GaoZhiShang.com Blog

We have 4 IQ Test facebook applications online, includes English, Chinese, Spanish and Portuguese. Those 4 applications launched before the site, we are getting users from there till now.

In order to cooperate with facebook application, getting users from whom already passed the test on facebook, we make the site as part of facebook application. We provide another “Login” way for users, allow them login via facebook. Don’t need to input email address and password on the site, but login from facebook and redirect to the site with logged status. Because the site is part of the facebook application, we are able to get part of information of the logged user, make it easy to transfer existing data from facebook to the site.

The basic concept is really simple, but implementation is a little difficult. We adjusted the authentication service of the site in order to adopt facebook SSO. And create a similar sign up page flow for facebook user, it looks no different for user who is from facebook or not.

During the development, we found a way to send notification emails to facebook users and can be scheduled to run on server-side. That’s really helpful for getting users.

We figure that a user of GaoZhiShang will mainly come back for the following reasons:

• ·         To find a person for dating
• ·         To find a person for business connection
• ·         To find a person for making friends
• ·         To discuss with others online be it inside discussion forum or with chat
• ·         To organize an event or get information about an event

We can categorize the above as either searching or interacting. We figure that interacting will be much less than as on a site like facebook. This is because facebook is primarily about staying in touch with your friends you got to know offline, GaoZhiShang is primarily about finding new friends online.

For that reason we estimate that the average GZS user will continue in the future to spend far less time on the site than the average facebook user on facebook. So this is a small problem for us but how big the problem is depends on the revenue model we are using, but we have not defiantly decided on the revenue model yet. If our only revenue model should be advertising, then this is not good at all for the financial outlook of GZS.

Before thinking about the time users spend on the site, we need to think about getting people to come back on a regular basis. And even before thinking about this we need to think about getting the registrant to come back at least once to use the site.

We have found that this is not that easy. Many of GZS registrants, have signed up after passing the IQ test  because we congratulated them and it seemed natural for them to sign up (over 90% of those who pass the test sign up!). But a large portion of these registrants never comes back.

This problem is partly inherent in the way that we get people to register on the site and partly made by our lack of incentives to make them come back, which resulted from our lack of focus on brining them back which resulted from an exclusive focus on getting NEW users.

These are ideas we soon want to implement to get users to come back:

• Send users mails on a regular interval especially shortly after registration inviting them to come back
• Modify the group section to get more interactive
  • Make the buttons post message post topic bigger
  • Call groups BBS or forum
  • Make discussion visible to non-members
  • Show the most active group on start page
  • Make every GZS member automatically a member of certain groups
  • Create many groups ourselves
• Modify the search
  • Provide a basic search and an advanced search
  • Divide the search results into dating, friends and business networking search
  • Make the search results so that the interaction is not exclusively focused on certain people and other are neglected
• Create additional functionality: for example, chat, virtual kiss, wall and events

We have many goals:

• ·         Getting returning users
• ·         Finding investors
• ·         Making the site useful
• ·         Getting PR exposure
• ·         Getting high SEO ranking
• ·         Having others be interested in our site

All of these goals are interrelated. At the origin of all of this is the gathering of new users. As you get new users, the site will grow in membership, making it more useful for others, which will lead them to use the site more, which will make investors and others be more interested in it and which will lead to more people writing about us meaning PR and leading to better SOE ranking.

So how do we get new users? : Either by buying them or by getting them for free. In buying them we need to watch for a cheap purchase rate. We are doing very well in that! We still need to analyze carefully how much we are spending for conversions of Google, Baidu, facebook, alimama and our email campaign. But it is clear that we are able to purchase users for little money compared to others sites and that we can expand our marketing without increasing the purchase rate per user.

However, we need to get people to register without paying for them. A good way for that is PR, SEO and word-of-mouth. All of these will go up as the site becomes more popular but all of them are not that easy to influence. But there is one more way that gets people for free to register and that we are able to influence and that is the viral email. Viral emails in form of invitations is send both after the user has used the site for some time and right after the registration. However, the emails send right after the registration are the easiest to be influenced, the most useful and the most likely to be successful. The user can send manual invitations or via a contact importer, but since the average number of emails send via a contact importer is so much bigger than via manual invitations, we can neglect the manual invitation and focus on viral email invitation via contact importer (CI). Thus Viral email invitations via CI is key to our viral marketing success and thus key to getting new users and thus key to achieving all our goals and being successful.

How do improve the viral email invitation via CI? The viral invitation depends on two factors the “acceptance factor” and the “passing factor”. The acceptance factor (AF) is how many registration you get out of 100 invitations send by a users. The passing factor (PF) is how many invitation you get out of one registered user.

Thus:

 AF = registrants/invitations and

PF = invitations/registrant.

