ZiTYZEN:81

iPhone application

2009-10-21T00:19:00.003+02:00

Last weekend we released our new free iPhone application, you can find it on iTunes here: iTunes, or read a bit about it here: Apptism.

Initially we made it such that most of the information available on ZiTYZEN also is available on the iPhone version, though users are not able to log in or contribute in any way. If the application is successfull we will expand it, such that users can also uploads pictures, write reviews, etc.

Moving the development to Silicon Valley

2009-08-21T00:31:00.001+02:00

As of the end of August, ZiTYZEN.com will move it's development to Sunnyvale in Silicon Valley.
There are a few reasons to this; We believe that if we are located in the midst of the area were most major international websites originate from and where the most innovative work is done. We will also become more innovative and learn much, ultimately making ZiTYZEN.com an even better website.
Furthermore Pia, one of the founders of ZiTYZEN, has received a shortterm job offer in Sunnyvale lasting until May 2010, hence she will be working part-time on ZiTYZEN until then.

Updates

2009-07-08T09:45:00.004+02:00

It's been a while since the last update, but many things has happened since...

Yesterday we put out our new "home" page. We have felt that we missed a "home" page for a long time, but it's notoriously difficult to make a "home" page, hence it took us a while... We made several iterations and agreed upon the current design. Currently it's only available to Denmark, since we don't have any pictures from Norway yet.

We're still not completely satisfied, but desided that the current design is much better than the previous, hence we put it out. This also implies that a few upgrades will come along in the coming months.

Around a week ago we got the translation to norwegian, hopefully we will get the swedish translation tomorrow :-)
The main reason to why we needed the norwegian translation is that since we operate in Norway it seems appropriate to have ZiTYZEN in the local language, furthermore it's much easier to perform SEO towards google.no if the website is available in norwegian.

Another upgrade we made was to enable history on the searchpage, this is a major problem on ajax pages. Luckily I found a plugin to JQuery, created by Mikage. By using this plugin it was fairly easy to enable history, though it still took a few days to make it work properly.
This upgrade has been suggested by numerous visitors, so a thanks goes out to all of them, for helping us make a better website :-)

Lastly a status update on our expansion to Norway, does it pay?
Since the launch on June the 1st, we've earned 1.5% of our ad earnings in Norway, and so far in July it's 5% of our add earnings we've earned in Norway. We expect (and hope) that this number will rise to around 45%.
Still we have had no affiliate earnings in Norway, affiliate earnings comprise 33% of our total earnings.

The birth of an international company

2009-06-04T09:42:00.002+02:00

ZiTYZEN is now officially an international company!!!
Cool stuff...

We expanded to Norway, beginning with 7.403 places.

I must admit that we came across a few surprises in the process: We thought Denmark and Norway were much more similar, than they actually were. (Still they were very similar)

The biggest surprise is that in Denmark every place has it's own address on a street with a streetnumber, this is not the case in Norway! A given place can have only a zipcode and a name, e.g. Jotums Farm, 8056 Saltstraumen. As one can imagine the yields problems when trying to find lat/long coordinates for Google Maps.

We've set Google Adsense, such that we can track which country each click comes from. This is done using channels, we've created ads for each country and put them the channels "Denmark" & "Norway" respectively.

We still need to to a few things in order for this to be fully finished, e.g. translate ZiTYZEN to Norwegian, and creating Norwegian sitemaps for Google. But now Google can start indexing our Norwegian places :-)

New pages

2009-05-06T08:32:00.003+02:00

Yesterday we added a few new pages to ZiTYZEN; "Links" and "http://www.zityzen.dk/?page=suggest". "Suggest a new place" because we receive numerous e-mails with requests for new places, and we feel it's easier this way. "Links" was created in order to share links with other companies we collaborate with.

In order to get all the new links available on ZiTYZEN we created a sitemap in the bottom of the page, nothing fancy but functional.

A fun usability bug we also found (and corrected) yesterday, was that the language dropdown menu had a link for the language that already was selected. Implying that if your current language was Danish and you clicked on the dropdown menu on the Danish flag, the exact same page would reload with the exact same language. Not a necessary feature!
The bug was discovered by a person that viewed the page while I was accidentially watching. Looking over the shoulder of unsuspecting users is highly valuable :-)

Updates

2009-04-30T07:55:00.002+02:00

Yesterday we released German as a new language for ZiTYZEN, and severel more (e.g. Norwegian, Swedish, French, Spanish) are on the way.

Another extension is that user-generated texts is auto-translated into the language you are using. If for instance you are browsing ZiTYZEN in German all reviews and descriptions are auto-translated in German, but it is stated what language the original text was written in and you can click to see the original text.

For the auto-translations we use a plugin to Google Translate called jquery-translate, which works really nicely.

Furthermore we've cleaned up a bit in the top-bar, making it look less cluttered.

Spring Cleanup

2009-04-08T15:31:00.004+02:00

We've just finished this years spring cleanup, which also marks our 1 year birthday (actually it's been a little more than a year).

We've gone through all the places in the database manually, and checked whether they existed or not, those that didn't exist was removed. A few places were lucky and were listed twice, but now this luck has run out as well :-)
We've also trimmed the total number of subtypes we have, so a few subtypes have been removed.
Lastly we have removed places that lie inside other place, e.g. a snackbars in a fun-parks, restaurants in hotels, etc. The overall rule was that if it's inside another place and it's not listed in the phonebook, it has to go...

Overall this reduced the total number of places on ZiTYZEN.dk significantly from 13.750 to 12.139 or 11,7%.

The are a few major benefits from the cleanup (besides having correct data): We got more than 600 new websites, so we now have 5.016 websites or 41% of the places and the number of phonenumbers rose to 11.752 or 97%.

All-in-all it's been a some though days, but we feel that it was well worth it.

