My pleasure reelnorth, here you go.
Translating Intertainment Media's Traffic Trends
May 5, 2011 | 15 comments | about: ITMTF.PK
Intertainment Media (ITMTF.PK) has been one of the most active stocks on the Venture exchange since the end of January, rising from $0.11 on January 25th to a high of $3.35 on April 19th. While the stock has sold off from recent highs, it hasn’t come all the way back to earth yet. The company has smartly decided to raise money while the stock is at this level, and press released on Tuesday morning that it is going to raise up to $48,000,000 at $1.20/share, plus a half warrant exercisable at $2.00 per full warrant. We are short the stock and I will try to explain why.
I do not have all the information I need to make a definitive statement and analysis. Everything in here is pieced together from publicly available information that the company has provided. I requested some information from the company that would have made certain things clear, but I did not hear back from them. If there are any errors or omissions in my analysis I am more than happy to hear about them and fix any mistakes in this commentary.
Intertainment Media has a few product lines, but most of the focus has been on its Ortsbo product, which is licenced from SaaS Technologies in some type of revenue sharing arrangement that is largely irrelevant. For the sake of simplicity, let’s just say Intertainment owns Ortsbo and is the primary beneficiary of its success (which I understand to be true). I have never used the product, but my understanding is that Ortsbo is a real time translator used for online chatting, also known as instant messaging. Watching some of the demos on the site give an idea of how the product works: You put in your MSN (or Facebook, or the chat program of your choice) username and password into the app at ortsbo.com and choose a language. Simple as that, you’re online and having your text translated into a different language in real time, and, I’m assuming, having other languages translated back into your native tongue.
As far as I’m aware, the company has yet to publish financial results where Ortsbo is a significant driver of revenue, although that is also largely irrelevant. This is an Internet (or if you prefer: social media) play and revenues aren’t relevant, users are. Intertainment Media understands this and has been posting weekly user data for Ortsbo, data which it says is collected by third parties such as Google Analytics (GOOG). This data can be found on the Intertainment Media web page under “Updates”. As an Internet play, it isn’t surprising that as the company approached 10,000,000 “unique users” (the company's term, see the link) the volume and share price started to increase rapidly.
It would seem that this is pretty straightforward. As users increase, so does the value of the company. The same logic universally applies to other Internet sites (like Facebook, for example), so why would the Fund short the stock? Simply because Intertainment Media is reporting absolute unique visitors as users. This means its user base IS NOT comparable to Facebook’s or Twitter’s user base. Unique visitors (and Alexa Rankings) can be artificially increased, which is why registered users is a better measure- although these can be faked too. Given that the Ortsbo app requires a login to use, it is puzzling why the company wouldn’t report unique successful login IDs as opposed to unique visitors.
It has been proven time and again that USERS are very valuable to someone, even if revenue from those users might be hard to come by. The key, in my opinion, is to differentiate between users and visitors. The largest public and private companies have a simple definition of users; people who have registered for their site. Facebook currently reports having 500 million active registered users who, in a given month spend 700 billion minutes using the site. Public companies like Netflix (NFLX) and Activision (ATVI) report metrics based on paying users with 23 million people using Netflix’s streaming movie service (some on free trial) and 11.5 million people paying a monthly fee for Activision’s World of Warcraft. Here is a list of Internet sites, the number of users they have, and their Alexa ranking. Other than Intertainment Media, I have never seen a public company use absolute unique visitors and users as an interchangeable term.
Using the terms interchangeably (although incorrectly) allowed the company to release these figures, which implies that it already has more users in 10 months than Facebook gained in its first three years. This press release is misleading as the company is comparing unique visitors to registered users. It is also wrong in that according to the site that was referenced, Facebook was launched in February 2004. So 36 months from February 2004 is February 2007. There is no data point listed for February 2007, although December 2006 is 12,000,000 users and April 2007 is 20,000,000 users. So a 36 month estimate of 15,000,000 might have been more accurate.
Google Analytics (which is what Intertainment is using to publish results) does not record users, rather it shows unique visitors. A visitor is someone who visits a web page. When the visitor goes to the web page, a “cookie” is installed onto their computer. A cookie is a small text file that allows the site to track some information about the user. A unique visitor (as defined by Google Analytics) is someone who has used a browser that does not have the cookie installed for that web page. The problem with using unique visitors as your method of tracking traffic to a site is that it just gives an estimate of traffic. It would be very easy to create a bot (computer program) that deletes cookies and refreshes a given web page. This would increase your unique visitor count dramatically. It is also important to note that if you access the site on different computer or different browsers (even if you aren’t botting) that you will count as a unique visitor on each browser and each computer. That’s why, even without botting, visitors and users are not exchangeable terms.
The easiest way to tell if a web page is using bots to artificially inflate traffic or if the traffic is natural (at least for a website like this) would be to look at the new versus recurring visitors. Google Analytics will provide these figures and the company could easily provide them. I can see no competitive reason why this information would need to be kept secret. I requested these figures from the company but did not receive a response. A web page using bots will have an overwhelming number of new visitors per day versus recurring ones, whereas a site like Facebook would have an overwhelming number of recurring visitors. Using bots to generate recurring site visits would be hard, if not impossible. It would mean deleting and recreating hundreds of thousands of cookies every day. So if you saw consistent evidence of recurring visitors then you could probably rule out the use of artificial traffic inflation, although one could also use a combination of manual labour out of the third world and bots. Of course, Ortsbo is a relatively new product, so you would expect to see a larger percentage of new visitors (as compared to Facebook). However, I am willing to bet (and have through my short position) that the number of new visitors on any given day (as measured by Google Analytics) is something in the 85+% range.
