Credit rating in e-commerce: Scoring today and in the future

Credit rating in e-commerce: Scoring today and in the future

Scoring influences the credit, trade and everyday life of consumers. From the order on the bill in the onlineshop via cell phone contracts and loans, many everyday business is dependent on a scoring. How does scoring work today and how could it look against the background of Big Data in some years? What are the advantages and disadvantages for retailers as well as for consumers? How does the so-called social scoring, which evaluates various user data from the Internet, affect creditworthiness checks in the future? And what does this mean for traders and consumers?

In e-commerce scoring is probably not as important in any other country as in Germany. This is due to the fact that the most popular online payment methods are charged for direct debit and purchase. Therefore, online buyers pay so much by direct debit and invoice, because they receive their order quickly and at the same time have a certain security: a debit note can be booked back by the customer, and he does not pay the invoice if the goods are in deficient condition or not at all Is supplied. This is practical for buyers but for dealers, there is a risk of payment default. In order to protect themselves against black sheep, shop operators can therefore carry out a credit check. In particular, traders whose goods can easily be resold, such as suppliers of electronic products or small furniture, bear a high risk of default.

Credit rating in seconds

If, for example, the operator of an electronic on-line shop decides to minimize his risk in the payment methods of direct debiting and purchase by scoring, how is the ability to pay for a shopper who would like to order an LCD TV device worth several hundred euros? To this end, it has to be said at the outset that it is a strictly guarded secret of scoring service providers how the process works in detail. So much is certain, however: the buyer does not even know that in the background of his order a classification of his creditworthiness takes place, the scoring happens within seconds. If it is positive, the buyer has the choice between the whole range of payment methods, including direct debit and invoice purchase. If, however, the analysis shows a high probability that the customer will not pay the invoice - whether intentionally or not, he will only be offered payment methods which are secure from the retailer 's point of view. This includes, for example, credit card payment or transfer: in both cases the money is collected by the dealer before sending the goods.

Information from several sources

In order to determine a scoring value for an online customer, a corresponding service provider uses several sources. These include classic credit bureaus such as the Schufa, Bürgel or Creditreform as well as various publicly accessible directories. For instance, insolvency courts must announce the opening of company insolvency proceedings as well as private insolvency proceedings, and public courts make an entry in the list of debtors if a debtor has to make an affidavit of his assets after an unsuccessful attempt to seize the court. However, classic credit bureaus and public directories alone do not yet provide a meaningful scoring value: on the one hand, it takes a certain amount of time for current information to be entered, and on the other hand the entries are either deleted by themselves or the person concerned can delete them upon request leave. Therefore, scoring service providers also use databases from companies to assess the creditworthiness of online customers. In the HWD database, for example, stationary and online merchant account numbers as well as bank sort codes of EC cards with which back-office invoices were issued. A further example is the EOS Group, a subsidiary of the Credit Group, which is also a shareholder of the Buergel information agency: EOS has access to all Otto customer databases and thus offers one of the largest sources of information on the payment morale of consumers in USA / UK / Germany.

What the residential realm tells about the moral of payment

In addition, another puzzling piece contributes to the scoring value: sociodemographic data. Behind this, data are primarily conceived about the purchaser's residential area. Does it reside in a high-quality rental area where a relatively large number of high-priced brands are registered? Whoever can afford to live here has a high probability of being able to pay. It is only with a high probability that the inhabitants of a certain area rarely form a completely homogenous demographic group, not to mention the fact that the customer does not necessarily indicate his own address when ordering, but possibly the development of his or her partner Scoring service providers, however, on the basis of all the data sources mentioned: classic credit bureaus, public directories, company databases and sociodemography. The result is a value that indicates the probability that a customer will pay his order. Depending on which minimum value the dealer has previously defined, the customer then automatically receives the payment types direct debit and invoice - or not.
Scoring in the age of Big Data

This is how scoring has worked so far in e-commerce. But the ever-evolving digitalization of everyday life - the keyword Big Data - has brought new potential for use in this area as well. Consumers leave a lot of data in social media: about their own person, the professional career, the social environment as well as personal preferences and interests. Many of this information also give hints about the creditworthiness of each person. Is she a friend of Facebook, especially with academics? Does it cost many pages of expensive consumer goods brands? Has she listed a company as a current employer on Xing or LinkedIn? All this suggests that potential creditors have little to worry about their money with this customer.