More events

2009-03-13T07:43:00.003+01:00

Yesterday we reached 75 places with events, and it's really hard work to make all those HTML-parsers!
On a good day I can make 10-12 parsers.
The strategy is to find the places that have events and is located in one of the top 30 largest cities in Denmark (http://da.wikipedia.org/wiki/Danmarks_største_byer)

The material we get from their websites is off course copyrighted, implying that it's illigal to collect and display it. Although we believe it's in their own best interest, especially since it's free, and doesn't require anything from them except that they keep their own website up-to-date.
Basically our approach is to contact the places via email, asking for promision to show their events. So far we have contacted 42 places, 21 of which have replied, and only 1 negative answer.

Event Calender

2009-03-04T08:01:00.004+01:00

The last couple of weeks we've been working on getting events on ZiTYZEN. Basically because we believe that more content drives more traffic, and since the content of an event calender changes on a daily basis, this also drives more returning users.

Our initial approach was that we could buy the events from other sources, so we had to do less work. We contacted Kultunaut, the largest event calender in Denmark and event calender supplier to several other websites. But for some reason they didn't reply.

Another possible approach was to make an HTML/XML-parser for each place that had a website and had events, and simply collect the content ourselves. I must admit I was very sceptic of this approach as it potentially could be an extremely large task.

After waiting a few days for Kultunaut to reply, Pia eventually persuaded me to give it a try. So we started off by finding the places which had events and an RSS-feed, as this seemed to be the easiest way to start.

It turned out to be pretty easy and fast to make parsers for the 10(!) places in Denmark which had an RSS-feed.

Since creating an RSS-parser was pretty easy I identified 4 different scenarios for websites with events, and for each of these scenarios I created a template, such that developing the actual parser would be much easier.

Currently I've written more than 30 parsers and it takes roughly 30-60 minutes to create a parser.

The beauty of it is that it runs completely automatical using a cronjob. The cronjob also removes duplicate material and allows places to slightly change the headlines for their events, e.g. Metallica -> Metallica (Few tickets left). Furthermore it detects whether a website is down or changes layout, and then lets me know.

Well, maybe the best part is that we are independent, we can display the events in any way we chose and where we chose (as long as it's on ZiTYZEN.

Creating websites

2009-02-01T10:22:00.001+01:00

It’s has been awfully quiet on the blog for a month now, so long that even I started to wonder, whether or not we were actually doing anything or were just vacationing.

Well we have been doing something; we created ZiTYZEN Websites, which launched today. The grand idea behind this is simply that we have decided to create websites for those places on ZiTYZEN which do not have websites. The only thing the owner of the place has do to is to send us some text describing the place, some pictures, and describe the overall theme he wants for the website, then we’ll do the rest. We host the domain, host the files, and, through Google Apps, provide email accounts.

A ZiTYZEN website is linked to ZiTYZEN.dk, such that pictures, events, texts, reviews, etc. uploaded to the website is automatically shared with ZiTYZEN.dk. The information is initially displayed on ZiTYZEN.dk, and in the weeks to come we’ll connect ZiTYZEN.dk to other websites such that relevant material is shared with relevant websites, e.g. pictures are uploaded to flickr. This implies that the owner of a place only has to keep his own website up-to-date, we’ll take care of the rest of the web.

So why bother doing this? - We can easily extract a list of places without a website, so the marketing becomes slightly easier. Currently we have around 14000 places on ZiTYZEN, with around 4000 websites. Given that we haven’t found the websites for some of the places, we estimate that roughly 7000 places in total have websites. This implies that 50% already have a website, but we believe 90% would like to have a website, if the price and ease-of-use was right. This makes 5500 places, which wants a website.

Another reason for doing this is that ads are generating a lot less than we thought they would be, I guess we were a bit naïve to believe that we could run solely based on ads. Currently we have 4000 unique daily visits – 120.000 monthly visits, which yield 1.200$. This is simply not enough to run a website for a longer period of time, hence we need to start earning some real money.

Happy new year, and we have a new record!

2009-01-02T10:25:00.000+01:00

Happy new year to those reading the blog...

Yesterday we had a new record, 5.077 unique visits to ZiTYZEN.dk, breaking the old record of 4.044, which was set on the 28th of December last year (only 4 days old!).

Improved UI and directions

2008-12-12T19:38:00.001+01:00

Today I improved the overall user interface with two things:

- I removed four tabs from the subsites of the businesses (map, images, contact, and events). The functionality previously found on these tabs is now found on the main page of each business. In order to prevent the page from being too long, while still maintaining frequent upload forms, I decided to hide some things from view. I did this using JQuery (found at jquery.com), a very powerful JavaScript library.

I was very impressed with JQuery, but there are many more libraries out there, lists of some of the major libraries and which websites that uses them can be seen here: speckyboy.com & royal.pingdom.com. As you can see most major websites employ some form of JavaScript library.

- The second improvement was directions in combination with the map. This was actually caused by the fact that I was searching for extension libraries to JQuery, and stumbled upon JMaps. JMaps really wraps all calls to Google Maps in really nicely, and makes it much more manageable. To set things in perspective, the total time spend on the integration of directions into the ZiTYZEN design was about four hours.

Zummary: The use of JavaScript libraries enables modest JavaScript programmers to quickly develop very powerful solutions. I can only strongly recommend the use of these libraries.

Developer API

2008-11-28T17:48:00.003+01:00

Today I finished the developer API i've been working on for a few days. You can see it here: http://www.zityzen.dk/api/ if you are interested :-)

Using the API it's possible to retrieve the images, reviews, upcoming events, and the most commen data associated with any given place, if you have the id.

Furthermore it's possible to retrieve the images, reviews, upcoming events, and the most commen data associated with any places in a given zipcode.