If you multiply the two you get the viral email factor (VEF) which is the number of registrants you get per one registrants; meaning how many new registrants does one registrant bring. If that number is more than 1 you know you have a viral super success. For example, one person brings in 2 people these 2 people bring in 4 and these 4 bring in 8, and these 8 bring in 16 etc.

VEF = AF * PF

So we need to optimize AF and PF. How do we do that? We need to break down AF and PF to see each component and in turn optimize each component focusing on the most important ones.

Acceptance Factor

In order for an invitation to lead to a registration the following needs to happen:

• ·         The email needs to be delivered into inbox or junk mail
• ·         The person need to open the email form inbox or junk mail
• ·         The person need to click on linked page
• ·         The person need to click on take test
• ·         The person needs to endure the test until end
• ·         The person needs to score over 113 or take the test again (going through cycle again)
• ·         The person needs to fill in registration details and click register

So we need to attempt to make each factor better:

• ·         Get emails delivered an preferably into inbox by for example getting our IP address white listed
• ·         Change the Subject and Sender of email so to get person to click on it
• ·         Make content of the email easy to read interesting and lead to click on link
• ·         Improve the welcome page design and content with help of web optimizer
• ·         Make the IQ test split in three parts and see if person abandons less

In these things we need to focus on the factors which are lowest since there we are likely to have the highest results. When cannot make test easier. Already we let people register at 115 temporarily and let other take the test again. Already over 90% of those who pass register so that once person clicks on take test there is little we can do to improve the percentages. It is most important that email deliverability and content including subject and sender are tested and improved.

The results we should aim for is 5%. If we can get 5% of all invitation to be converted to registered user than we are on our way to success!

Passing Factor

In order for a registrant to lead to many invitations via CI the following needs to happen:

·         We need to have a contact importer for the email address with which the person registered or the person needs to have another webmail for which we have a contact importer

·         The person needs to input correct username and password

·         The person needs to have  many email addresses stored in his webmail

·         The person need to invite a large percentage of these contacts

In these things we again need to focus on the factors which are lowest since there we are likely to have the highest results.  Already we have contact importers for most Chinese webmails. One that is missing and which is likely important as many users have it and as likely many contacts are stored in it, is QQ. We should get it! We should also focus on getting registrant to enter password and once he entered it get him to invite all contacts. We need to use web optimizer here to find the best design and content.

 The results we should aim for is over 20 invitations per user. If we are able to achieve this and we are able to get 5% of all invitation to be converted to registered user than we are really on our way to success! Because it means that we can stop doing any other advertising and even if SEO and PR and other words of viral marketing do not work at all we will still grow without paying for new users.

Today I met at lunch with Tiago Guerra. Tiago is a part-time investor in start-up companies. He has seems to have many contacts to other entrepreneurs and investors in Beijing. It was a pleasure talking to him. I am thinking about two points we talked about:

1.      there is money in dating

2.      we need to improve our team

There is money in dating

I presented to him the different ideas we had about monetizing GZS and he seemed very open to monetizing the site later with a dating service. He said that there is money to be made with dating and that dating was growing in China. I am thinking about this, because this is the same thing Markus from web2asia mentioned yesterday: monetize with a dating service. I guess I will try to rewrite the business plan in this area.

Improve our team

I know already since a long time that it would be better to have another really smart person join our founding team. However, i feel that getting in a person that is worse than optimal is not good. In order to have an excellent company we need to have an excellent team. And it is better to have two excellent people than three people out of which only two are excellent. The problem is that this person has not been found yet.

It is a bit like a chicken and egg problem: good people want to join the company first after it is successful but in order to be succefull we need to have good people on board.

As I think about it both the team and the monetization question are important but not as important now as the question of how to get more users. So I guess I will go back to concentrate on this question.

I updated the site very often during these days, fixed numerous bugs. In order to keep the site launch, we don’t have so much time to care about the quality. While the code base is growing quite fast, it has some bad smells, you see duplicated code, copy & paste the same code in lots of place, ugly CSS design, bad naming convention, etc. It’s time to clean the code, improve the quality and also performance. I managed to apply methods I learned from book <<Refactoring>> to the current code. Finally, I reduced 10% of the code, extract lots of common methods and template snippets. Looking at the clear code, feel happy.

Finally group section launched. The group section was supposed to be ready more than year ago. Then defiantly by January, then by February then March. We paid many freelancers to get it done but no result. So for the last 10 days ShunJia decided to do the whole group section himself and we launched it yesterday. That meant that many other things could not get done. But group section is just so essential for members to come back. Currently we are attracting many new visitors every day but few people come back.Thinks that are in the pipeline and we classify as urgent are translating the IQ test detailed results fully to English, Installing google “website optimizer” and testing the performance of different pages, getting the email that a friend of mine gave me send out and testing & bug fixing.

