receives compensation from some of the companies listed on this page. Advertising Disclosure


Crystal Knows All: You're Being Profiled editorial staff editorial staff

Crystal combs your online social media and compiles a personality profile that it sells to businesses trying to reach you.

  • Crystal is a proprietary personality detection technology that makes it possible to find the best ways to approach a client or potential prospect.
  • Studies have shown that Crystal predictions are accurate up to 75% of the time.
  • Although Crystal predictions can be very useful, it can also be beneficial to get to know your potential clients the old-fashioned way.

You don't need a crystal ball to anticipate the best way to approach a client or potential prospect. Crystal knows.

What is Crystal, and how does it work?

Crystal is a proprietary personality detection technology originally developed in the Harvard Innovation Lab with the self-proclaimed mission of working at the intersection of social science, technology and communication. According to Cultured Digitally, by mining various social media platforms for information, Crystal helps companies better determine how people communicate and how they prefer for others to communicate with them. In other words, Crystal technology simply provides an automated method of companies to get to know their potential clients without wasting tons of time and energy to do so.

Crystal can only access information that is being publicly displayed. It mines data that the user has posted as well as data that has been posted about you by others. Crystal's personality profiles are based on analysis of public data obtained from user and peer reviews, LinkedIn accounts, Google, Facebook pages and blog posts. 

The app matches this information to one of 64 unique personality profiles. A person's profile tells you how best to communicate to them, including what words, phrases and even the tone of voice you should use when trying to connect with individuals.

Selena Larson of the Daily Dot describes Crystal as a "psychic for online communication. It walks the line between innovative and super creepy." But as Rachel Metz notes in MIT Technology Review, Crystal is no creepier than the data mining performed by companies such as Netflix, Amazon and Facebook to figure out what movies or books you like, or what ads to display that most interest you.

Apps such as Apple Magic Sauce also translate people's digital footprints (primarily your "likes" on your Facebook account) into detailed psychological profiles. As Kashmir Hill writes for Fusion, "In the digital age, it is very, very easy to profile people … thanks to our ever-increasing online footprints and the explosion of offline data collection."

Personality email checker

What makes Crystal stand out from all the other digital psychics is a "personality spell checker." This is a Gmail extension for the Chrome browser that analyzes email content as you compose, highlights wording it finds ill-suited to the recipient's personality profile, and suggests phrasing more "simpatico" to that person. 

It tells you to use "Hello" rather than "Hi there" when emailing with Jane Doe, and add friendly lines like "hope you are doing well." (There are currently no plans to integrate with other email services, though there is an email assistant that allows you to copy and paste email text for Crystal to examine.)

You also get hints about what to do when working with someone (focus on goals) or speaking to them (don't leave detailed messages); there are also hints on what does or does not come naturally to that person (she feels comfortable making quick decisions; he does not like spreadsheets). Fast Company reports that the long-term plan is to branch out into social media and other forms of online communication.

Crystal's goal, which may seem odd for something that is the result of applying web crawlers and algorithms, is to create a new technology built upon an age-old principle: communicate with empathy.

For business communication, that ability to foster empathy can make the difference between an email that is read and one that isn't. Moreover, it can lay the groundwork for better customer relationships and long-term brand loyalty. Even within the office, better communications among team members and peers can reduce petty squabbles and misunderstandings that can hinder productivity and effective cooperation.

Assuming, of course, that the thing actually works.

How accurate are Crystal's predictions?

According to University of Maryland professor Jennifer Golbeck, personality inference technologies tend to be right about 75% of the time. Crystal hedges its bets somewhat, saying its estimate of accuracy is based on how much information is publicly available for a specific individual.

Elle Hunt analyzed her own Crystal generated profile for The Guardian and found it about right: "Though I am not quite as robotic as Crystal suggests, I am a fairly to-the-point person and, though I delight in rambling, expressive emails from friends … I'd be less enthusiastic to receive one from a stranger." 

Crystal works, but is it a good idea?

There's nothing new about automating customer relationships – there's an entire CRM software industry based on it. Given that an ever-increasing percentage of customer (as well as internal workplace) interactions are conducted online, any tool that can give you an edge is a plus. Besides, your competitors are probably using it. 

But there's something to be said for learning about your customers the old-fashioned way, by just getting to know them. Consider the company that proclaims the policy that every time you call its customer service number, you talk to an actual person and not a computerized answering service. You don't need to read anyone's digital profile to know that's the start of a winning conversation.

Image Credit: Chainarong Prasertthai / Getty Images editorial staff editorial staff Member
The purpose of our community is to connect small business owners with experienced industry experts who can address their questions, offer direction, and share best practices.