Using this API you can e.g. pull all images out, together with their coordinats, so you can put them on a map.

Now I hope someone will use it :-)

Errors...

2008-11-25T15:52:00.004+01:00

After a few days of heavy refactoring the inevitable happened; ZiTYZEN.dk had an error. The worst thing was that it stayed there for more than 2 full hours without us noticing it... And we're sorry for any inconvenience this might have caused.

The error was caused by a stupid sql-error, caused by a hasty change in the database done earlier today. This also shows that even though we employ version control of the files and test on a test-server before issuing any new releases, things can still go bad.

To prevent this from happening again, i've refactored all calls to the database, so now they go through a central function, which, in case of errors, sends an email to us, and if the error persists, sends a SMS as well. Hopefully this will prevent such errors to occur in the future...

A tip for refactoring on Windows systems is to use a tool such as Windows Grep, which mimics The Grep function found on Unix systems. Otherwise the process of refactoring can be extremely tiresome to perform if you only have windows search at your disposal.

Zummary: We're very pleased with the refactoring which has been done, although not very pleased with the unfortunate breakdown. But the refactoring lead to improvements both in speed and correctness of the search and paved the way for future extensions.

Rewriting the search engine

2008-11-22T07:12:00.000+01:00

One of the comments we got from 2 different sources was that the search engine had a few errors on specific searches were both string and place type was used.

Furthermore we had some discussions whether some of the techniques used in the search so far was beneficial. This concerned the expansion of the geographical region searched and the use of aliasses, and decided to remove them.

This off course called for a rewrite of the search engine, and I decided to also restructure the database, so to better allow for multiple languages (an upcoming expansion for ZiTYZEN.dk).

Due to the restructuring of the database the number of files and tables that needed to be altered was roughly 50% and 10% respectively. I have had the guts to do this without SVN which I installed last month, and it wouldn't have been possible at all to do without a test-site.

Zummary: the result of the restructuring was very good, the search is performed faster and more correct, and the database is prepared for multiple languages. My advise: "Rewrite, restructure, and refactor as often as you can!"

Direct mail

2008-11-21T05:55:00.000+01:00

We have heard from several sources that what one of the things ZiTYZEN.dk lacks is pictures of the places.

Although we have pictures of 1772 places, it still only corresponds to 12.83% of the places, which is too low. So we thought about what to do about this and a friend of ours came up with the idea of direct mail to the places that didn't have pictures.

We decided to start out with the cities of Århus and Odense, the second and third largests cities in Denmark. This totalled 483 places without pictures.

It should be mentioned that the workload of creating these letters, printing them, folding them into envelopes, and stamping them, before shipping them, was well over what we had expected.

So far 33 places have logged into ZiTYZEN.dk, and 7 have uploaded pictures. 14 letters have been returned with address unknown. Furthermore we have received numerous phonecalls, mostly positive about ZiTYZEN.dk.

Zummary: Since only roughly 7% of the places that received a letter responded to it by logging in, this was an extremely expensive way of advertising. Although we got some connections and a lot of comments on the website.

Conference in social medias

2008-11-20T06:40:00.000+01:00

Today I attended a conference in social media, På Herrens Mark (translated it means something like 'lost'). It was an event held primarily for the businesses in the area that have or develops technology products.
There were 5 speakers and lots of free food and cake but the quality of the speakers were very fluctuating. Some very good, whilst others were directly frightening.

Though some of the speakers were good, like new were said. The conference, i guess, was targeted marketing employees not yet familiar with web 2.0. Things like twitter (still not a big thing in Denmark), facebook, etc. was discussed.

This frightening quote was said by a guy from Aab (a football club, which had recently renewed their marketing strategy): "Don't trust the users, always make sure you have complete control of all material on your website"... Pretty dump words coming from a man in charge of running the marketing campaigns of a football club which relies 100% of it's fans (users)!

Zummary: I think the primary reason to come to events like this is to network, and to see what the other not so web 2.0 companies of the region is doing.

The value of newspaper articles

2008-11-10T19:16:00.000+01:00

In the last few weeks, we have been trying to get more users to come directly to our website, we have tried to get some local newspapers to print articles about ZiTYZEN.dk.

We hope to increase the portion of users who goes directly to ZiTYZEN.dk, currently we have around 3000 daily users, where around 30 (1%), comes via direct traffic, around 50 (1.75%) comes from referring pages, and the remaining 97.5% comes from Google and other search engines.

We had some help to write a story from a communication student, as we are primarily engineers, and not experienced in writing press releases for the newspapers. She wrote a nice article which she sent to a few newspapers. The process of writing the press release took roughly 14 days, so it was a fair amount of work.

The first paper which published a story about us, was a small local free newspaper called 24timer. They gave us a small article on page 18, which was largely rewritten from the press release, being on page 18 was not satisfactory. In the article they described ZiTYZEN.dk as a great new site with very few flaws, so that was satisfactory.
The actual payoff of the article can be partially read in the direct traffic which also climbed on the day of the article (tuesday the 28th of October), and the following day. The total amount of additional users were 30-40, which was a little disappointing, furthermore we believe that being in the newspaper also increases the overall click-through-rate on search engines, as people are more likely to click on URL's with familiar names.

The second paper we know the article were published in was Midt-Vest avisen, another free local newspaper. The two papers are also part of the same overall company. Hence the article was exactly the same as the one printed in 24timer. The second article was printed on wednesday the 5th of November, 8 days later on page 24. It's a little hard to say something about the impact of this article as the spike we encountered in the direct traffic statistics occurred on tuesday, the day before. We don't know the exact reason for this, but suspect that the paper might have been out already that day. If this was the case this article gave around 50 more direct users.

