Friday, August 7, 2009

When Competition Actually Helps

Opera CEO Jon Stephenson von Tetzchner was recently asked about competition from Google Chrome and how that effected his company's browser. See NPR aricle. Surprising, he stated that Chrome's entrance into the marketplace actually helped Opera. "The net effect of Google Chrome has been 20 percent more downloads of Opera every day since [the Google browser's release]." How can that be so? Greater competition lead to more sales for one of the competitors.

In the case of the Opera browser, I believe the answer lies in the shared mega enemy known as Microsoft. Anything that prompts longtime Microsoft Internet Explorer users to try another browser helps to break the monopoly and encourage users to experiment with other browsers. The theory goes, after having tried Chrome, the same user is more likely to try yet another established brower in the marketplace. Another way to state the effect would be that two smaller competitors are both delivering the same message to the public that strengthens and reinforces the communication--i.e., MS Internet Explorer is not longer a cutting edge browser, better free software is out there so go try it.

One my stock business models relating to this dynamic is phenomenon of restaurants huddled together in a neighborhood. Two neighborhoods in my hometown of St. Louis exemplify this: the Italian quarter known as The Hill and the cluster of ethnic restaurants located on South Grand. There must be 20 Italian restaurants located on The Hill, a relatively small, working class neighborhood. How do they all survive? Would it make more sense for the restauranteurs to spread out? The cluster of Italian restaurants on The Hill draws patrons from all over the city (including tourists). The cluster of restaurants creates a critical mass focusing on a narrow market segment. The Hill dominates this small segment of the restaurant market and, thus, the competitors help each other. However, this only goes so far. At some point, the market will not bear additional restaurants (however, this weeds out the worst of the bunch making the survivors better).

A five block stretch of South Grand Boulevard near Tower Grove Park is packed with ethnic restaurants: Vietnamese (3), Chinese (2), Thai (2), Sushi (2), Persian (1), Afghani (1), Ethiopian (1), Italian (1) plus three coffee / sandwich shops and an American style dinner. This clustering of affordably priced ethnic restaurants upon one small stretch of road brings in customers from all over the metro area. But too much of a good thing can turn into a bad thing. With the recent economic downturn, two Vietnamese restaurants closed. One was the highest priced Vietnamese restaurant in the neighborhood and, consequently, the weakest competitor.

To sum up, the grouping of small competitors into the same space can, under certain circumstances, actually be mutually beneficial to those in the group; however, this strategy has its limits.

Thursday, May 14, 2009

Twitter Bomb, How It Works

I've seen a few blog posts on the new phenonomon of twitter bombing (here and here). We're all familiar with the google bomb. The twitter bomb, given that twitter is now something of limited universe search engine, apparently is an adaptation of this older tactic. I don't claim to be an expert at this so feel free to jump in and correct anything you read here in the comments section.

Here it is in a nutshell. This is where you search twitter posts, aka "tweets". If I flood twitter from multiple user accounts with the same tweet, such as "Harvard sucks", anyone searching twitter for the words "Harvard" or "sucks" is likely to come across my message. I could further attach a link to my message, such as "Harvard sucks". This might get additional traffic to It's a form of online guerrilla marketing. The problem with this is that other tweets are continually pouring into the twittersphere and those containing the words Harvard and sucks will degrade my results. The new tweets from outside sources (lacking my link) will push mine out of the search results. It's like trying to swim upstream against a swift current. You can only pull it off for a short burst, if that.

Here is the twist making the tactic more powerful. What if I and my confederates in the twitter bomb use the the one word preceded by a hash mark--"#sucks"? We are presumably the only people out there with tweets using "#sucks" so we own that result but nobody is searching that term. How does this help us? The answer: Trending Topics. Within each twitter user's home page is a section on the right tool bar by this name (see graphic). See the terms with the hash mark in front? Those are twitter bombs. The people behind those bombs have flooded enough tweets using their term into twitter to cause it to show up on the "Trending Topics" section published on every twitter user's home page. If a curious user clicks on the link, they see the full message that can include the link you are promoting.

Someone marketing the candy Skittles recently did a twitter bomb to help with a free promotion for a laptop giveaway. The whole point of the google bomb was to get readers to this product promotion page. I have no idea whether this is a legit promotion or a scam but, hopefully, you get the point.

Saturday, April 4, 2009

Open Letter to Twitter

Talk about taking a niche startup (micro blogging) and blowing up into an industry phenomenon, one can't help but tip the cap at the founders' vision for Twitter. The beauty of their plan was not just the concept of micro blogging but, also, early adoption of cel phone tweets and rich availability of APIs enabling others to spread twitter content around the web. Twitter is now a communications appliance.

But every entrepreneur must tweak his tweets. The plan must evolve based upon changing market conditions and better understanding of user desires. The following is a humble suggestion that I believe is shared by many in the twittersphere. The first commandment of Twitter is "Thou shalt not tweet more than 140 characters". Awesome, should have been on the stone tablets God handed down to Moses. Just maybe the Twitter Gods should chisel an addendum to the first commandment: "Thou shalt not tweet more than 140 characters exclusive of URLs. Here is the problem, let's say I want to tweet about yesterday's Twitter story in BusinessWeek. Here is the URL: For those counting, that's 102 characters leaving me 38 for the content of my tweet. Quite constraining but at least I get six or seven words. What about URLs that are themselves more than 140? There is no way to tweet about that web page unless the creator possessed the divine wisdom with which to append TwitThis code to the page. One of the magic features of the TwitThis code is that no matter how long a URL may be, it condenses it down to 26 character if necessary to get under the 140 limit: i.e., when posting a tweet via a TwitThis link, the user always has 114 characters to work with.

Why isn't this convention followed in all twitter posts? Why does URL length so hamstring the tweet author? I can think of no good reason other than the Twitter Gods are busy (perhaps 'chatting' with Google). Suggestion 1: automatically crunk every URL inputted in a tweet down to 26 characters leaving the user at least 114 characters (as is already done with the TwitThis code). Suggestion 2: create a separate text input field below the input box that the the Tweet goes in to be used solely for URL linking. The tweet content goes in the text box and any URL the tweet refers to goes in a text box immediately below. The users gets 140 characters in the tweet box but any URL inputted in the field immediately below would not count against the limit. The tweet input format under option 2 would be akin to reddit. Suggestion 1 is the easiest to implement although from my personal perspective I like #2 better. But either solves the problem.

Signed @jjray

Update: In the comments here and at reddit, some posters attacked me for not mentioning the tinyurl option. I responded that it was stupid for a service as ubiquitous as twitter to require its users to navigate to another website for something so basic as compacting a URL's length. I argued that twitter should itself compact URLs for its users to help them better fit messages into the 140 character limit. Several tests confirm that twitter now compacts URLs under certain circumstances but not as well as tinyurl. It's a half measure.