$175 Million and Four Million Fake Users: The Story of Frank’s Fraud

Frank was a financial services application which made it easier for students to fill out the FAFSA. Last year JPMorgan Chase filed a lawsuit against Charlie Javice, the founder of Frank alleging that she had misrepresented the number of users and suing her for fraud. Charlie Javice has now had an opportunity to file her counterclaim and argue about why her actions were not fraud.

Read more: $175 Million and Four Million Fake Users: The Story of Frank’s Fraud

If you would like to subscribe to my free newsletter then please go here.

If you want the posts from this blog delivered to your inbox:

I also have a Discord server that you can join here.


00:00:00:01 - 00:00:37:15
Bennett Tomlin
In 2021, JPMorgan Chase paid $175 million for a company called Frank. Frank was a company that helped college students fill out the FAFSA. JPMorgan presumably purchased Frank because they hoped they would eventually be able to turn those college students into profitable Chase customers. And there is nothing Chase likes more than profitable Chase customers. JPMorgan has since filed a lawsuit accusing Frank and their founder, Charlie Javice, of fraud, specifically, they alleged that the 4.3 million users that Frank claimed to have were actually only about 300,000 users.

00:00:37:15 - 00:01:00:10
Bennett Tomlin
When JPMorgan initially asked for proof of their users. They allegedly sprung into action. Charlie, the founder, along with Frank's chief growth officer, Olivier Amar went to the director of engineering to try to convince them to generate the list, but the engineer refused. This allegedly led Frank to seek out a data science professor who they contracted with to make the user list.

00:01:00:13 - 00:01:24:03
Bennett Tomlin
The data science professor was able to generate convincing looking synthetic data for the needed 4 million users, but was not able to generate convincing email addresses, which, speaking as a former data scientist, if you needed them to pass any kind of integrity check, then yeah. This could be a really thorny problem. Like just thinking about it now, you would need to randomize between several different formats, right?

00:01:24:03 - 00:01:49:09
Bennett Tomlin
Like first name, last name, first initial, last name. And then sometimes you need graduate years or you need birth years or just random significant numbers. And you need some nonsense emails too, because sometimes people still use the email they made when they're 12. Even though it's an embarrassing Hotmail address. Then, because it's FAFSA applications, you have to imagine some of these students signed up with their institutional email.

00:01:49:09 - 00:02:11:23
Bennett Tomlin
But then if you want your data set to be really good, you need to make sure that those fake ones you generate for those institutions are in the same format that those institutions actually use. Otherwise, a more detailed integrity check would come back with problems. Some of the emails between Charlie and the data science professor actually point to some of these really salient issues in generating synthetic data.

00:02:11:24 - 00:02:34:29
Bennett Tomlin
Javice asks the data science professor, will the fake emails look real? To which the data science professor responded, They will look fake. At one point, the data scientist talked about how they were adding a check to ensure none of the sample names were real. There's another point where they're debating how to handle addresses, which, again, as someone who has made synthetic data sets before, let me say geospatial data and address data.

00:02:35:05 - 00:02:56:15
Bennett Tomlin
It's the worst. It's the worst. If I never have to deal with another geospatial data set for the rest of my life, I will not be disappointed. The data scientist asked Charlie if they should attempt to fabricate the addresses to which Charlie said, I just wouldn't want the street to not exist in the state. Eventually, the data scientist let Javice know that real addresses may not be doable.

00:02:56:23 - 00:03:30:08
Bennett Tomlin
Like I said, geospatial data is the worst. Anyways, the data scientist was not able to make convincing emails and so they also sought out to marketing firms to purchase lists of college student email addresses. And then they married that with the information that the data science professor had synthesized and presto 4.3 million users. The data scientist sent over an invoice to Frank for $13,300, which was itemized and showed each of the fake data fields they were asked to create.

00:03:30:11 - 00:03:58:20
Bennett Tomlin
Javice then emailed the data scientist to let them know they should change the invoice to just say data analysis. And if they do, then they can increase the invoice amount to $18,000, a nice 35% bonus, and all you had to do was no longer itemize your work. Oh, and that 35% bonus comes after the data scientist allegedly quoted their usual rate of $300 per hour, and Frank apparently decided to double it to $600 per hour.

00:03:58:21 - 00:04:34:01
Bennett Tomlin
So 100% bonus, followed by an additional 35% bonus. Frank was feeling generous, apparently. Whatever this data scientist was doing, it must have been important for the future financial success of Frank. This list was apparently good enough to convince JPMorgan to close the deal, and Frank was acquired again for $175 million. This facade would not last long because when JPMorgan started trying to email the customer list, they ended up with a 28% delivery rate and a 1.1% open rate.

00:04:34:13 - 00:05:03:28
Bennett Tomlin
Among the delivered emails, which is truly awful. JPMorgan does a little more digging into the company they already purchased and then started to put together the pieces, eventually filing this lawsuit. So that's where we were when this lawsuit was initially filed right before Christmas. A couple days ago, Charlie Javice filed her counterclaim. Now, part of the defense is centered around legalese surrounding the merger and the trust and the damages in the contracts.