We also sent press releases to two other Danish newspapers: iByen.dk and Midtjyllands Avis but we don't know with a 100% certainty whether or not the article was printed, although we believe that it was printed in Midtjyllands Avis sometime after the 5th of November, although this doesn't show in the statistics.

Zummary: we spend around 10-15 hours on the press releases, with a very little gain to show for it, only around 100 users spread over 10 days, yielding 10 users per day. Nothing that shows when you already have 3000 users per day. But despite the poor result we still believe that it was worth the effort, as we got some experience with press releases, as well as we believe that it partially comes down to luck.

This time we targeted 3 media bureaus, next time I think we will try to hit even more using some press release agencies, e.g. get2press.dk or pressport.dk. In order to save time and hit more papers. Unless we have a story which is only relevant in a small geographical region, in which case I doubt it is worth spending much time on writing an entirely new article, but since we have an article now we could simply rewrite it slightly, and send it to the local paper.

Subversion comes to ZiTYZEN.dk

2008-11-07T16:54:00.000+01:00

Today I completed the setup of the version control system that I initiated more than a month ago. It's not that it's hard to do, it's just that I needed some time to pull myself together and get the job done. It's one of those tasks that you don't really see the benefit of today, but the day catastrophe strikes you hate yourself for an entire month. So it's basically an insurance (and a really cheap one, because it's really easy to setup and maintain).

I chose to use SVN (subversion), mainly because we've used CVS in university and was a little tired of the missing functionality of not being able to delete directories and not being able to rename files.

The installation is really basic, you type "apt-get subversion", and you're more or less done. The you setup a repository first by creating the directory "mkdir /home/zityzen/repos/", then you create the repository "svnadmin create /home/zityzen/repos/". Now you want to import the files you already have into the repository, which is done using this "svn import /home/zityzen/repos/ file:///usr/local/svn/repos/some/project \ -m "Initial import"", were some/project is the location of the existing files.

Then there is the trouble of how to structure the layout of the repository, I've chosen a fairly standard way, in which there is a trunk (corresponding to the test-server) and a number of releases (the latest release corresponds to the running version of the production-server). This implies that whenever something is committed it's done to the trunk and whenever a new release is required the export function of SVN is used to create a new release. Note that the old releases are not deleted, they are needed in order to perform roll-back operations. The export command looks like this: "svn export /home/zityzen/checkouts/trunk /home/zityzen/checkouts/release/2008-11-07--1", this denoted that this is the first release on the 7th of November 2008. I've also create a little bash-script that figures out the name and does all the copying for me, nice and easy... I also used another of the features of SVN, hooks. These are scripts run after e.g. commit, the hook can then be used to copy the content of the trunk over to the running version of the test-server, such that it's kept up-to-date whenever something is committed.

Furthermore I've created the real insurance namely the roll-back script, which deletes the latest release and replaces the running production-code with that of the second-latest release, again: nice and easy.

Now you might argue: you don't need SVN to perform that, you can just run those bash-scripts, they will still perform the same operations and be equally insuring and useful. Partially true, I don't really need SVN, but it's nice to have this added insurance on top of the multiple releases, furthermore using SVN forces me to use those scripts, which I otherwise could forget. Additionally having SVN is a benefit when you work in teams, and although we're not a large team yet, SVN makes the job of sharing files much easier.

Stupid users...

2008-11-02T08:20:00.000+01:00

We had a problem with the number of reviews we generated, as the number was too low; a few reviews a week even though we could see that users were writing reviews, they just wouldn't sign up, so the reviews could be seen. Hence we decided to make it even easier for the users to write reviews and sign up.

The solution was that when a user whom is not signed in writes a review, he/she must also write their name and email. Whenever the review is submitted an email is sent to the email address explaining about their newly acquired membership of ZiTYZEN.dk along with a generated password.

This also implies that it will be hard to lower the barriers for writing a review even more, and next to impossible if we still want to maintain at least some control over who writes what. This is very important due to the business owners who doesn't trust internet users at all, hence they need some reassurance that whoever wrote this is a real person.

One of the more annoying problems this gave us was that people wrote all sorts of crap in reviews. So we decided to take 2 actions: the length of reviews must be more than 50 characters, this takes care of the most crap, as well as the quarantine-system we have also got an upgrade. So now reviews written by users and ip's that are in quarantine, also is put directly in quarantine. Reviews can off course be easily un-quarantined but we need to read the review first before we do so.

Zummary: On a web 2.0 website users are your greatest asset but also your greatest nemesis. This implies that if you want to be a respected website you need, make it as easy as possible to create content but also be sure to have some kind of control mechanism to manage the content. This could either be a centralised approach, such as ours, or when the website becomes more popular and the centralised approach becomes unfeasible, a decentralised approach were the users themselves govern the content.

I personally believe that the reason for all the stupid users around is the history of the web as a thing on which one was believed to be anonymous. This has off course never been the case, and is still not the case today. But some users still believe they can browse the web without leaving a trace. This far from the case as most websites have Google Analytics, 1-2 different sources of ads, log-file statistics, and maybe even internal statistics, implying that any single page view is tracked by up to 4 different companies. On top of that comes all the add-ons installed in the browser Google Toolbar, Alexa Toolbar, etc. which also track numerous things.

Affiliate marketing

2008-10-29T09:04:00.000+01:00

As most other websites we what to make money from our users, well not directly from them but through ads. It's not that we what a ton of money, we just need some to keep fund the marketing campaigns and hopefully I'll get paid some day.

Well anyways, the fact of the matter is that we've had Google Ads running ever since the publication of ZiTYZEN.dk and we wanted to see whether or not other affiliate marketing sources could perform equally well. Another reason for trying something other than Google Ads was also that they just look plain dull.

So we found (and got recommended) a few companies, and I simply wrote them. Four of these we tried, adBrite, adPepper, Guava, and MikkelsenMedia.

