Shine we together: A innovative dating site using 2012 Nobel Laureate Roth's algorithm
Abstract
Our dating site introduced scoring and its related functionalities innovatively, conforming to people's natural inclination to score people of the other sex, while other sites simply ignored. In addition, we use the 2012 Nobel Laureate El Roth's algorithm to compute a fair match set, and recommend the match to all relevant users. This recommendation can help the users getting closer quickly. With the introducing of scoring and Roth recommendation, a new and alternative usage pattern of dating site is given birth to, such that, the new pattern is more efficient than the old pattern. Thus users will perceive our site as to better serve their ultimate goal in using dating sites, i.e. to get the best possible mate in an as large as possible scope. This is especially so true for dating users in large cities. We also have a true unique 'settle-down' recommendation, which digs out values in users' past dating experiences, which we consider valuable since it takes users' energy and time to get, while other dating sites simply dumped as garbage. We have also many other important shining points, like, viral propagation mechanism, powerful and flexible search, eyeball catching ads words, etc., some are good for attracting new users, some are good for retaining old users, when combined together, we shine brightly and kick ass. Now it's time to join or invest us, so we shine together.
(-> 中文:Chinese version)
It is true there are many dating sites already. But users of dating site have the least loyalty to the site, and they are constantly seeking new ways and exploring new possibilities, they are just not satisfied with current dating sites. Dating site also has clear business model and is a pretty good market in terms of size
(see: online.wsj.com/article/PR-CO-20130430-916851.html?mod=googlenews_wsj). Dating site usually has big internet traffic flow and people awareness, so is good for portfolio building and online presence building. We believe if a dating site can achieve the followings, then it is not hard for a new dating site to stand out, possibly even come out on top. These are:
1. Novelty.
2. Usefulness, addressing users' true need, helping them to reach the ultimate goal, providing convenience and some type of saving (whether it be time saving or effort saving), or adding value.
3. Good viral propagation mechanism. (This will save ads cost and shorten the process to stand out)
4. Shining. It must shine to the users.
I now show you that our site achieved the above 4 points substantially, plus, our site uses sophisticated technology and uses patent protection, so it is hard to copy.
Our dating site innovatively and better addresses some fundamental mentalities of online dating users
First off, I want to raise some fundamental questions about current dating sites for you to think about:
1. Have you ever scored your opposite sex classmates in student time, even argued with your roommates in dorm over who should be scored higher? You have, I guess. But such natural and common human nature or behavior is simply turned a blind eye to by all dating sites and no dating site (as far as I know of) provides the scoring feature? Why?
2. Why are people using dating site (although they are not quite satisfied with them)? What is the users' goal in using a dating site? What is the fundamental value of online dating for users? How well or badly current major dating sites do in providing those values?
For the second set of questions, it is that dating sites vastly increase the scope from which people can get a best possible match, and at a better efficiency than the offline dating could ever reach. The users' goal in using dating sites is to get the highest possible scored (in another words: preferred) mate that also scores back him/her high, in an as large as possible scope. This basically reflects human's greedy aspect of natures. It does so by having a large users database, and the search, browse, and interacting functionalities. The usage (user activity) pattern is: finding, probing, waiting for feed-back, repeating the foregoing, choosing the best among the positive feedbacks. User can relatively quickly find many that look good to them, but a match is a pair that look good to each other. To find out each prospects to whom you(the user) look good also, you have to probe (by wink or message etc) and wait for a feed-back, this usually requires many round-loops, and is time-consuming and often results in setbacks. Furthermore, user can get many positive feedbacks, and there could exist still more potential positive feedbacks in the un-explored, un-probed or un-answered-yet crowd, so it is usually hard to know if a prospect returning a positive feedback is the best possible match for the user or close to the best possible, so the user is still a mile away from the final goal of getting the best possible match. So, the conclusion is, current dating sites do satisfy users' fundamental needs to a certain degree, but not quite well. It is at most a half-done job.
Our site, by providing two key unique features sets, scoring and Roth (2012 Nobel Laureate) recommendation, satisfies the above two sets of basic or natural human natures better. The scoring features set, satisfying people's natural inclination to have preference on people of opposite sex, allows users to set score back-to-back, overview and compare prospects in a score-sorted list, adjust the scores, and reuse the scores (saving the user from having to remember his/her preference order). The Roth recommendation, using the stable matching algorithm (Roth and Shapley's major contribution area for the prize), compute a match set from the users that have scored each others (i.e. they see each other as prospects), and then present the result to the users. Here "fair match" means, the matched user in a pair has the highest possible score in the others' score list. So, how do these features help user to achieve the ultimate goal of getting the best possible match in an as large as possible scope with better efficiency? Simple, with our powerful and flexible search, you(the user) find as many as possible prospects, and then score as many as possible of them by your preference, you can send them winks simultaneously to remind them to score you back. Then you are done. When the other users receive your winks, the system prompts him/her to score back if you look good to them also. Then the system uses Roth algorithm to find the best possible fair matches for all users and present the matches as recommendations to all the relevant users.
