Who are your most valuable business customers? Which ones can you trust to pay their invoices on time? Are customers in certain economic regions worth targeting more than others? Which ones are proving a drain on customer services resources?

Most businesses would like to have answers to such questions at their fingertips. But that will only happen if they put in place a foundation for monitoring and analyzing customer behavior and channeling that into a mechanism for customer scoring.

So, what is customer scoring? What steps are involved when it is applied effectively in a business-to-business (B2B) context? And which customer scoring models will best support your future business success?


• Customer scoring can help to drive greater revenue and profitability while reducing business risk.

• A wide variety of customer scoring techniques fulfil business goals, from creating confidence in credit decisions to reducing customer churn rates.

• Using customer scoring models based on comprehensive, accurate, and up-to-date data results in customer credit scores that inspire fast, confident business decision-making.

• Customer scoring can empower your e-commerce business by enabling you to offer Buy Now, Pay Later (BNPL) facilities to customers, knowing the level of credit risk involved for each.

Customer scoring is a means of segmenting your customer base so decisions related to individual buyers can be made with a high level of confidence, speed, and efficiency. By statistically analyzing how each performs against a set of important criteria – year-to-date sales, payment history, repeat orders, and so on – each can be allocated a single, easy-to-grasp score that reflects their value to your organization. 

Some scores are designed to be used right across the business by sales, finance, senior management, and so forth; others are set up by different parts of a company to fulfil their own specific purposes. 

A customer service department, for example, might have criteria based on the number of service tickets raised by a certain customer while factoring the renewal date of its contract into the score. A credit management team may use customer credit scoring based on the size and payments history of the buyer, its recent financial performance, bank references, and the health of the industrial sector in which it operates. And a manufacturing unit might want its customers scored to show which orders, discounts, and deliveries it should prioritize.

As that suggests, there are many answers to the question ‘What is scoring in business and why is it worth doing?’ But the shared goal is to segment customers so you can treat them differently, deepening the relationship with the ones you know are reliable and valuable while reducing the risks and overheads associated with those that are less so. In a nutshell, when well-executed, customer scoring guides your business towards better outcomes. 

Different organizations and industries implement a wide variety of customer scoring models to suit their purposes, but a handful stand out as the basis for sound business decision-making:

  • Credit scoring
    Few businesses can function without embracing credit risk. Traditionally that has meant providing goods or services on net terms, with an expectation that an invoice will be paid within a set period, such as 30, 60, or 90 days. But few businesses offer such a ‘grace period’ without first qualifying a customer’s ability to pay – in other words, most businesses engage in some kind of customer risk scoring.

    Whether that is simply based on an established relationship with buyers or a more rigorous analytical approach, such customer scoring is key to establishing their creditworthiness. It protects your cash flow and profitability from the negative impact of delayed payments and defaults. But it has another side too. It allows your business to deepen its trading partnerships and develop loyalty with select customers by offering them higher levels of credit than you might otherwise feel comfortable with.

    Building a credit scorecard on a given customer typically involves weighing up an array of buyer information, including the customer’s size, ownership, and subsidiary structure; its historical financial performance, operational data, and public filing; the markets it is active in and their current stability; its debt burden, payment history, and any collection issues.

    As e-commerce has expanded the addressable markets for many businesses, the need to obtain a clear picture of the creditworthiness of their customers has become even more critical. E-merchants, especially those dealing with larger B2B transactions, increasingly want to offer deferred payment options to online customers in the form of BNPL credit. But in doing so, they need the assurance that such an approach does not expose them to greater risk. (see below : Customer scoring powering B2B BNPL)

    The basis for that is reliable credit scoring, a process that requires casting a wide net of credit assessment – certainly wider than traditional credit ratings. Produced by agencies such as S&P, Fitch, or Moody’s, credit ratings are a forward-looking, independent categorization of a company’s ability to meet its debt obligations in full and on time, usually ranging from AAA for the best performers to D for likely defaulters. While credit ratings are an important feed for credit scores, they are, by their targeted function, only providing part of the picture.

    Of course, there are countless other customer evaluation methods – from the casual to the highly scientific. Some accounting teams may have the time and expertise to dive into a customers’ financial reports and filings to spot red flags, such as the late filing of accounts or weak liquidity, and come up with their own solvency ratios. But most rely on expert partners to gather the right data and turn it into actionable grades.
  • Churn risk scoring 
    In itself, calculating a churn rate is straightforward – the customers you’ve (unintentionally) lost over a given period divided by the total customers at the beginning of that period, expressed as a percentage. But predicting which customers are likely to stop buying from you and grading them on that basis is another matter altogether.

    A churn risk score is typically built by surveying satisfaction levels among customers or monitoring help/complaint lines to gauge sentiment on aspects such as value for money, product fit, and customer experience. But churn can also be induced by outside factors, such as geopolitical or economic events.
  • Fraud detextion scoring 
    Fraud detection scoring helps businesses put a value on the level of risk associated with a particular customer transaction before the order is processed. To achieve that, fraud scoring typically couples data analytics with machine learning to spot patterns of fraudulent behavior and minimize transactional losses. In many e-commerce settings, it often involves carrying out checks on customer orders in near-real time and scoring those based on a set of rules that assess whether a transaction is suspicious.
  • Lifetime value scoring 
    Lifetime value (LTV) scoring seeks to measure the total value generated by a particular customer across their entire engagement with your business and to represent that as a single grade. In essence, it sums up the degree of customer loyalty, signaling when a relationship should be closely tracked and nurtured by accounts, marketing, or customer support teams.