The first one the respond was Guava who responded the same day, and after a short conversation on the phone we were affiliates, so we set their ads on ZiTYZEN.dk. They looked very fine, but there was a slight problem you had to determine which ads you wanted to display, there was not automated rotation. Well that's part fine part annoying, why don't they do it, i guess they are more qualified for that task than I am. Well, the ads performed like crap when it came to make money, but they looked good...

adBrite was also quick to respond, as it's basically just an online form you need to fill out. They work much like Google Ads and I put in the javascript-code, but nothing happened. They showed an ad saying something like: "We're still collecting data, the ads will come when we know which ads to display" (don't take it literally, but the point is that they needed some time to adjust). I wasn't to happy about this, but fair enough. After roughly 1500 displays I pulled the ad, it still hadn't showed any proper ads just proclaimed that it needed time to adjust.

MikkelsenMedia was next in line. When I put their ads up, I initially tried them in various browsers. Even the text ads displayed differently in different browsers!
So I took some screenshots, pulled them immediately, and sent the screenshots to MikkelsenMedia so they explain whether or not it was just me. Furthermore ad displays are registered automatically in the database, so I was a little puzzled by the fact that in their system the number of ad displays was still 0, even though we could see that it had been shown well over 300 times.

adPepper worked much better and is the system we use currently. The money generated by the ads is much lower than that of Google Ads but still fair. Furthermore the ads look professional and we're just generally pleased with the result.

Zummary: After having tried a few other sources of affiliate marketing sources, I'm not surprised Google has such a good business, most other companies suck! Off course it's a hard task, matching a website with appropriate ads, but still... If it's the only thing you do...

Flame put aside, I'm pleased with the result, ZiTYZEN.dk looks more professional. Although it came at a cost, but as long as Google Ads are displayed on each page view alongside other ads, the cost is minute.

The JAOO conference

2008-10-02T10:48:00.000+02:00

This week I attended the developer conference JAOO in Århus. I attended it a few years ago as well, and I must say that it's really worth it. It really was the superstars of computer science that was there...

Anders Hejlsberg talked about the next generation of programming languages and started the conference (and the hype of functional programming which was very outspoken this year).

Then I saw Glenn Vanderburg and Ben Goodger, talking about how to create a functional UI and about the development of the UI for Google Chrome, respectively. The afternoon of monday was spend listening to presentations of cloud computing, presented by Gregor Hohpe, René Schmidt, John Davies, Dave Thomas, and Stefan Tilkov.

Monday is also the night of the conference party, and Erik Meijer of Microsoft had the party keynote, which was outstanding. Although i'm not a big fan of functional programming he sure understands to make a presentation.

Tuesday started quite exciting as Lars Bak presented the newly developed V8 javascript engine, one of the corner stones of Google Chrome.

Wednesday I saw a few not so interesting presentations and went home a little early.

Zummary: I'm a big fan of JAOO, and it's close to awe inspiring to attend these kinds of conferences. It's like refueling pure energy. I can only recommend attending conferences like this!

I went home with a notepad full of new ideas. Now I have something to implement in the next few months :-)

Contests and online marketing

2008-08-25T05:54:00.000+02:00

We have been working furiously for over one year now on ZiTYZEN.dk, and so far we have primarily been focusing on online marketing and SEO.

As a means for online marketing we are currently only using Google Ads but we have also tried Yahoo Search Marketing, and creating an easy contest for an iPod.
Yahoo Search Marketing did not work particularly well but this is most likely caused by the fact that Google is responsible for over 90% of the searches performed in Denmark. Furthermore the budget we gave Yahoo was very small, but compared to Google Ads we got less displays and less clicks for each Euro.

The iPod contest, was very simple: You win an http://www.apple.com/ipodtouch/ if you can answer a simple question to what service ZiTYZEN.dk provides. Is it a restaurant, is it a realestate agent, or is it a nationwide source of inspiration for consumers?
We then put the contest out on the 10 largest websites targeted such contests in Denmark, and we soon started to receive a few clicks but not that many. We also put the contest in a newsletter send out to 100K+ receivers. It turned out that only roughly 600 of those actually opened the newsletter and we received around 400 clicks. So in total the contest campaign costed roughly 4 DKr pr. click while Google Ads costs roughly 3 Dkr. pr. click, with the added bonus of providing users with a higher number of pageviews, as they don't come for a prize but because of information. This implies that if we want to do this again, we should either buy a smaller prize, or appear in more newsletters with the same contest in order for the prize pr. click to be smaller.

Of the three sources of traffic (search engines, referring websites, and direct), we have only been focussing on search engines, and feel that we now have to put more focus on either referring websites or direct traffic. One way to put more focus to referring websites could be to make a contest with a prize, as we have already done, and it actually raised the traffic coming from referring websites to 5,7% for a brief number of days, and to a staggering 9,3% for the best day. But this strategy is unfortunately to expensive for a company such as ours. Another cheaper way could be to provide all the businesses on ZiTYZEN.dk with some additional free services if the would link to us from their website. The downside of this is that they have to do some work and it seems that most are willing to promise they will do it, but reluctant to actually do it.

This all leads to the fact that we wish to increase the direct traffic from the current 0,5%. We envision that this could be done by a more comprehensive marketing campaign, which off course should be relatively cheap as money is scarse.

Mobile website part II

2008-08-14T09:49:00.000+02:00

Along with the new work we done in the are of sms's we also wanted to make the mobile website, ZiTYZENMobile better. We did this by asking people who used it, and looked at it ourselves (with relatively new eyes as we don't use it ourselves).

This lead to a major decrease in functionality as we removed several pages, and tried to stick with the most used functionality, i.a. search and places.

We simplified the way the search is performed and made the result easier to overview. Then we removed most functionality regarding members, implying that members cannot login but they can write reviews, by simply providing their name and email.