Mosaics from thousands of data points
Pioneer in the field of this so-called social or big-data scoring is the Hamburger Start-up Kredch. Founded in 2012, the company already has more than 200 employees and subsidiaries in nine countries. The business model: Kreditch lends money to consumers, even to those who would not be lured by a bank due to a negative entry at a credit agency. However, Creditsch grants these persons small loans between 50 and 500 euros if they reveal to the Hamburg company income as well as employers and allow him to crawl the web after various information about the applicant. Complex, intelligent algorithms evaluate data streams that they have left in web browsers, databases, social networks, and platforms such as Ebay and Amazon. Technical data about the device used - the latest iPhone or the old cheap PC - are just as much in the evaluation as features of the Facebook friends and the way the application form is filled: It takes one minute for the applicant to think about his employer From? Or does he often press the delete key? From around 20,000 data points, a total picture of the person has emerged after a few minutes, from which a meaningful forecast value can be derived about their future payment behavior. Social scoring
Complement to established methods
In Germany, Kreditch had stopped the granting of loans - on which the start-up raises 15 to 30 percent interest - within just a few weeks in order to forestall a review of the business model announced by the Federal Financial Supervisory Authority (BaFin). In other countries, however, Kreditch still offers the service and is working parallel to a second line of business. The aim is to build up an international scoring database and, on the basis of this, to offer credit ratings, especially for e-commerce companies. Is this the beginning of the end of classic scoring procedures? The fact that social scoring will replace conventional credit assessment methods is currently unpredictable. The proven scoring methods are largely based on hard facts, on actual events in a customer's payment history - this information will continue to form the basis of credit assessments. Classic scoring providers will, however, explore the possibilities of supplementing their procedures with social scoring. This is particularly important because of the actuality of the data and the analysis of thousands of data points within a very short time. As a result, social scoring offers a possibility for classic providers to reduce the number of false positives: Customers who get a negative scoring value because of missing or obsolete data, although they are (again) completely creditworthy. Unfortunately, such cases occur again and again, although the frequency is difficult to quantify. On the other hand, social scoring also carries the risk of a good assessment of people with poor credit ratings. This is because when someone has a kind of digital mask, and social networks seem to be someone who is not. In Germany, social scoring will rather complement the established procedures rather than replace them. In international e-commerce this is different, there are in many countries, nothing comparable to the German Schufa. There, social scoring could be a way for traders to get at least low-cost credit ratings at all.

Creditworthiness check by Google soon?
In any case, credit checks in e-commerce are likely to become less important as payment methods such as PayPal and credit cards are used. In the case of electronic wallets such as PayPal, the corresponding providers would have sufficient data on the payment behavior of their customers in order to ensure the creditworthiness check themselves. It is also conceivable that companies such as PayPal, Apple or Google, their payment solutions should be widely used, even to providers of credit checks. Paying by credit card is also very safe for online merchants, as a credit card transaction is not completed until the bank of the cardholder authorizes the transaction. For this purpose, the system automatically checks whether the card is valid and whether the amount does not exceed the credit card limit.Conclusion
It is worthwhile for online retailers to have a scoring carried out by their payment service provider - with the risk of unpaid invoices and re-debited direct debits. Current credit assessment procedures mainly use information that is directly related to the financial situation and the payment behavior of the person concerned: open claims, insolvency proceedings, garnishings, etc. Social scoring offers a completely different approach in the context of Big Data. Intelligent algorithms combine all the data that the Internet provides over a person to form an overall picture, and in turn derive the anticipated payment morale. Even if this happens anonymously, the idea is likely to cause discomfort among many German consumers. For this reason, and because there are other good practices in creditworthiness testing in Germany, social scoring will be more difficult for us than in other countries.