I am really tired because today we had so many meetings and phone calls with potential employees and business partners. But it is really difficult to find suitable people. Maybe it is that my thinking is not the same to Chinese. But so often I can decide within a few minutes that the person we are meeting is not suitable.

Yesterday there was a start up meeting at ZhongGuanCun. It was good to be there.

Our strategy to get investment for the further growth of GaoZhiShang is as follows:
Get as many users as possible, get appearance of a real company and improve our company presentation.

The company presentation is already quite good. We have experience in writing business plans and giving power point presentations and we are regularly updating the business plan. We should, however, update the financial section and that is it. The power point presentation is very simple. It does not include images which we would like it to. We had envisioned to get somebody to transform it to look more like the “Meet Henry” presentation mentioned in the blog of Guy Kawasaki but after searching for long for freelancers to do this job we gave up. It was too hard to find these design experts.

The appearance of a real company still requires us to not only register a company in the UK but also in Hong Kong. A Chinese company, I guess will be difficult because we are more than 30% foreign owned. Also we are still looking for office space. Currently we meet either at the home of one team member or at Café.

Today I hope we will get business cards so that tonight when we attend a start-up event at ZhongGuanCun we can give some out to our business contacts.

We will start contacting investors hopefully in two weeks. By then we are aiming to have 1,000 registered users. We are looking for investors at early stage. We have already talked to an Australian angel investor and to China Venture Labs. We have researched several other investors in Beijing and will start soon sending our power point presentation, business plan or executive summary.

The IQ test has now been launched since almost a month. It is still giving us errors if there are many users. This is really annoying but it is something that we accept because if we were to focus on the IQ test more, we would have to take away our focus a bit from the GaoZhiShang which we currently do not want to do.

But what we have done as a very simple measure is reduce the ads spending form $150 per day to $60 per day. This has reduced the traffic a bit and thus made the application more stable. This obviously leaves some free money for other things.

We have still many features to be implemented for the IQ test. The most important one is probably the finishing up of the detailed results. Once we have the detailed results made we have an incentive to convert our facebook users to GaoZhiShang users by offering them to see their detailed results. We also have a way to ask GaoZhiShang users to invite their friends: if you invite your friends and thus help us we will give you the detailed results of your IQ test for free.

Facebook ads

Today we launched facebook social ads for GaoZhiShang. For the last three weeks users did respond very well to our facebook ads, but they only went to our facebook application not to GaoZhiShang. A user of facebook is not that valuable to us. The primary value lies in the fact that we have his data and can send him an email message to invite him to become a member at GaoZhiShang. Thus it was very important for us that users are able to take the IQ test directly on GaoZhiShang and not via facebook.

We hope that the percentage of people who have clicked on the test and actually take the test does not decrease. It probably will decrease a bit as the user clicking on the ad goes outside of facebook. But we have tried to look at the psychology of the user clicking on the ad and tried to make the percentage of people actually taking test stay high. Most importantly, the user does not have to register on GaoZhiShang to become a member. A user from facebook will be led to a spate start page which has one difference: when he clicks on take test he does not go directly to a page where he has to enter his email address but he is directed to the facebook authentication. Within one click he can register on GaoZhiShang via facebook and all the available facebook profile data is imported to GaoZhiShang.

Google ads

36 hours ago we already launched Google ads. It is working quite nicely. There is a constant stream of visitors that come to GaoZhiShang via Google. The average cost per click is only 3 dollar cents. So far 500 people coming from Google have taken the IQ test. Unfortunately only 55 people have passed. That is lower than we expected. While the percentage of the population having an IQ over 120 is only about 12%, among internet users it should be higher. What we have decided to do is to let people who almost pass give access to the site for some time until they have to take the IQ test again. Those that do not pass will get an email from us with a link where, if they want, they are able to give it another try. They will obviously not get the same IQ test, but one that is similar.

What we found out for Google ads is that for many keywords for the Chinese location an advertiser can pay the minimum and still be on the first page. Thus what we have done is to choose the minimum payment but choose thousands of keywords. The reasoning being that even if a person is not looking for a keyword like IQ or IQ test, if he clicks on our ads which clearly states it is a free IQ test, the person is likely to be interested to take an IQ test.

SEO & Baidu ads

Our search engine marketing strategy only relies to a small percentage on Google. More important for us will probably be Baidu and search engine marketing. We have not done any search engine optimization yet but we hope to do so very soon. AS for Baidu ads the only problem we have faced currently is that we need a Chinese company in order to make ads on Baidu, at least this is what I was told. Guo is working to see how it is possible to make Baidu ads.

Ayawawa

Ayawawa is a Chinese internet celebrity who became famous by establishing the saying “there is no girl more beautiful than me that is also more intelligent, and no girl more intelligent that is also more beautiful”. She has tried to establish Mensa in China, and has a forum for intelligent people. We hope to cooperate with her in the near future. We will meet after she comes back to Beijing soon.