We also removed redundant code segments that the mobile version shared with the PC-version, e.g. the search algorithm, and the script for uploading reviews.

We hope the new and simplified version will be more useful to everybody :-)

Zummary: It's nice to get rid of redundant code, especially when it's as important as the search algorithm. Furthermore I think the simpler mobile version seems to work much better and I think even I will start using it now, so remember: "whenever you go mobile, go simple" :-)

SEO (Search Engine Optimization)

2008-08-03T17:42:00.000+02:00

One thing I've been working very much on is SEO or Search Engine Optimization. I have spend countless hours reading books and browsing websites on the subject, and it has been paid out fully.

One book I can recommend is "Professional Search Engine Optimization with PHP", it's not that is groundbreaking in any way, it just sums up all the different things you have to keep in mind very nicely. I must admit that I skimmed most of the book as I had already performed that particular optimisation. The information can be found on various websites, e.g. www.websitegrader.com, www.seomoz.org, www.vaughns-1-pagers.com/internet/google-ranking-factors, but there are many others, these are just some of the most prominent.

The search engine optimisations I've performed ranges from minute optimisation such as prolonging the subscription time for the domain from 1 to 5 years to more complex things such as creating a sitemap and making URL-rewriting.

The hardest thing to perform was URL-redirection I did in in the .htaccess-file, implying that it really dictates that you write everything correctly otherwise your whole site breaks!
It took me a few days to write those few lines that eventually got the URL "http://www.zityzen.dk/moonlight_torvet_1_odder" to be rewritten into "http://www.zityzen.dk/?page=place&place=7294" without the user ever knowing it...
This blog from Corz really helped alot!
URL-rewriting is supposed to help in search engines, because the search term might be in the URL, which might be true, though I doubt that it has a great effect. But I'm pretty sure it increases the click-through-rate, whenever a result is displayed in the search engine. So definately worth the effort, and it also looks way cooler than all those id's :-)

A really easy and basic thing was the sitemap,
which basically is an XML-file with all the URL's of your website in it. You can then rate the importance of each URL as well as how often it is updated. This should allow search engines to crawl your site faster as they know were to go. You just need to setup a cronjob that run a script that generates the XML-file. We've set it up such that it runs each night, and I can see in Googles webmaster central that they pickup the file shortly after. I can only recommend doing this, as it is extremely easy to implement.

Furthermore (as I mentioned before) Googles webmaster central is also very helpful, and provides valuable information.

Currently we been "live" for a few months and have 2K+ daily visitors, 97% of whom comes from search engines and 1,5% comes from other websites, leaving a minute 0,5% to direct traffic.

Creating an SMS-service

2008-07-02T22:29:00.000+02:00

One of the major problems of most websites is how to generate enough money, so we have thought a bit about how we could do it and came up with the idea of providing the places with a sms-service. Functioning much the same way as an email-newsletter. The businesses collect the phone numbers of their customers and write sms's to them whenever they have events coming up, a new menu, or something else of interest.

The idea is basically to make it as cheap and easy for the businesses as possible, so we also want to produce marketing material for them.

There are numerous sms-gateways, so the first issue was to find the best and cheapest one, ours is called hardwired.dk, and is a provider based in Denmark. Unfortunately they only use 1 service provider, so to ensure that we can always send sms's we have another gateway as well. The switch between the 2 is made automatically if the primary cannot send sms's.

Another issue was that we wanted to take credit cards. In Denmark this is a 2-step process as one first needs a deal with the company responsible for most common credit cards in Denmark, PBS, then an agreement with a credit card gateway, FRIbetaling.dk. The integration between all the credit card systems and an easy UI on ZiTYZEN.dk proved the hardest issue to resolve.

Zummary: Until now it has been hard work, lets hope it pays off in the long run...

Finished with university!

2008-06-24T12:02:00.000+02:00

A few days ago I had my final exam at university, hence I'm now a free man :-)

It's a really strange feeling, but nonetheless a great feeling!

The completion of the search engine project, didn't go exactly as planned, as the test subjects who tried the search engine found it very slow, due to all the additional algorithms that are calculated. This implied that I had to remove most of them again and we're now left with a nearly standard search engine. It still uses query rewriting, while the ranking of search results is done solely based on the content-based agent, combined with an agent who scores each result according to the actual search query. This is much faster than using both the collaborative-based agent and the preference-based agent. Especially the collaborative-based agent turned out to be exceptionally slow.

In the end I found that users will rather have great speed, than great functionality. At least when the functionality isn't 100% correct. The reason for this was most likely the small dataset, as we only had a small number of users and that they often only visit our site a few times, hence it becomes a nearly impossible task to correctly identify their preferences based on 2-5 page views.

Zummary: Build your applications for speed, not for some unique functionality only computer scientists really can appreciate...

Expanding the geographical area in searches

2008-04-05T22:25:00.000+02:00

I decided to use query rewriting in the searches such that it will extend the query result. The idea behind this is that in Aalborg there is a very nearby city called Nørresundby, and e.g. turists are not aware of the fact that these cities are in fact two very different cities and not one large.

Hence it seems appropriate to in some way make the places from Nørresundby appear in the search result from Aalborg.

If we extend this thought slightly, we end up with the idea that the places that lie within a given range from the area in question, are also returned as part of the search result. This also enlarges the result set by potentially inserting garbage in the result set, we made a study that claimed that roughly 80% of the original result was garbage, while 97% of the added result were garbage.

This implies that they should be penalized such that they end up somewhere in the back, and those contained in the original result set are listed first.

Follow-up on "going live"

2008-03-29T00:18:00.000+01:00

ZiTYZEN.dk went live on friday the 14th of March and have ever since been correcting the flaws, bugs, etc. our friends found.

