This is, in all likelihood, the worst Twitter client ever created.

At the same time it makes me wonder how tweets might be displayed in public places. The new geolocation APIs allow for each tweet to carry location information along with it. There's no reason you couldn't have displays in stores, at events, on billboards, etc. that just display a constant stream of tweets in the immediate area. Like a real life shout box. I wonder what that would look like? A application that simply displays local tweets could be a pretty big deal for digital signs.

If you look "spam" up in the dictionary, you'll find two definitions. There's the "canned meat" and then there's the "unwanted email." At Twitter, we see spamming as a variety of different behaviors that range from insidious to annoying. Posting harmful links to phishing or malware sites, repeatedly posting duplicate tweets, and aggressively following and un-following accounts to attract attention are just a few examples of spam on Twitter. Like it or not, as the system becomes more popular, more and more spammers will try to do their thing. We’re constantly battling against spam to improve the Twitter experience and we're happy to report that it's working.

Really? I still see just as much spam as I did before. Just as many robots following me as there were before, and trending topics is still dominated by garbage and nonsense. Sure, there aren't as many bots hitting the API as their used to be, but the signal-to-noise ratio still favors the spammers.

They should hire the team who built Gmails spam filter. Those people are brilliant.

A fully interactive map showing everyone who follows BBH Labs (@bbhlabs) on Twitter. Also visualized as a heatmap.

Twitter is still something of a mystery to those of us in advertising and marketing. Everyone thinks they need to be on top of it, but no one is completely sure now to use it. Even fewer people have an idea of how to measure whether or not they’re using it effectively. Most of the time brands think about Twitter like this: Create an account, start tweeting, and then measure success by looking at how many followers we have. But that doesn’t tell you the whole story. In fact, that tells you almost nothing.

Here's a little something I wrote up about how to use the Twitter API and Google Fusion Tables to draw insights about those who follow you (and others) on Twitter. It's amazing how much data is publicly available, and the kind of analysis you can do with a few lines of code. If you ever wanted to know how to map 12,000 points on Google Maps, this is your chance to learn how.

Gelocation information provided by modern mobile devices puts analytics data into an entirely new context. Not only do you know what someone did and when, but now you know where. I'm particularly intrigued to see how these kinds of APIs develop in the future, and at what point brands start using public API data to aggressively target their competitors consumers in a very hostile manner. We aren't far away from the breaking point where you can challenge massive incumbent brands by targeting their most vocal consumers right outside of their house.

The study found that 65 percent of the largest 100 international companies have active accounts on Twitter, 54 percent have a Facebook fan page, 50 percent have a YouTube channel, and one-third (33 percent) have corporate blogs. Only 20 percent of the major international companies are utilizing all four platforms to engage with stakeholders.

Do the world's largest multinational firms even need to be utilizing Twitter, Facebook, YouTube and a corporate blog? Seven of the top ten companies on Fortune's list are oil companies. I'm not sure that ExxonMobil is really going to be able to elevate it's brand with a new Facebook fan page or have any kind of meaningful "conversation" with consumers on Twitter. Are there things that they could do? Sure. But you shouldn't get a pat on the back just for showing up.

I'm also bothered by "studies" like this which make zero attempt to calculate the value added from social media. It costs money to create and maintain a voice on major social media platforms. The service might be free, the content isn't. It's only worth the cost if the return is real value, in whatever form that value might take (PR, sales, awareness, etc.). When you start looking through the list of Fortune 100 companies that are using multiple social media platforms you realize that, while they might be there, they don't have anything to say.

First, we've created an open source directory for the entire company. This lists all the public software that the engineering teams have created or contributed to. Much of Twitter's success has been enabled by open-source software, and we want to give back. Everyone is welcome to use this software for their own projects, and if the project is Twitter-related, so much the better.

Wow. I knew that Twitter was built on a lot of FOSS, but I didn't realize they contributed to this many projects.

I wish the attitude towards FOSS was different in advertising. It's an incredibly hostile industry that uses open source code constantly but rarely gives back. Most of this has to do with the way the contracts between agencies and clients are written. Most clients don't understand open source. They want to "own" everything so their competition doesn't get a hold of the valuable code, ignorant to the value and PR that would have been created had they decided to license things under an appropriate open source license. If you were really concerned that competitors would copy your concepts, your could license things under the GPL or another license that doesn't allow for proprietary software. At worst, your competitors could improve on your code.

What would happen if you had a digital agency that set out, from day one, to make as much of their work open source as possible? If you could get clients to see the benefits, and sign on to such an agreement, you could elevate the entire industry. Silicon Valley understands this. Why doesn't 5th Avenue?

We categorized each incoming tweet as about the Super Bowl itself, about the brands or the commercials, or neither. Dividing each group by the total volume of tweets, we produced the graph below which represents a minute-by-minute reflection of people's thoughts and emotions during the game.

I was reading over this post by Twitter's Kevin Weil and realized that what they are actually looking at is variations in the noise that happens on Twitter, and trying to draw some kind of insights from it. There isn't really much substance. It's akin to looking at your spam folder and noticing that there are more ads for Viagra than there are for Cialis. The noise generated by people I know isn't that much better than noise generated by random strangers. It's noise.

