My wife, a brilliant numbers person, spent many years in the classroom as a teacher, until I finally convinced her to forgo the incredible pay and astounding benefits in exchange for life as a highly-in-demand math tutor. As a tutor, she typically gets two kinds of kids: the ones who can use a little (or a lot of) help, and the ones who are already killing it but have parents that want them to kill it even harder.
Math is neither good nor bad, unto itself. It simply IS. You can’t say that “2 + 2 = 4” is a good thing. It’s simply a fact. But the ability to perform that calculation, and calculations far more complex, is in fact good.
One standard line my wife doesn’t hear nearly as much as she used to is, “When will I ever use this?” It was always the last excuse for the kids who just plain didn’t want to deal with the subject.
The knee-jerk response is typically, “You need to be able to balance a checkbook.” Although the mobile banking app we’ll all be using eventually (as many of us do already) will take care of that, right?
A little over a year ago, I checked into a hotel in San Bruno, CA. Their system was down, and the customer service rep at the desk was having a hard time subtracting my deposit of $40 from my total check-in amount of $180.47. I gave her the answer, but she didn’t seem to trust my number. She then complained that she had a new phone and hadn’t yet found the calculator app. I ended up doing it on paper for her, and convinced her I wasn’t trying to commit fraud. A little math capability there would have gone a long way.
Just for laughs, I recently sent my wife a picture of a whiteboard from our headquarters, on which a subset of the data science team had been working out the start of a new anti-fraud model. It was a doodling of various formulate in which the solutions to various calculations bubbled into one last calculation. As our CEO likes to say, “It’s just math!” Yeah, and it’s just brain surgery.
I realized that this represented another response to the “When will I ever use this?” question. We use data sources to validate that an individual data element is a real thing: a legitimate phone number, a legitimate email, a real name and address. And then we validate that the data elements belong together. A fraud actor might use your good credit info, but his own email and phone, so he can speak to a lender on your behalf to commit fraud.
Applicants for loans, credit cards, bank accounts, etc. are a mixed lot of quality; good, bad, and who-knows. They can be good as in, they have decent enough credit to successfully apply. And presumably they’re also good in that they’re not crooks. They can also be bad in that their credit is lousy, or that yes, they are in fact crooks. As for the who-knows crowd, they might be good, but their application is incomplete, their background is incomplete, or perhaps their background isn’t very deep because they’re millennials.
So this is where our math kicks in for identity verification. We calculate the likelihood that an identity as presented will commit fraud. You could say that this is only a quantitative thing, as it’s based on numbers. It’s neither good nor bad, it simply is. Well, sort of. Based on our comparative results, our numbers are far and away superior to any competitor’s.
We calculate the likelihood that the applicant is who he says he is. We can see if the various bits of data he’s supplied all belong to the same person, to ensure he’s not posing as his brother-in-law, who has better credit and no rap sheet.
We calculate the quality of an applicant who is honest and hard-working, but who doesn’t have a very deep credit file, perhaps fresh out of school, or only in the working world for a couple of years. He may have a good job and good intentions, but doesn’t have history.
The quality of the math means so very much to our customers. Good numbers of good applicants, who might get erroneously rejected just because their applications aren’t up to snuff, mean more business. Bad numbers related to fraud hurt the bottom line. The goal is to eliminate the bad applicants, and find as many good applicants as possible.
That’s our Socure secret sauce - the AI-based machine learning we employ to generate the most accurate models in the industry, to make customer applications a qualitative proposition. We turn a user profile - the data passed by an applicant for credit - into true value for our customers. If we help reject an applicant because he is clearly a risk, then we save our customer a loss. If we help avoid a false positive that would otherwise reject a good candidate, then we bring our customer additional business.
Stopping bad guys has value. Allowing worthy applicants to get a loan or a credit card has value. Just math? Sure, it’s still increasing good numbers and reducing bad numbers, but these numbers are a good thing. They have quality.
So when am I actually going to use this stuff? To operate an AI-based company. To stop the bad guys. To help the good guys. Yeah, that’s a good thing. A very good thing.