The crawling process of the search engines have been much more slowly than I had expected. I made a major error when I set up the lock for ZiTYZEN.dk, as I simply put up a log in screen for ip's not logged in. This proved very bad!
This strategy had two effects:
- Crawlers thought our site still had only one page and are very hesident about returning again
- Our Google Ads still focuses mostly on the keyword "login", which is very bad...

We still have great difficulties getting listed in the search engines, but after 16 days of very low number of visits, ranging between 13-32... But yesterday we had 62 visitors, so hopefully it's slowly starting to change

Zummary: NEVER set up a login splash screen!

Starting a Startup

2008-03-25T15:33:00.000+01:00

Let it be said, there is a long way from the initial idea to the startup is a reality (and most likely an even longer way before it becomes a success.

When we first got the idea of ZiTYZEN.dk, we though that it would be pretty easy to take the idea and transform that into a finished product, ready for the world to use.

Not quite so...

Initially I started the company on my own, the first name was "ZityZen v/Tobias Christensen", implying that it was a company owned personally by me. We enlisted in a program the University of Aalborg called SEA (Supporting Entrepreneurship at Aalborg University). It's basically a program in which the students can participate for free and receive some help in the startup phase. We participated in a course over 10 lectures, in which we were taken through some of the potential pitfalls of creating a company, tax-rules, hiring employees, type of company, marketing, etc. We also received two mentors, persons who had tried to make a company before, so they knew the process and were much more experienced than us. Furthermore we received 40 hours worth of free time with layers, accountants, marketing people, etc. And lastly there is a person associated with SEA who writes press releases for free.

So this was all good, and the course paid of really well, the 40 hours were cut to 20, but they were still free, and although we could have used many more, we feel that they were well spend.

As entrepreneurship is currently a buzz-word in Denmark, everybody wants to help small businesses get airborne. Hence the local region (similar to a county) also has a nearly similar service called NiN (Nordjysk iværksætter Netværk). Although they don't have courses as such but offer some events occasionally. We also received 24 hours from them, for the tiny price of 1000 dkr, which corresponds to the price for one hour at an accountant.

Zummary: We had great help getting airborne, and we're definitely not there yet. But it was helpful to have some experienced people look over the business plan and get help from experienced people.
So if I should pass on some advice it would be to try to find someone who has tried it before and convince them to help you get started (simply be regularly asking them for their views on your business plan and ideas), they will most likely love the challenge and probably do it for free...

Going live!

2008-03-15T01:12:00.000+01:00

Today I removed the "locks" and ZiTYZEN.dk went live. We sent mails to all our friends and hopefully we'll get some positive feedback :-)

In the first few days we hope to get feedback from our friends, and get ZiTYZEN.dk registrated on as many search engines as possible so the crawling can begin...

We expect that the process of being listed on search engines might take a few days or maybe a week, but we'll see about that...

More university projects

2008-03-03T16:26:00.000+01:00

This semester is my last at the university, so again I will spend it on ZiTYZEN.dk. This semester there are again two potential projects: providing each place with it's own webshop or providing a better search algorithm, one which takes the users profile into account, so the website can perform more qualified searches for the user. I chose to play around with the underlying search engine of ZiTYZEN.dk.

The main object of the project is to enable users to log in to ZiTYZEN.dk, locate themselves on the interactive map, and then the proper results (e.g. restaurants) will pop up. This shall be accomplished solely by looking at the history of the user, hence by silently observing the user, the behaviour of the user must be learned. I guess this is the wet dream of most search engine engineers :-)

Well I have started by first establishing all the statistics that should be collected about the behaviour of the users on ZiTYZEN.dk. To briefly mention all the stats that are collected:
- time spend on each page
- whether scroll was used
- page displayed (e.g. the place displayed)
- all properties of each search
- all the standard stuff (geo-location, user agent, time, etc.)

The most interesting statistics are the time spend and whether scroll is used. These two stats (together with the number of displays of the given page) are used to determine how interesting a given page (or place) is for the user. This implies that the more active a user is, the more interesting the user is likely to find the place. In this way it should (hopefully) be possible to automatically determine the users profile without the user having to do anything. In other words; the algorithm provides an implicit rating for a place by a given user.

Basically the search engine should be combined of three different search agents;
- the content-based agent (determines the value of a search result based solely on the amount of information of a given place, e.g. number of pictures, whether a description is given, etc.)
- the preference-based agent (determines the value of a search result based on the implicit ratings of the places "given" by the user making the search)
- collaborative-based agent (determines the value of the search result by first matching the user making the search with other users in the DB, the places that the most similar users like are then assumed to be well suited for the user)

These three types of search agents are then combined using ensemble methods, which hopefully will create a great result :-)

Well, more to come...

Now with fancy graphics...

2008-02-21T08:28:00.000+01:00

One of the things we received from SEA was around 10 hours with a marketing consultant. One of the things she initially said was that we needed a powerful logo and a more professional expression on the website. Luckily she knew a graphics designer named Randorff, who were experienced doing this.

He was extremely skilled and helped us very much, the development of the site from the initial website, over the first sketch, to the second more detailed sketch, to the "finished" website.

We paid 15,000 for the graphics, which was quiet cheap. The result is much much better than the initial design. One of the great things about having a more professional design is that people automatically take your company more professionally. I can strongly recommend to spend some money on the design and logo.

Zummary: Always spend some money and time to get the design and logo right, it helps enormously whenever you need to show it to someone, i.e. you are taken more seriously. This should be one of the things you do before you "go-live" with your website.
Furthermore you need to find a professional designer, not just someone in your family who designed the family-website two years ago.

Server setup

2008-02-04T16:03:00.000+01:00

One of the goals we put up, when we started ZiTYZEN.dk up, was that it should be a well-driven and professional company (website). This also implies that we had to think a somewhat professionally from the start, namely by investing in our own server.