As described in previous paragraph, we actually created a new usage pattern of dating site. This pattern only consists of: finding, scoring, repeating the foregoing, getting recommendations. The new pattern is extremely short and efficient when compared with the old one. Of course users of our site can still follow the old usage pattern. Users can also use a mixture of old pattern and new pattern at their will in our site. According to the new pattern, users can keep on finding more prospects and scoring them until the users are tired or satisfied, do not have to probe and wait for a feed-back, saving many extra steps. This means an increase in efficiency and user can reach a even larger scope of prospects, which is in line with people's greediness. This is especially an advantage for dating users in large cities like Beijing, Tokyo, New York, etc.
Regarding the ordinary dating site's value, "vastly increases the scope from which people can get a best possible match, and increases the efficiency of the process to get this", with our new usage pattern, the scope and efficiency are increased to a next level !
Now come back to my first question. The reason that all dating sites are not providing the scoring feature, is that they do not connect the dots of the scoring and Roth recommendation. Without Roth recommendation, the value of scoring is limited. Scoring does have some value, but it does not shine alone, on the other hand, it also need a little extra effort for users to input data and for them to develop the feature.
With the adding of Roth recommendation, scoring plus Roth recommendation has the effect of 1+ 1> 2, the combined benefits of scoring and Roth recommendation outweigh the effort for user to input the data, the net effect we achieved is the considerable increase in efficiency for user to get the best possible match from a larger scope in a large dating site's users database. When the two features join forces, they shine.
A truly unique "settle-down" recommendation feature
In online dating, it is quite common for users to date multiple prospects over an elongated period without a success of settling-down with someone. It is also common that, among those a user has dated, there are some for which there is only some minor issues or no obvious issue between the two, but unluckily, the two just did not make it or just shelved there temporarily by one or two sides, however they are actually acceptable or close to acceptable to each other. A Chinese idiom: "好事多磨", which translates to English as: "good things often take turns and twists to finally succeed", accurately describes this phenomenon. The system can look into users' score lists, so can get all pairs that are of this kind, since they left each other with a good or good enough score after they have dated. From the set of all pairs of this kind, the system compute a fair match set and recommend it to all the users involved. This settle-down recommendations (temporarily so named) might be a good choice to users. Without this recommendation, the people involved in those pairs could very possibly never know that they are actually mutually acceptable and could miss a good match forever. All other dating sites probably assume the peoples users have dated and not resulted in a settling-down with are trash.
The "settle-down" recommendation digs out value in users' past dating experiences. Users' past dating experiences take users' time and energy to get and are truly valuable, while the question of how to manage prospects that users have already dated and shelved temporarily is left to the users themselves totally by all other sites. Basically all other dating sites let the user to always look forwards, never look back, that is to say, someone better is always in the future, in un-explored crowd, never in the past, which may be true if users can live an infinite lifetime.
Furthermore, this settle-down recommendation, can also be used as a reference point even if user is making dating arrangements with new prospects.
We have a privately developed parallel stable matching algorithm to deal with the problem size issue of a dating site.
A back-to-back preference list is probably an important contributing feature for BangWithFriends' initial success.
Now I show you some screenshots so you can get more concrete ideas about how the above process works.
(Please note, the pages are prototype pages just out from the programmers' hands, not touched up by a web page designer yet, and the touch-up can be done any time later relatively cheaply and easily, so do not bother with the color, layout, font stuff etc. I know this since I once worked in a team mixed with developers and designers.
Also note the currently screen shots is only from a past design. A new design has been laid-out, the new design is cool and even fun to play with. This is work in progress, and is being constantly improved, as is the case for any innovative attempt.
The whole thing is also fairly complicated. Any questions or doubts, communicate with me please, maybe it's not problem at all.)
The Contacts Page
The contacts page by itself is unique among dating sites, it gives the user an all inclusive view of other users having various different types of interactions with the current user.
