AI, Big Data, and Your New Credit Score
One thing that is common among most Americans is that they love to buy isht on credit. That inclination is only growing with the passage of time, in part, due to low interest rates and because there is just so much cool stuff to buy and so little money on hand. Here’s a look at this trend, courtesy of 60 Second Statistics:
Now we’re not sure how it works in other countries, but the way it works in the U.S. is that you have a credit score which is a basic indicator of how likely you are to pay back a loan. When an institution loans you money, the likelihood of you paying back that loan helps determine the interest rate as well. Higher risk begets higher interest rates. For that reason, it is extremely important that any lending firm accurately assess your financial picture. That’s kind of hard to do with just a 3-digit number. This is where “big data” steps into the picture and you may not like what we have to say. While there are no doubt loads of firms out there that claim to evaluate your ability to payback a loan using big data and/or artificial intelligence (AI), for the purposes of this article, we’re just going to focus on one startup called DemystData.
DemystData is helping clients cut risk by up to 60% and increase acceptance and straight-through processing rates to convert more customers. We provide the key to accessing valuable data 10 times faster while adhering to increasing privacy constraints
Founded in 2010, Singapore startup DemystData has taken in a total of $12 million in funding so far with their latest Series B of $7 million closing last October. The firm uses big data to help the world’s largest telcom providers, insurers, and lenders to “optimize their workflows” which essentially means determine your potential as a customer using big data from over 150 different sources.
Update 08/22/19: DemystData has raised $12.5 million in growth capital to further accelerate its data on-boarding, invest in data platform development, and grow its product and client teams. This brings the company’s total funding to $31.5 million to date.
Here are some categories:
This isn’t just about making more accurate lending decisions, but it’s also about cross-border lending. As the world becomes smaller and smaller, people are moving about and credit scores just don’t exist in all places like they do in the U.S. where most our readers live. In fact, the company founder, Mark Hookey, founded the company after moving to Hong Kong and realizing that he couldn’t use his U.S. credit score in another country.
Instead of that single U.S. credit score consisting of 3 numbers, DemystData uses over 5,000 different attributes which they claim is more than any other provider out there. Their technology isn’t exactly just entering the market as it has already been used on over 400 million customers. Are you dying to hear about what sort of information they are gathering about you in order to assess your credit worthiness? So are we. Here are three examples:
Income Consistency – Come on now, we’ve all been tempted to fudge our income just a bit. It’s not even about wanting to get approved, it’s just that our egos make us want everyone to think that we make more than we actually do. The Demyst algorithms compare your self-reported income with things like your job title (LinkedIn maybe?), your life status (kids cost a lot money), the industry you work in (radiology isn’t looking too hot these days), and your purchase behavior (restaurants you frequent as an example).
Telcom Data – This one took us by surprise. There is a wealth of data that can be pulled from your phone records. As an example, Demyst lists “phone location matching” as an attribute under “fraud protection”. We would speculate that this means if your credit card is trying to buy a new television in Florida but your phone is currently moving around Los Angeles as it always does, something isn’t right. You can just let your imagination run wild for a bit here thinking about what sort of information can be gleaned from your phone records.
Conversion Triage – Excellent name for something that’s plain scary. Taken right from their website, Demyst states “Use email history, demographics, social profiles, and more to identify customers likely to qualify and be approved for your products“. Honestly, we racked our collective brains trying to think how “email history” is acceptable but we just couldn’t come up with a suitable use case for that.
Those are some of the attributes that Demyst came up with but we can probably think of some more. You can bet that some of these are being actively used today by some firm out there:
- County and neighborhood – All your neighbors make minimum wage but you’re rolling 6 figures? Come on now.
- Schools you attended – This one is a no-brainer
- Sexuality – Gay people often have more spending power as many choose not to have children
- Friends you keep – If you plaster every detail of your life on Facebook, don’t be surprised if people are using that to make decisions with
- All Criminal Offenses – Like it or not, much of this information is public record and it will be used
- LinkedIn profile – This is a good source of info because you are compelled to tell the truth on your profile. Your colleagues are the checks and balances.
You can probably think of a whole bunch more. Going forward, your actual credit score is going to have less and less a say in whether or not lending firms give you a loan. As an investor you might be thinking, why not short the publicly traded credit agencies? It’s not the worst idea, but it’s hard to pull the trigger on that trade when Equifax (one of the biggest agencies) has a stock price chart that looks like this:
The reality is that the big credit agencies are thinking about this stuff, if not doing it already. In a presentation by Equifax titled “Navigating the Big Data Super Highway“, the below chart was presented showing that the highest growth financial institutions are using far more data sources than just your credit score:
Yes, Equifax is well-aware of what direction the “new credit score” is heading and you can bet that they have their eye on fintech startups like DemystData. Eventually, we may see a world where there is a customized interest rate for every single loan transaction. For those of you who have picked up on the fact that we haven’t mentioned artificial intelligence (AI) yet, you’re right. There is just far too much obvious information to glean from this data. We’ll let the AI algorithms go consume all that delicious big data later.
Of course with AI comes a certain ambiguity as to why particular decisions are being made. For those of you who will scream for “transparency” in the process because you believe that some underrepresented group is being discriminated against, good luck with that. There will always be a valid reason to give the customer as to why their loan was denied and the almighty dollar will continue to rule as it always has.
At a Chicago Booth big data forum in Hong Kong recently, the CEO of Spanish firm Taiger said that your personal data profile is worth about $20 USD. With the amount of information being gathered and the rate at which that data package is growing, we’re thinking that number has nowhere to go but up. The next time you post something on Facebook, just remember that some firm out there is probably using that to “optimize their workflows”.