Can Crowdsourcing Prevent Credit Card Fraud?


Yaron Samid believes the best way to combat credit card fraud and errors is not just to use data, but use lots of it. Samid’s start-up BillGuard, just came out of stealth mode and is looking to build a fraud protection system that leverages a consumer’s own data history along with signals gathered online and through banks. By constructing smart algorithms that digests complicated transactional information and pulls in alerts from both existing members and the web, BillGuard believes it can build a crowd-sourced anti-virus system for bills.

The idea has won the New York company $3 million in angel funding from Ron Conway, Chris Dixon, Howard Lindzon, Bessemer Ventures, IA Ventures and Yaron Galai. And BillGuard, which also has a research and development office in Israel, was recognized as the top start-up at the Strata data conference last week. This is the latest start-up for CEO and co-founder Samid, who previously co-founded Pando Networks, a scaleable content delivery solution. Raphael Ouzan, a big data mining and security expert, is the other co-founder.

BillGuard works like this: consumers register their credit and debit card accounts with BillGuard, which ingests all the transaction information and organizes it in one database. The company will allow users to flag each transaction for questions, errors or fraud. Those user alerts will get combined with other bank data and notices on recent fraud activity along with web chatter from Twitter and consumer forums. Customers will then get notified via e-mail if they have a questionable, fraudulent or erroneous charge on their credit cards.

For the vast majority of credit card users who don’t scrutinize every transaction, BillGuard brings an extra level of protection. Even power users who check every bill can benefit: BillGuard provides added peace of mind and also the ability to share the fruits of their vigilance with others. It’s not just for fraud but it also addresses merchants who pack in extra charges on to bills knowing that most consumers often won’t notice.

“We’re discovering lots of possibilities by crowdsourcing consumer vigilance,” said Samid in a phone interview. “Its an entirely new data-set that we’re structuring but the real IP is the back-end algorithms that make that data actionable. We’re training computers to understand the language of transactions. It’s a lot of messy data coming from the banks.”

Banks already have fraud protection, but Samid said much of that involves broad pattern detection that often can’t uncover more complex schemes or persistant errors. For example, Samid said his debit card number last month fell into the hands of scammers, who quietly placed a legitimate looking $8.95 charge on his card. BillGuard looked up the transaction information and when it couldn’t identify the merchant online, flagged the payment, allowing Samid to correct the situation.

BillGuard is currently in alpha with friends and family and is looking to open up to a beta in the second quarter with a more fuller featured offering. The company has not announced pricing but Samid said there will be a freemium model, so users will be able to try at least some of the services for free. He said while a standalone service will always be available, BillGuard has been talking to banks and financial institutions, who are receptive to the idea of incorporating BillGuard into their online services.

BillGuard is a good example of the ways companies can take small bits of consumer information and apply big data techniques of analysis, visualization and machine learning to create valuable services. The company will need to get consumers to agree to hand over their financial data, which could be affected by how much the premium services costs and how much they trust BillGuard’s security. And the service’s effectiveness may be somewhat dependent on its ability to recruit hyper vigilant users, who can flag transactions that benefit the entire user base. But if BillGuard can scale up and tap into a large customer base, it could be a very powerful way to help combat online fraud, not just for bank and credit card bills but a wide variety of billing services. If you’re interested in big data, GigaOM will be hosting its first annual Structure Big Data conference in New York March 23.

Related content from GigaOM Pro (subscription req’d):



Good lord. Crowdsource my personal financial information in order to absolve myself of basic financial responsibility? And won’t this just give scammers a working sandbox to test new techniques on their own credit cards?

Maybe next will be a webcam-based “ChildGuard” in my kids play room so I can go get groceries and crowdsource the responsibility of ensuring their safety?

Yaron Samid


To clarify, you do not crowdsource your personal financial information. No one sees your card activity but you. Similarly to how your email service scans your inbox for incoming spam identified (flagged) by others in *their own* inboxes, BillGuard scans your bills and alerts you when a charge on your bill has been flagged by others on *their own* bills. You’re still responsible for your bills, we just help bring the questionable stuff to your attention. Does that make sense.

CEO, BillGuard


I guess that is clearer: I give you access to my financial data in order that your service can identify patterns based on charges that others have flagged as suspicious. But isn’t that predicated on users (*some* users, others if not you) still reviewing their financial transactions carefully?

If they already do, they probably don’t need to share their sensitive financial data with you and are unlikely to require the services – Vlad’s point I guess. If your users don’t typically review personal financial data, then how do you ensure enough flagging activities to make the system effective?

I gather the crowdsourcing ingenuity is in a very small number of diligent users doing the work for all the rest? But again, if I were one of those diligent users, it would take some seriously compelling evidence that your system would add valuable security to make it worthwhile to share financial data. Seems like a bit of a bind to me.

As a service built into internal credit card security monitoring, ok. But “hand over [my] financial data” for a poor value proposition? Some people, other people – lots of them I hope for you – but not me.

Yaron Samid

@Dave: This is an excellent question. I asked myself this very thing when I first heard about a ridiculous idea called Wikipedia. Why would anyone take the time to share their knowledge with others with seemingly no direct benefit to them? The answer in Wikipedia’s case is that someone, somewhere out there is really passionate about Hawaiian Bumblebees and wants the world to have an accurate account of what they really are. Turns out that with just 10% of folks creating/curating content and the remaining 90% just enjoying it, you’ve got yourself the most accurate, comprehensive database of information in the world.

In BillGuard’s very similar case, diligent users (“spotters”) that anyway check every line item of their cc statements (about 10% of Americans) have several incentives for using BillGuard and flagging/editing/commenting transactions:

1. Save Time: Any input you give BillGuard about transactions on your bills makes BillGuard smarter and better at protecting YOU. The system learns the DNA of what you consider to be ok, unsure and bad transactions and helps prioritize the stuff you should look at first. The UX is akin to Google Priority Mail. With a few clicks you’re going to see your transaction list in a whole new way and be far more efficient at checking your statements.
2. New Clarity: BillGuard decodes cryptic transaction descriptions (that only a bank computer could love) into human-speak and adds rich metadata and media to help you easily make sense of every transaction. Some really cool work here that I can’t announce yet.
3. Tell The World: Turns out when you’re pissed about being taking advantage of by a merchant/scammer, you want to tell the world. You don’t have a voice today to do so on scale. BillGuard is your amplified voice to tell every other customer that’s been hit.
4. Get Credit: We will be recognizing and rewarding top spotters and their contribution (knowledge and savings) to the community. More on that when we launch our beta.

That said, verifying transactions in BillGuard is completely optional and the service does not rely on just crowdsourced data. There’s a ton of data and complaints online about charges that currently appear on your bills and BillGuard is structuring all that data into one massive, actionable database. We’ve also got patent-pending data mining algorithms being diligent for you; auto-magically spotting bad transactions without you doing a thing.

Someday your bank will have “BillGuard Inside” to alleviate any hesitation you might have with a startup, but until then we want you as a user Dave. We’re facilitating this consumer advocacy movement but BillGuard is being built collaboratively with our users. We need more of your feedback. Please email me directly at if you’re interested.



I can spot fraudulent transaction myself, w/o any complex data mining algorithms.

Thanks. This service has no value

Comments are closed.