    This is a critical metric but one that isn’t always easy to assess. The volume, frequency, and value of orders are an obvious feed for an LTV, but so too are less quantifiable aspects of customer value creation, such as a customer’s willingness to be an advocate for your value proposition – with other customers, in marketing materials, at conferences, and with the press and analysts.
  • Customer satisfaction scoring
    Customer satisfaction, or CSAT for short, is an aggregate of different measurements of customer sentiment on a business’s products or services. Measurements can come from numerous sources: post-sales customer feedback surveys, data on the frequency of repeat purchases, or less-structured feedback, such as comments on review forms or via social channels.

    It differs from the widely used Net Promoter Score (NPS). NPS is designed to reflect customer loyalty to your business or offering, based around a single fundamental question: ‘How likely are you to recommend our organization/product to a friend, associate, or colleague?’

1. Data access 
To build accurate, up-to-date, comparable scores on customers, you need access to the right data, in sufficient quantities, and from all the right places. Some of that data is going to come from your own systems, showing buying patterns, payment histories, customer satisfaction reports, and so on. But that is only part of the picture needed to develop a customer score. It is vital that you also tap into reputable third-party sources. For instance, at Allianz Trade, we grade companies around the world by tracking and enriching many thousands of data sources, as well as engaging directly with buyers themselves as part of our core trade credit insurance business. The need for independent third-party guidance is especially important when you are trying to allocate a score to first-time buyers or those from countries or regions unfamiliar to your organization – a situation that’s particularly common in e-commerce settings. 
2. Scoring criteria  
A rounded picture that informs a customer score needs to be based on a broad mix of attributes. Those might include a customer’s bank information, debt history, and financials. But it might also consider sector- and country-specific risks they face, details of their trading patterns with other companies, and much more besides. And that takes expertise – both international and local.  
3. Scoring model development  
Customer scoring models are developed to support different decision points. By weighting criteria, they can be designed to, for example, rank the most loyal and the most problematic customers, the best payers and the companies that are always late, the companies where after-sales service is easy, and those that are high-maintenance. A simple example is the Recency, Frequency, Monetary (RFM) model. This allocates a score of 1 to 5 to each of the three attributes, based on a customer’s buying patterns, to produce an overall value. A model for a credit risk score, on the other hand, might include up to 15 variables.

4. Scoring and segmentation 
By assigning each a score, customers can be segmented into groups that reflect their current situation and help you predict their behavior. That might, for example, show you the level of risk your organization faces when selling to a specific customer and so influence the credit terms you deem appropriate.

5. Decision making and action 
With a score flagged up for each customer, decision-makers can act with new levels of speed and confidence. And when informed by insight from third-party experts, that decision-making can drive low-risk sales growth among both existing and new buyers.

Customer scoring has countless use cases in business, but one of its most important applications is in reducing the risk your company faces from customers whose scores would indicate they may be slow or fail to pay an invoice. The flip side of that, of course, is that customer scoring can also allow you to offer higher credit limits to customers you can trust to pay.

Scoring the creditworthiness of your existing customers – and getting a handle on prospective customers – requires the monitoring and analysis of a number of key attributes, such as:

  • Solvency information to gauge the financial stability of customers and their ability to repay debt on time
  • Different types of risk assessment, including market risks, commercial risks, political risks, reputational risk, and ESG risks
  • Company data, including financial results, balance sheets, payment history, the use of invoice factoring, governance, management turnover, credit ratings, claims histories, and collections activities
  • Credit data, assessing the use and mix of credit over time.

At Allianz Trade, we use state-of-the-art credit assessment models, an internal risk platform, and an information pool that tracks 85+ million companies worldwide to create the most accurate, up-to-date, and actionable credit scores. Our economic research and country-based teams also feed in market, country, and political risk factors to further enhance the reliability of company scores.

Our portfolio of models, built and maintained by a team of insurance actuaries, credit analysts, statisticians, and data scientists, is constantly evolving to take advantage of the latest technologies. It combines logistic regression, fuzzy-logic, and AI-based machine-learning technologies to generate cutting-edge grades. 

Credit scoring has always been a powerful and important tool for helping to decide who you want to do business with, and the credit levels you can safely offer them. But as a larger proportion of business transactions have moved online, its role has become even more critical.

Although over one-third of all B2B purchases by large- and mid-sized business are now made directly online 1, the practice of offering credit terms has not made the leap to e-commerce. The vast majority of e-merchants in the B2B space still expect immediate, upfront payment by card or bank transfer for new purchases, with only a small faction able to provide the kind of payment options that are the hallmarks of B2B trade – from payment on net 30 to spread payments. 
The result: too many frequently abandoned baskets and lower transaction volumes than might be expected if credit terms were available.
But many e-merchants are now rising to that challenge by taking a leaf out of the consumer finance playbook. They are implementing B2B Buy Now, Pay Later capabilities that can offer business customers appropriate credit terms by running a real-time check as buyers move through checkout.
It’s a leap that is underpinned by customer scoring. Sophisticated B2B BNPL e-checkout solutions combined with real-time credit scoring and backed by financing and credit insurance can help you grow your business by allowing you to offer favorable terms to reliable customers and minimizing the risk of non-payment associated with others.

As the world leader in trade credit insurance, Allianz Trade is partnering with major banks and B2B Buy Now, Pay Later pioneers to allow e-merchants to confidently embrace B2B BNPL. Our integrated solution, protected by Allianz Trade pay, is designed to help your business achieve:

  • Higher e-commerce conversion rates at checkout
  • Fuller checkout baskets and more repeat customers 
  • Real-time, automatic decisions through best-in-class credit scoring
  • Cash flow that is not impeded by slow or late payers 
  • Quick and easy set-up of B2B BNPL through an API that plugs directly into your e-commerce system
  • And, ultimately, peace of mind – knowing we collect from your buyers when necessary, and you always get paid in full.
  1.  The new B2B growth equation, McKinsey & Co