In response to a recent Globe and Mail article, Intertainment has published screenshots of its Google Analytics results, shown here. As shown in the screenshots, the company aggregates the data for Ortsbo.com, Ortsboqq.com, and Ortsboii.com. As far as I know, Google Analytics cannot tell if a unique visitor to Ortsboqq.com hasn’t already visited Ortsbo.com. Meaning that if I were to visit all three pages, I would count as three unique users- according to Intertainment’s terminology. This is not an acceptable methodology for counting unique visitors, let alone users.
Since we don’t have the new versus recurring data to analyze on the Ortsbo.com site, there are a couple of other measures we can look at to try and approximate new versus recurring users and therefore try to determine if bots are being used. Basically we would expect to see a large number of unique visitors on the first day of the time period reported, many visits per unique per month, and exponentially increasing weekly page views if the traffic were “natural” and not artificial.
Using total visits versus unique visits as a proxy for new versus recurring visitors (in a given time period) we see that for the time period reported (March 25th to April 25th 2011) there were 6,398,253 total visits against 5,482,944 unique visits at Ortsbo.com. I have manually calculated the 6.4mm figure and while I double checked this figure, it is possibly inaccurate. This means that of the total visits for Ortsbo.com, only 915,309 (or 14.3%) were recurring (i.e. visited the page more than once in that month). To make that clearer, if I visit the site every day for a month then I count as 30 visits but one absolute unique. Each absolute unique was averaging 1.17 visits for the month. Once again, this is extremely low for an instant messaging service.
Page views are another thing that you would expect to see a lot of if you had 40 million unique users and a desirable instant messaging service. If you had 20 million users logging in just four times a month, that would generate 80,000,000 page views in a month. From March 21st 2011 to April 25th 2011 the company had just 55 million page views despite having 22,430,538 “unique users” on March 21st and increasing that number to 39,526,381 by April 25th. If you discount the new “unique users” who had to visit at least once to be counted, then during that time period 22,430,538 “unique users” generated a maximum of just 38 million page views in a month... very low for an instant messing service that requires you to visit the webpage to use.
The last piece of data provided by the company that shows that its users don’t appear to be coming back to the site is the traffic to its alpha HTML version of Ortsbo. It appears to have gone live in late March (see screencaps here and has since seen 177,491 page views with an average time of 13:22 (emphasis by the company) on the page. While the number of page views seems extremely low relative to the company's claimed number of users, it is an alpha version so I can understand that maybe people don’t want to use it. What I find puzzling is that of the 177,491 page views, a full 173,094 were unique. That means only 4,397 (2.5%) of the page views were people using the alpha version twice.
If we look at its Alexa ranking, which shows Ortsbo.com with a daily reach of 0.015%, we again see that it does not have a traffic rank consistent with 40 million or so recurring users. Facebook has 500 million users and a reach of 42%. Using basic math this would imply that Ortsbo.com has 178,000 recurring users (again, this is users and not absolute unique visitors). In fairness, maybe Facebook has more stickiness so people visit the site more often (in terms of visits per month), but no matter what multiple you want to use to adjust these figures, 178,000 is a far cry from 40,000,000 (225x or so).
Given all the doubts the company’s own data gave me, I decided to check into how I would go about generating similar traffic figures. The first step would be creating a bot to improve my Google Analytics numbers (a program that would clear my browser’s history and then reload the page over and over again). The next step would be signing up to this free service that automatically has members visit participating sites in order to improve Alexa rankings. Alexa only records data from people who have the Alexa toolbar installed, meaning it is a much smaller subset of the population and easier to trick into a top 10,000 to top 15,000 ranking. The final step would be signing up for a service similar to this one to increase Facebook likes. I would now have a web page that could stand up to a cursory (but not in-depth) glance and look quite appealing both in terms of absolute uniques and fans.
Ultimately there is no way to know for me to know for sure how Ortsbo is generating this traffic. However, there are literally dozens of translator programs or apps available (and have been available for years), some of which work with instant messaging services. Creating the program from scratch would also be relatively easy (use the Web software development kit for each instant messaging service and plug in a translator app) and there are already competitors in the market with offerings that work both on IM and mobile (VoxOx). VoxOx appears to be a much more polished product and the interest it is generating seems to be more inline with what one would expect for a translating messaging system-- that is to say, low.
I am not saying Intertainment is misrepresenting the business, only that they are not providing the market with adequate data to make an informed decision; i.e. specifically registered users versus unique visitors or alternatively recurring visitors versus unique visitors. This lack of data leads to inferences (from Intertainment's supplied data) that there is a significant deviation between unique visitors (window shoppers) and the more valuable registered users (shoppers). After all, you would never value a retailer based on how many people walk by, or even look in, the window.
Disclosure: I am short ITMTF.PK.
Disclaimer: The Fund is short the shares. Again, I do not have all the information required to definitively know what is going on with INT’s traffic. The above is just my opinion based on the facts that have been presented by the company. If you feel that anything has been overlooked or there are any errors in the commentary I would appreciate your insight.