Join now Safety Press Kit Tour Affiliates Contact Blog Login

Scametrics: What is it? How does it work?

WARNING: Scammers operate on our website!

Shocked that we have scammers on our site?
Shocked that we actually tell you we have scammers?

We do not see the point in pretending that everything is lovely and nice. We believe the best way to combat dating scammers is through eduction of our members on what to look out for. In the same way you cannot stop spam email, you need to learn what's genuine and what's not. The romance scam is big business and figures suggest many millions of £/$/€ is lost every year.

Scametrics helps

Scametrics is our in-house system that monitors and learns the patterns and behaviour of the romance scammer. This system automatically deletes hundreds of profiles every day. Where we feel we lead in this area is that we don't hide it behind the scenes. Every member can see other members' Scametrics scores. It's visually shown on a member's profile page as a circle. A full circle is a 100% clean record. An empty circle means the account will be automatically removed.

Scametrics circle examples

Identities and location protected for privacy reasons

How does it work?

The most important part of the system is its ability to learn. For example, a scammer may use a disposable email address (an address that needs no registration and therefore cannot be traced). There are thousands of disposable email services online and knowing all of them is near impossible and in addition, new ones will appear every week. Scametrics takes the second half of an email address used to create an account and stores it within a database. If it finds a new one it adds a new entry, if it finds an existing one it increases the count for that email domain. When an account is banned for being a known scammer the ban total for the scammers email domain is incremented by one. Over time this builds a trust value for a particular domain. All registered users are then compared with the database and are deducted Scametrics points based on their email address. The following email domains have been detected by Scametrics automatically as prolific contributors to fake profiles.

@trbvm.com

@yopmail.com

@sharklasers.com

@grr.la

@armyspy.com

@dayrep.com

@einrot.com

@cuvox.de

@gustr.com

@rhyta.com

@dispostable.com

@teleworm.us

@jourrapide.com

@superrito.com

@vomoto.com

@fleckens.hu

@imgof.com

@naij.com

@qisoa.com

@abyssmail.com

@anappthat.com

@grandmasmail.com

@tafmail.com

@soisz.com

@kiois.com

Email address is just one component of Scametrics. A similar process is carried out for almost all types of data that could be considered unique for an individual users. Some of these we are happy to share, others remain a trade secret to ensure the system is as effective as possible. Tag line (short description of the member) is another area that is tracked, the top 'scam like' tag lines that Scametrics has learnt are:

Tag	Ratio of bans : usage
hello am looking for soulmate	0.92
iam looking for serious relationship	0.85
looking for my soulmate	0.82
hello am looking for my soulmate	0.75

Example of tag lines used by scammers that have been learnt automatically by Scametrics

Not everything is equal

Scametrics monitors and learns from may different types of data but the penalties vary depending on data. For example, IP addresses are not consider unique enough. Many people think that the IP address is the secret to banning people but it's not. If you use a mobile phone network (about 70% of our members do) then the IP address is often shared by thousands of people and may change with every connection and very few home broadband providers give a customer a static IP address. Therefore our scoring of IP addresses will have less impact than tag lines.

Old fashioned reporting

We also don't ignore basic functions such as a report button. If another person reports a member then they will also loose Scametrics points but there are additional safety measures to stop this being abused to protect members from being deleted by repeated use of the button.

Big data

All of this processing happens on every page load, every message, every login. At present Scametrics makes about 30 decisions a second and has a knowledge base of over 1.2 million lines of information. More server CPU power is used in scammer prevention that actual online dating!

Summary

It's important to note that Scametrics is not a recommendation of who's genuine and who's not. It offers our members an extra piece of information that they can use to make an informed decision of the legitimacy of another member. Scametrics ability to learn ensures that it can keep pace with fraudsters. Exposing a visible score and not hiding the fact that scammers operate on our site is a very important part of helping members have a safer online dating experience.