“Ever wonder if someone you like likes you back? Or how your friends, or even family members feel about you?” Chances are, most of us have asked at least one of those questions to ourselves at least once in our lives. One way to find out would be to ask, but that wouldn’t be very 21st Century of us, would it? Realistically, we already have the answers sitting among the 6 billion texts we send each day in the US alone. Rather than analyzing the texts ourselves, or passing them to a friend to analyze over brunch, there is of course at least one startup that specializes in that kind of snooping.
Crushapp uses some smart algorithms and machine learning to perform a pretty straight-forward job: interpret your texts to determine how much people actually like you. Crushh is just like that pal to whom you hand over your phone for honest feedback on a conversation (“What do you think this means????”), except now that our phones are our best friends, Crushh is a data-driven app, not a human.
To use Crushapp, the user first selects a contact from their phone. The app then analyzes the text messages sent between the user and that contact. From the data, a score is generated, complete with charts and graphs enabling users to not only see who likes who how much, but precisely when the relationship balance changed over time. In addition to the automated metrics, users can also manually enter details (gender, age, location) to improve the accuracy of the results.
All the analysis is done “secretly” so only the Crushapp user knows the analysis is being performed. Once the messages are analyzed–which apparently takes only seconds–a score from 0-5 is generated, along with some other measures of interest and engagement. The scoring scale is as follows: 4-5 = they like you way more, 3-4 = they like you more, 2-3 = they like you equally, 1-2 = you like them more, and 0-1 = you like them way more. Ah, yes–the ol’ “if you like them more than they like you, that’s a bad thing” implication. Where would our self-esteem be without that handy perspective?
“Effectively, the app uses algorithms and machine learning to interpret the many aspects of a text conversation,” explained app creator, Shi Wen. “It has the ability to learn the behavior of the user and then with the information available, try to make an assessment of how much each party likes the other, relatively.” If the user is certain that the application analysis is wrong, they provide feedback to that effect and the app will incorporate that data to improve accuracy over time.
According to Wen–and due to the reliance on machine learning, not surprisingly–user adoption is the key to success with the app. “We think we’re providing a unique service that everyone can relate to at some level,” says Wen. “So having Crushh become a household name would be the dream outcome. In the near term, we just want people know such a service exists.” At a higher level, Wen hopes that the insights gained from the app will help people better understand and improve the way they communicate with those about whom they care.
At the moment, Crushh is in open beta testing with a full release to follow (no specific dates available for that). It has been in development since January 2016, and is currently available on Android only. When the final version is released, the app will become invitation and referral only.
Charles Costa is a marketing specialist helping technology companies grow, one word at a time. You can learn more at CharlesCosta.net
- This Smart Home Sensor Auto-Adjusts Rooms To Who’s In Them - October 12, 2017
- Meet Sunu Band, The Wearable Helping the Blind Navigate Their Surroundings - August 21, 2017
- Powerful Startups: Ameren Accelerator’s 7 Latest Energy Innovation Inductees - August 4, 2017
I would be completely nervous to even log into this thing!
Great, another something else online that is telling me that I suck 🙂
You are thinking the exact same thing that I am!
I have so many friends that complain that nobody ever replies to their messages. The only thing that I could tell them, before now, was that they send too many texts to keep up with 🙂 Now I can direct them to the app and the app can give them the bad news