The contacts page facilitate users to sort (including sort by score), overview, make comparison, and edit the preference list, and interact with users in the list. These things are real helper if users interact with many prospects. Without the help of this functionality and some other related ones, when the number of other users that users have interacted with gets bigger, they often forget the preference context and/or interaction history.
We also have many other pages that display views of specific types of user interactions, such as sent/received winks, etc. Our users' backstage photos permission management are also more convenient to users than other sites'.
The User Profile Page
Where users view all information about other user and interact with other user...(temporarily reserved.)
The Dash-board page
Where Roth recommendation is presented.
Our truly unique and exceptional search system, with the following characteristics
- Powerful and yet flexible search system, catering into users' search needs at all different levels.
- More ways of sorting, supported by scalable architecture and highly optimized code to ensure a good performance at large users load. (Technically difficult to implement for site with big user base)
- A very unique Best-Match search, in which the match degrees to both sides are weighted and summed up, and then sorted and displayed in a paged view. (Any one in this technical field know this feature is really hard to implement, it is an un-matched feature of our site only.)
- A search manager list all types of searches. All types of searches can be saved, duplicated, edited, re-used. This makes it really easy for users to try different searches, users can just play with them.
A well designed virus propagation mechanism
This ensure users to take initiatives in promoting our site, lower our advertisement cost.
...... (un-published product, temporarily reserved)
The users groups that favor our site mostly
All types of user will like our site to a certain degree. Some special users groups will lie us more. These are:
- Men and women who remained single for a long time
Some of our unique functionalities are especially helpful to them. This group of users are ever increasing in population. - Nitpicking users, e.g. mid to high end users
Our especially powerful search cater into the taste of this group of users. Roth recommendation and the efficiency of new usage pattern are also attractive to them. High end users, although small in quantity have traction for ordinary users. - Online dating users in big cities
Roth recommendation and the efficiency of new usage pattern are attractive to them
How do we get new users?
- Our Ads words is: "Let Nobel Laureate Roth help you to find the one you deserve from prospects of millions". How does this sound to you? It is well known that dating users are always curious to dating site with novel concepts. A small amount of money with this ads words will quickly get us to a thresh-hold users base.
- A well designed viral propagation mechanism ensuring we roll bigger and bigger like a snowball once we have an initial users base.
How do you retain users?
Users will perceive our site as to better and more efficiently serve their ultimate goal in using dating sites, i.e. to get the best possible mate in an as large as possible scope. This plus other our shining points, when comparing with a site with equal users base and fee level, our site will be better at retaining users.
Dating or Marriage is fundamentally a fair deal, we address this fairness directly and efficiently. It is interesting to compare the relationship between our new usage pattern and the old usage pattern of dating site, with the relationship between the fully-automated New York stock exchange clear-house and a flea-market. In the first member of each duo, you have a fully automated deal brokerage which guaranty fairness with high efficiency, in the second member of each duo, each user is both a customer and a vendor, they explore, probe, bargain, one by one.
My opinion about personalities match algorithm of some dating site
I think personality test and match of some dating sites are marketing gimmick, it may be good for attracting new users, but not really useful. Do they really have scientific experiment data to show that the pair recommended by their algorithm will really get along better than average, and will live a better marriage life? Do users really believe what they are told? Our Roth recommendation will not lose in both attracting new users and producing real value for existing users.
Conclusion
Now I have shown you some of our shining points. Unfortunately, right now, I have to keep it secret some very important our shining points temporarily, until the time that we get to know each other better (I am looking forward to this, how about you?). I think you do understand our concerns.
Many investors are looking for disruptive product in big market. I think dating is big enough. But is our product disruptive? I can't make a 100% sure promise, but I am pretty sure about one thing, i.e., our site with so many shining points, some are good for attracting new users, some are good for retaining old users, especially those features which touch and serve online dating users' mentalities better, when combined together, it shines so brightly, such that this venture, at the least level, will not fail, and in the best case, we disrupt, you win big! It is also to note the ratio of size of dating market and the capital needs of our site.
Right now is the best opportunity to invest in this kick-ass product or join this kick ass team.
Appendix: Preliminary business analysis
A plain site like this: www.okcupd.com , is worth of $50M, check here: en.wikipedia.org/wiki/OkCupid, you can also google it yourself.
I think my site will do a lot better than that, so I estimate the final value will be from $50M to $5B. And suppose the risk factor is 10, the minimum value is around ~$5M.
With our truly unique features, eyeball-attracting ads, viral propagation mechanism, it is highly possible we can achieve a middle point, say we make it to somewhere around $500M.
