors tend to be the highest paid and most specialized resources within a tax agency. Agencies have a limited number of auditors, and therefore must be careful with allocating these resources. Analytical models provide an opportunity to increase both revenues during the current fiscal year as well as future voluntary compliance.
Analytical models can identify audit candidates that have a statistical similarity to past productive audits. These models can score for dollars assessed, dollars collected, or even dollars generated per audit hour, allowing the agency to maximise their return. They can also be used to minimize the impact to compliant taxpayers, trying to reduce audits that result in no or limited changes.
Tax Compliance – Tax Analytics Opportunities for Collections
Collections is another area with many opportunities for analytical models. Collection management invariably has more inventory than staff available to work cases. Therefore, they must determine which cases to assign for treatment and when to assign the case. Typically, cases are assigned based on dollar amount owed, but as I described in an earlier paper, predictive models can also help determine who will self-cure and who needs more attention.
Models can also help determine who to contact, when to contact them and what message to deliver. Studies have shown that even subtle changes to letters can result in Millions in additional collections. By combining analytical models with behavioral science (varying notice texts) you can vary your notices by taxpayer, using the language most likely to result in success for each taxpayer communication.
Finally, analytics can predict which taxpayers are most likely to need the most aggressive collection techniques. While the Government must provide due process to all debtors, if a segment of taxpayers are unlikely to pay through billing, then the analytics can help the agency move these cases through to enhanced collections faster, without expending significant staff time on phone calls that are unlikely to result in collections.
Tax Compliance – Tax Analytics Opportunities for Customer Service
Tax agencies also have limited resources to provide proactive customer service. Analytics can act almost like a crystal ball to predict which taxpayers are likely to encounter future tax issues. For example, if statistics show that a new business in a specific industry is more prone to under-reporting or not registering for specific tax types, the tax agency can proactively send letters, make phone calls or organize industry workshops for businesses that meet that profile, helping them understand their upcoming tax obligation before they encounter an issue. The taxpayer will better understand their obligation and they will know the government is paying attention, which will lead to higher voluntary compliance before encountering problems.
Tax agencies can also use analytics to measure the impact of their customer service, the effectiveness of different channels (e.g., in-person meeting, phone call, or letter campaign), and the effectiveness of different messages. These can help shape future interventions by testing different strategies and utilizing the most effective one.
In conclusion, tax agencies are under continued pressure from their legislatures to do more with less. Innovations in deploying analytical models provide the opportunity to enhance customer service, revenues collected, and achieve greater compliance, not by working harder, but by working smarter.
###
Enjoyed this blog? Why not read some more on the topic of tax and analytics:
Improving Tax Office Data-Matching
Tax Evasion: Have We Learned the Panama Papers Lesson?
Identifying Tax Fraud through Social Network Analysis
Tax Agency Optimization – What Are the Benefits?
FICO Survey: APAC Banks Expect Rise in Tax Evasion