I think this is what Google might end up getting right with Buzz. They've built a recommendation engine into their service similar to the one in Google Reader. And since it's actually integrated with Google Reader, it already knows what kind of content I find interesting. If everyone I follow starts spamming some nonsense and it's not relevant to me, I'm either going to see it once in a stack, or I'm not going to see it at all. I'm not going to see it flood my feed.

Facebook's main focus was community and Twitter's main focus was conversation. Neither of them really gave any thought about content or context. For better or for worse, Google already knows who I am. Google knows what I know, and it knows what I don't know. It knows what I find interesting, and it's just starting to learn about who my friends are. It definitely knows to much, but everything it returns to me is useful. The same can't be said for Facebook and Twitter.

As Twitter evolves, and more people share what’s happening in their own world, we want to provide another way for people to discover topics that may be relevant to them. Last week we began to slowly roll out a new feature called Local Trends to expose what people are talking about on the state and city level, and today we've fully launched so everyone can use it.

I'm starting to believe that Twitter is really just a big joke. I figured that a localized version of trending topics would have fixed the issue (that being that trending topics is usually full of nothing but junk and spam), but it appears that the NYC area trending topics is somehow even more junk ridden than the global setting. The ratio of normal users to marketers and spammers is completely out of whack. That's really the only way to explain it. Most tweets don't provide any utility—to anyone.

And I'm starting to wonder if perhaps all the hype over social media is general is completely overblown. I've been looking at a lot of Facebook Pages recently with fan counts over one million, and the level of noise is through the roof. You would think there would be some level of control with 100,000+ "social media managers" on LinkedIn, but there isn't. How do you manage millions of people anyway? When you look at some of the leading brands on Facebook most of the "interactions" are spam. It's junk. People are looking at things like follower count and the amount of comments, but aren't looking at what's actually being said. Look at the Wall for Starbucks or Coca-Cola. For every interaction with a consumer worth anything at all, there are at least 10 pieces of spam. Is this really elevating the brand?

December 11, 2009

The Defense Advanced Research Projects Agency (DARPA) released 10 red weather balloons across the country and offered $40,000 to the first group to locate them all. Riley Crane, of MIT, explains how his team bagged the prize by locating the balloons in fewer than nine hours.

This week's Science Friday covered the recent DARPA Network Challenge, which celebrates the 40th anniversary of a little thing called the Internet (which DARPA helped create). The interview is a little over 10-minutes long, and definitely worth a listen. MIT's Riley Crane explained how his team won the contest by creating a "recursive incentive" – what amounts to a social media affiliate program. As he explains it,

We offered a recursive incentive, which allowed us not only to reward the people who actually found balloons, but we could also reward the people who helped us find those people.

The first question you might ask yourself is this, "Why did DARPA fund this, and how is this even remotely useful?" Well, as Mr. Crane goes on to explain,

… imagine that a building collapses during a natural disaster and you need to rapidly find 10 people who can operate heavy machinery in a certain location. I think these are a lot of the challenges that are facing society that nobody really knows yet how to mobilize people on a large scale, and I think that's what our team was able to demonstrate, that these things are possible and that we've built the technology to try to do it.

I can't wait to see what the results of Mr. Crane's analysis. Hidden in this experiement, they may be some data points that help reveal exactly what the thresholds are (and the incentives need to be) in order to cause people to take an action.

Today, when you present an advertising concept to a client, you will inevitably launch into a discussion on how your experience, application or content spreads virally, and what your strategy is for making sure that something "goes viral." Doubly so if it's a digital concept. And while you can't guarantee that anything goes viral, I've long thought about the idea of incentivizing sharing. A monetary incentive, in most cases, is the best incentive. But in the digital realm, virtual currencies may be just as good.

This is a topic I intend to explore with more depth in a later post. Hopefully, after I get my hands on MIT's official report.

Now, immediately after conducting a search, you can see live updates from people on popular sites like Twitter and FriendFeed, as well as headlines from news and blog posts published just seconds before. When they are relevant, we'll rank these latest results to show the freshest information right on the search results page.

Google has now officially joined the real-time web bandwagon. Starting today, they are going to be integrating sites like Twitter and Facebook into search results. The key word that they use in describing their approach to real-time search is this: relevance.

And I hope it works, because right now that's the problem – search on Twitter and Facebook usually returns irrelevant nonsense. When you go onto Twitter, trending topics is filled with spam. Is Google going to start injecting spam into search results? I hope not.

If you look at the "Hot Topics" on Google Trends, and click on any one of them you'll see tweets start showing up in-line and under other search results. For the most part, they seem pretty irrelevant to me. Looking at it now, it seems like it's just keyword matching. I would presume that, as Facebook becomes more open, Google will start indexing Facebook status updates and user photos as well.

I could be proven wrong, but I'm not sure that social media has a place when running everyday search queries. For the most part, the more interesting part of social media is finding out what's relevant to you – based on your relationships to other people – not necessarily the keywords you searched for. And when I want to ping the Twitterverse with a question and get a response, I don't do that through search. I do that by actually sending out a tweet and seeing who picks it up.

Perhaps the future of social search will work like this: I go to Google and type in a search term. That query gets posted to my Twitter feed automatically. As users reply to me, these are the tweets show up in the search search results. It's this level of personalization that seems to be missing in Google's current real-time web approach, and why they seem to fall short with their goal of brining relevance to social media.