Currently the site is all major functionalities done. Need Chief Front End Engineer, Chief Operating Officer or Marketing Director, and Investment. Contact me at:http://weibo.com/u/1999451503
( All rights reserved, No parts should be reproduced without explicit permission from the author. )
Shine we together: A innovative dating site using 2012 Nobel Laureate Roth's algorithm的更多相关文章
- 机器学习实战(一)kNN
$k$-近邻算法(kNN)的工作原理:存在一个训练样本集,样本集中的每个数据都存在标签,即我们知道样本集中每一数据与所属分类的对于关系.输入没有标签的新数据后,将新数据的每一个特征与样本集中数据对应的 ...
- k近邻法的实现
k近邻法 模型 使用的模型实际上对应于特征空间的划分.模型的三个基本要素:1.距离度量 2. k值的选择 3. 分类决策规则决定. k值的选择:k值的选择,k如果选择的过小会导致过拟合,模型会变得复杂 ...
- POJ 2752 Seek the Name, Seek the Fame [kmp]
Seek the Name, Seek the Fame Time Limit: 2000MS Memory Limit: 65536K Total Submissions: 17898 Ac ...
- POJ 2752 Seek the Name, Seek the Fame
传送门 Time Limit: 2000MS Memory Limit: 65536K Total Submissions: 14761 Accepted: 7407 Description ...
- poj-------------(2752)Seek the Name, Seek the Fame(kmp)
Seek the Name, Seek the Fame Time Limit: 2000MS Memory Limit: 65536K Total Submissions: 11831 Ac ...
- poj 2752 Seek the Name, Seek the Fame【KMP算法分析记录】【求前后缀相同的子串的长度】
Seek the Name, Seek the Fame Time Limit: 2000MS Memory Limit: 65536K Total Submissions: 14106 Ac ...
- 【POJ2752】【KMP】Seek the Name, Seek the Fame
Description The little cat is so famous, that many couples tramp over hill and dale to Byteland, and ...
- poj2752 Seek the Name, Seek the Fame
Description The little cat is so famous, that many couples tramp over hill and dale to Byteland, and ...
- Seek the Name, Seek the Fame(Kmp)
Seek the Name, Seek the Fame Time Limit : 4000/2000ms (Java/Other) Memory Limit : 131072/65536K (J ...
随机推荐
- Delphi的String内存结构(够清楚) good
变量s的内存结构为(字符串编码)A8 03 (字符宽度)01 00 (引用计数)FF FF FF FF (字符串长度)0A 00 00 00 (实际内容)31 32 33 34 35 36 37 38 ...
- 内核空间和用户空间的分界 PAGE_OFFSET
PAGE_OFFSET 首先看看PAGE_OFFSET的功能 内存映射 | 用户空间 | 内核空间 | |——————+———— ...
- HTML 页面载入 Flash 插件的几种方法
前言 之所以写这篇文章,主要是由于组长给提的一个新的需求--使用浏览器调用电脑的摄像头,来实现即时拍照的功能.在网上查了非常多资料,由于这样那样的原因,终于选择了使用flash插件来调用pc的摄像头. ...
- 参加2013中国软件开发者大会(SDCC)会,听软件开发趋势
1.SDCC 盛大召开的会议,既然参加了,就写篇博客记一下. 2.蒋公子 首先向大会主席台走来的是csdn老大...... 额,好像不是走过来的.蒋涛采用了个特殊的上台方式呢~ ...
- js获取并设置<p></p>的显示的值。
原文链接:http://www.nowamagic.net/librarys/posts/jquery/23 html()方法 此方法类似于JavaScript中的innerHTML属性,能够用来读取 ...
- go编程基础
可见性规则: go语言,根据函数名首字母大小写区分private,和pulic. 函数名首字母小写为private 函数名首字母大写为public
- jQuery EasyUI API 中文文档 - 链接按钮(linkbutton)
<html> <head> <script src="jquery-easyui/jquery.min.js"></script> ...
- delphi 自我删除和线程池(1000行代码,需要仔细研究)
unit Unit4; interface uses Windows, Messages, SysUtils, Variants, Classes, Graphics, Controls, Forms ...
- 取CPU序列号,获取网卡,取硬盘系列号,获取目录下的文件,强制删除目录
unit Unit1; interface uses Windows, Messages, SysUtils, Variants, Classes, Graphics, Controls, Forms ...
- easy_install MySQL-python
python - Why can't easy_install find MySQLdb? - Stack Overflow easy_install MySQL-python