00:05:04:05 - 00:05:31:05
Bennett Tomlin
And we're going to skip most of that because I'm not a lawyer, never been to law school, never even played one on TV. So I'm going to stay away from that, Charlie, categorically denies all of the allegations, except she also kind of doesn’t. The first part of the document is all about how she denies the allegations, but then we get into the nature of the counterclaims in the response becomes far more interesting.

00:05:31:15 - 00:05:58:28
Bennett Tomlin
I'm just going to read you paragraph eight of that section because it is wild. JPMC now says that despite its extensive diligence and professional bona fides, it failed to uncover a key misrepresentation that Ms.. Javice told its executives about Frank during diligence, and that it only later discovered that Frank only had around 300,000 registered account holders rather than the 4.25 million users that Ms..

00:05:58:28 - 00:06:22:13
Bennett Tomlin
Javice said represented. But this is simply implausible given the public information and market comparables available to JPMC and the detailed information obtained during diligence. As far as I could tell, and again, I'm not a lawyer, this part of the defense is basically it doesn't matter what Shirley told Morgan, they would have to be stupid to believe it. Are you stupid, JPMorgan?

00:06:22:25 - 00:07:00:20
Bennett Tomlin
Is that what you're saying? Are you guys a bunch of stupid idiots? Which is a new fraud defense technique for me. It continues in paragraph ten with even a cursory look at publicly available data. Makes JPMC’s allegation that it was led to believe Frank had 4.25 million registered account holders implausible. This allegation falters quickly when one looks at the value of comparable companies that do have that number of registered users and the market share that this number of registered student accounts would have implied, both of which would call the purchase price that JPMC paid for Frank into stark question.

00:07:00:21 - 00:07:23:13
Bennett Tomlin
That number of registered account holders would also be close to impossible given the limited number of employees and marketing spend that Frank represented, which would both have needed to be many multiples of Frank's numbers to support such a larger enterprise. This paragraph seems to basically be saying we would have been worth far more money than you paid if anyone actually believed the user numbers.

00:07:23:13 - 00:07:51:14
Bennett Tomlin
We were promoting. And I mean, obviously we weren't big enough to handle the number of users we said we were handling. Are you stupid? JPMorgan. Paragraph 13 is where Charlie reveals she was repeatedly investigated at JPMorgan, first for spending irregularities in her account and then for relationship building activities that she was engaged in. I have no idea what that could possibly mean, nor do I see how including this benefits her case.

00:07:51:14 - 00:08:19:13
Bennett Tomlin
But again, I'm not a lawyer. I mean, okay, she's including it because JPMorgan says she was fired with cause and she needs to prove it was not for cause so that she can retain more of the money from the merger. That's what it comes down to, really. So basically, Charlie maintains that JPMorgan knew they were buying a company who didn't have the users it promised and that because they knew that or should have known that this whole thing is manufactured.

00:08:20:03 - 00:08:39:08
Bennett Tomlin
And if anyone knows about manufacturing things, it's Charlie. Okay. So now that we've covered the facts, let me give you what I think happened here. Chase needs more young people so they can keep paying their bankers massive bonuses and handle the fines from their myriad lawsuits. So they set out to go and buy some fancy new, fast growing startups.

00:08:39:11 - 00:09:08:03
Bennett Tomlin
Charlie alleges this was specifically a Jamie Dimon initiative and that because it came from him with urgency, the teams were really interested in closing this part. Frankly, seems quite plausible to me. Deeply entrenched companies love to grow by buying the markets they like. The next thing I think happened and Charlie kind of alleges this too, is that JPMorgan's due diligence was kind of half assed because they really wanted to close these deals.

00:09:08:03 - 00:09:31:26
Bennett Tomlin
As someone who has found cryptocurrency white papers with a math literally does not add up. We've already raised millions and millions of dollars from top flight firms. Frankly, I believe that JPMorgan's due diligence was bad. So JPMorgan decides to purchase Frank, even though, as Charlie pointed out, even cursory due diligence would highlight how Charlie had been lying to them.

00:09:31:27 - 00:10:00:14
Bennett Tomlin
Eventually, things go badly, possibly including some legal issues related to changes at the Department of Education and suddenly for JPMorgan, the embarrassment of failed due diligence becomes less than the potentially recoverable value from Frank and Charlie. And so they start their lawsuit. Now, fundamentally, the truth is the JPMorgan lawsuit includes multiple public statements and Frank team emails that lay quite clearly that there were misrepresentations made.

00:10:00:14 - 00:10:22:18
Bennett Tomlin
Even Charlie's defense seems to basically be I deny everything except the text of the things that I can't deny because it's objective. And also it doesn't matter what I say because only an idiot could believe me. I cannot speak to the specific legal issues at play in this and how the contracts were actually created. But this has to be one of the funniest defenses for fraud I have ever seen.

00:10:22:24 - 00:10:27:15
Bennett Tomlin
Sure, maybe I lied, but God, you idiots believed me.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.