Since I don't have much experience in either setting up a server or running it, we found that I was best to let others take care of that. Luckily one of my friends from university is much more experienced than me in this field so we enlisted him to set up the server. The server is located in a Danish hosting center called EasySpeedy (they use four providers and are well connected to the European backbone), there are others who deliver a similar service, but my friend who has tried others claims that they are the best. This solution seemed best as a friend of mine has a website in the top20 largest websites in Denmark, and they have on two occasions tried to run their own servers, but they failed miserably both times.

The service EasySpeedy provides allows us to have complete control of the server. although it's located in another part of the country. It's very simple even for an unexperienced server admin, as myself, one click and you have your operating system installed as well as the LAMP-package.

Once the OS is installed you're on your own, luckily we had our guy to step in and help with all the important stuff, on which he spent roughly 20 hours.

The most important things to get installed is:
- security (a decent firewall script, which doesn't allow cross-site scripting)
- log-file web statistics (awstats)
- a mail-client
- backup (we take backup each night to the local drive, and once a week the backup is transferred to an external location as well)
- proper settings for the web-server (allowing browsers to cache images)
- and the most important feature: two web-sites, a production-site and a test-site

The test-site allows for testing before the code is put on the production-site, which is very very important. A single error in the code could otherwise break the site. I have some inside knowledge of one of the top20 sites in Denmark which doesn't have a test-site, they simply uploads the code directly to the running website. Which occasionally breaks the site.

The morale of this is: If you believe that your website will ever be a success you will at some point need a web-server of your own, why not get it from the beginning, it's not that expensive (remember to let others run the server!). The only thing you need is an experienced server-admin who can set it up for you.
And the most important thing, always take backup and always test your code in some environment before you upload it to the running live version of the website!

Starting up using my projects at the university

2008-02-02T08:37:00.000+01:00

We started brainstorming ideas for ZiTYZEN.dk in the summer of 2007, and the programming started shortly after. It was built on the framework of 9000bodega.dk but it needed much rework and upgrade.

I started my last year on uni the 3rd of September 2007, and I knew we had to find our own projects and our own supervisors, and we had to to it all in the summer. Hence I created a project from the brainstorm we had done for ZiTYZEN.dk. There were initially two ideas which could have become a project, we wanted a mobile-version of ZiTYZEN and a webshop for all the places on ZiTYZEN.dk included for free. Coming of a very hard year writing my master thesis I decided to go the "easy" way by creating a mobile-version of ZiTYZEN.dk, ZiTYZENMobile.

The idea for ZiTYZENMobile was to look and feel like the PC-version, and with much the same functionality. The project was done as a usability project, in which I had three usability tests.

The idea of a usability test is actually pretty straight forward; you make a series of tasks that you expect users would be likely do perform on your site, you find some likely users (not only computer science students!), you record the users while they (struggle to) perform your tasks, and record what they think and do. Based on the finding you can create a prioritized list of usability issues, which you can address in a minor re-design of the website, such that it becomes even more usable for your likely users.

This rather simple technique actually worked very well, and it was easy to use, though you should note that it was very time consuming.

After trying this technique in practice I find many websites out on the web today pathetic. You still find the line "best viewed in Internet Explorer"... That's just the work of a webdeveloper who is too lazy and should be fired immediately!
You also find forms that are nearly impossible to fill out, multiselect lists where you have to push CRTL in order for it to select multiple items, which is clear to those who have studied computer science but noone else, menus that are impossible to navigate, and many more examples of webdesign designed by a lazy (or inadequate) webdeveloper...

Another way to achieve user feedback is to simply set up a feedback form, so you can get feedback by mail. You can also go for the slightly more fancy solution of putting up a kind of feedback-forum, in which users can see the proposals of other users, and maybe even rate the validity of their proposals. A very small number of users a likely to use this method of providing feedback but if you have enough users it's not a problem. Another drawback is that it's only a certain kind of user who will give feedback.

The morale of this is that your need user feedback, and that you can do it in two ways, allow the user to give feedback while online or pull some users into a lab where you ask for their feedback. For a small website with a limited number of users it is vital to perform usability studies in the lab as they don't have enough users on their websites, but as a website grows in makes more and more sense to also have an online feedback possibility. But remember to perform usability tests, as you get that very rare opportunity to push a button called RUM (Read Users Mind).

Introduction

2008-01-31T22:55:00.000+01:00

ZiTYZEN.dk is an idea we (Pia & I) came up with some time ago, because whenever we wanted to go out in Aalborg, we initially wanted to explore a restaurant or cafe we had never gone before. But due to lack of options ended up in the same old 3-4 places because it was very hard to find information about what other places there were.

We had also previously created a website for the bars in Aalborg, 9000bodega.dk. On this website all the bars got a free profile page, onto which they could upload pictures and edit pictures and reviews. Furthermore all users could rate and write reviews of the bars. The bars liked this idea and some even got more customers, so it was free online marketing that actually worked. Hence the idea was basically that we wanted something similar, but much bigger and much more professional.

So since we had an idea and some experience with a similar idea, we got right to it and started creating a new website from scratch. This time we started directly from a dedicated server, and designed the system such that it largely could be run without any human "interference".

It was a major dream of mine to start my own welldriven company, which was profitable enough for me to live of it, and I felt that the time was right now, as I finished studying computer science this summer. Furthermore I can use this last semester as a startup phase for ZiTYZEN.dk, for two reasons; first I wrote my master thesis on the previous year, hence this year is "free-time", and secondly because I get some money (roughly 1000$ per month) from the state (SU) until I graduate and when I do I still haven't gotten used to earning 7000$ a month.

Hence seen from an economic point of view as well as an logical point of view, now's the time for greatness/the best time to try out my baby-legs as a CEO :-)