Serving the Underbanked in China - An interview with CredoLab
May 9, 2018
Artificial Intelligence: The Gamechanger in Improving Persistency Ratio and Boosting Customer Experience
First appeared on Asia 7.5 Degree.
Michele Tucci, head of product development and data partnerships, was interviewed by Caixin Global, one of China's most influential news outlets on the economy, finance, and economics. Excerpts from article translated to English below:
Singapore-based CredoLab completed a $1 million Series A round in the first week of May 2018. The company, established in January 2016, helps banks, consumer finance organisations and mobile-based Internet financial institutions strengthen risk control by analysing their mobile phone information footprint.
Today, CredoLab has laid out nine countries, China and Southeast Asia, Singapore, Indonesia, the Philippines, Thailand, Vietnam and Myanmar. The funds raised in this round A will be used to expand regional and national operations, such as Latin America to Brazil, Mexico, and Africa and the Middle East.
The credit rating of users through data mining and user behaviour analysis has become one of the hot trends of financial risk control in recent years. However, under the turmoil of Facebook personal data leakage, how to obtain data and how to commercialise the utility of data has become The real problem that cannot be avoided in this field.
To this end, at 7:5, I visited Michele Tucci, head of CredoLab's product and development strategy department in Singapore.
After Facebook's data breaches, the country's increasingly strict protection of personal data will become an industry trend. Will CredoLab's credit analysis by analysing the data footprint of users in mobile phones will be affected?
All information collected by CredoLab is anonymous, and we do not collect sensitive data such as mobile phone numbers and email addresses of registered users.
Banks and financial institutions that work with us will send text messages through their system and advise customers to download CredoLab's app. The number used by the registered app is not the personal mobile phone number or email address, but the user account number or loan slip number given by the bank. CredoLab only analyses the mobile device usage habits of the registered users, and the whole process is automatic and there is no manual participation.
This way to ensure the security of the data, even if the hacker hacked into our system, I can only find a bunch of behaviour data that can't confirm the identity, unless he invades our partner bank and matches all the data successfully. But the possibility is so small.
What information does CredoLab collect on the phone? Nowadays, users have increased their awareness of personal privacy protection, and they are more cautious when they authorise the App to access the collection process.
Many people may not realise that CredoLab needs to obtain only one-fourth of the user's authorisation, Alipay needs to obtain 52 licenses from users, and we only need to get 12. We analyse the user's usage habits instead of actually looking at the specific content of the data: for example, the user's photo information, we only get the time and where to take the photo, but not the content of the photo. For the user's address book, we will only read metadata such as the number of contacts, and will not see who these contacts are.
By automatically analysing the metadata of the user's “behaviour habits”, CredoLab scores a credit rating for each device, and the bank and financial institutions we serve are buying such a score.
User ratings are a very personal service. Each bank and each region has a different rating model. How does CredoLab match the service to the customer's needs? How do you let the bank believe in your service?
For each different bank, we will develop different credit rating scorecards, but the scoring criteria are generally recommended by us, because we are more experienced and efficient than banks in the analysis of equipment usage habits. Before we reach the cooperation, we will design this set of scoring system. After that, the bank will get the final score from us. If we need to update this system afterwards, we should cooperate according to the cooperation between the two parties. The effect depends on.
Our founder team has worked for big banks or financial institutions. I myself came out of Mastercard, so CredoLab has research on the needs of banks and the behavioural analysis of user data. Now that our products have only been launched for one year, we have already been recognised by 26 partners, and we can get one customer in two weeks.
Can you tell us what are the main customers of CredoLab now?
One of our clients is the top 3 banks in Singapore, and some will be announced gradually next month, including several major banks in the Philippines, Indonesia and Thailand.
Of course, our functions are more flexible. In addition to serving these banks, we can also serve private lending or Internet-based financial institutions.
The analysis of user information in China is already common, and more data mining is carried out through other trading platforms. For example, Alipay will rate users' payment information. Is it difficult for CredoLab to develop in the Chinese market?
China does have a lot of innovations in credit ratings, but there are still a large number of people outside of the tools such as Alipay who do not have a reference standard for credit scores. CredoLab's core technology is the analysis of the behaviour of the device, as long as it is used by users who use the smartphone. And e-commerce transaction records or recharge records, although you can know whether this person has the ability to repay, but can not assess how this person's "repayments", these are all analysed from behavioural habits, on this basis After adding the analysis of the transaction record, the score will be more accurate.
Will CredoLab also collect data on these e-commerce platforms and payment wallets? Will you check the user's purchase of the product to conduct a credit analysis of the user?
We do plan to introduce transaction information for these platforms in the future, but all data on our mobile phones are anonymous and limited to the user's behavioural characteristics, such as when users buy online, how often the online shopping software is opened, but purchase We do not collect personal information of this kind. CredoLab believes that the use of data is mutual, to take care of the interests of lenders, but it does not infringe on the privacy of users. The Facebook incident is a lesson.
Does CredoLab have contacts with companies in Southeast Asia for P2P or cash lending?
There are several talks about cooperation, and one country in Southeast Asia is already at the stage of data collection. The big banks and P2P platforms actually serve two different groups of people. The former has a relatively high credit rating, while the latter has many credit records or even people who do not use banking services.
And we found opportunities in both groups. In the former, we found Under-served users who are not well-serviced. Generally, there are many such people in developed countries. They actually have financial service needs. However, due to the incomplete use of credit products, the credit records are incomplete. Banks have not paid much attention to them before, but using our scoring system, we can still find quality users in them: for example, newly graduated students, new immigrants, Or the individual business owner; while the latter is a large number of Unbanked (no credit record) population in developing countries , the country's credit system is not perfect, and more need to find data support from other places. So whether it is a developing country in Southeast Asia or a developed country such as Singapore and Australia, there is a market for CredoLab.
What is CredoLab's corporate vision and future plans?
CredoLab hopes to dig every user's anonymous data to the extreme, transforming the company's products and services from a single wind-control product form to a financial institution's “equipment library”, providing our customers with, for example, anti-fraud, user information collection and identity. Simple-to-operate systems such as certification, especially to help these organisations expand their reach to people who have not yet received credit services, or to help increase the loyalty and loyalty of their existing customers.
We aim to achieve a credit rating of 7 million sets of user equipment by the end of 2018, and will complete the next round of financing in December this year. It is currently estimated that CredaLab's B-round valuation will be around $100 million.
Former banks have little enthusiasm for the user base that lacks credit records, but with the services of data mining companies like CredoLab, some banks that are willing to try new businesses can expand their business space. They will further improve their service population by reducing the bank's audit labor costs while improving audit efficiency.
This trend seems to be a bad news for Internet finance companies that target people outside the banking service. However, although the high-quality customers who have been missing from the bank's admissions will erode the market space of these enterprises, on the other hand, through the deep excavation of data behaviour analysis, these network finance companies can also help them improve their risk control capabilities. Reduce business costs and find new opportunities in markets where there is not enough credit data.
Read the whole interview on Asia 7.5 Degree.