The Case For and Against Google's Rumoured Twitter Acquisition

Posted on April 5, 2009 by amit

There’s been a lot of talk the last few days about Google’s rumored “late-stage” acquisition talks with Twitter. While current consensus is that they are merely discussing mutually beneficial product strategies, it doesn’t mean that this deal isn’t being thoroughly considered and dissected by executive committees (not just of Twitter and Google, but MS & Yahoo as well) and tech pundits alike. In this post I will briefly give some background, describe the potential best and worst case scenarios for both Google and Twitter, and give a final opinion.

Background: Why is Twitter Important?

Microblogging (whether it be Twitter, Facebook Status’ or anything else), was always thought to be about about marketing. The business appeal of social media is that you can have a two way conversation with customers directly, at almost no cost, and let them do the dirty work of promoting your product “virally.” Personally, our iPhone-toting, attention-lacking, cause-of-the-day promoting, have-it-your-way generation thinks our opinions are important and should be heard…now. Equal parts narcissism and genuine interest in enhancing communication.

Now that Twitter has achieved mainstream popularity people are no longer think of it as a niche, social network. If it’s possible to organize, make sense of and data mine these millions of thought bubbles, Twitter becomes an incredibly powerful, real-time, social, search-engine. In this light, it is obvious that Google wants a piece.

Why Google Needs Twitter

Although Google dominates search and remains increasingly profitable, there are still a number of factors which could contribute to a Google’s fall.

Why Could Google Die…

View more presentations from Ouriel Ohayon.

With the belts tightening and infinitely high expectations, Google is struggling to find their next big source of revenue. YouTube has the viewers, but no revenue. Other products like Lively (which always felt very pets.com bubbly to me), Jaiku (microblogging) and Dodgeball (location based services) were all killed off. Not only is Twitter (real-time search), a natural compliment to Google’s core product (static search), but Google also has the knowledge and experience to monetize this effectively. The sticking point is the cost. Twitter turned down a $500mil offer from Facebook a few months ago (though mainly in overvalued FB stock). Right now, a compelling offer from Google would have to be in the $750mil – $1bil range

What if Google Buys Twitter?

Best Case Scenario: Google develops an Ad-Sense for Twitter, and is able to elegantly incorporate Twitter into its search engine. Twitter-sense comprises a significant amount of search revenue, the real-time search offering creates a huge gap between Google and it’s closest rivals, and the perception of Google’s superior innovation and foresight lives on (profitably).

Worst Case: Twitter is unable to live up to the hype, a true monetization plan never pans out and people finally realize that no meaningful information is actually contained in a tweet. Investors are angry that Google paid a $1.65bil for YouTube and $1bil for Twitter and haven’t made a dime back. The next new thing is all the rage, Twitter is left to journalists, b-grade movie stars, politicians and retro-nerds.

What if Twitter remains independent?

Best Case Standalone Twitter: Twitter remains open and partners with Google, Facebook and every other major site on the planet to publish conversations streams on every page of the web. IM, SMS and newspapers die. Consumers and brands are able to interact directly, in a powerful way to improve products, services, politics and communication. Twitter is able to effectively generate revenue through a mixture of paid services for corporates and search monetization.

Worst Case Standalore Twitter: With VC money drying up, web valuations plummeting, and no clear revenue model Twitter collapses under it’s own inability to scale it’s services to meet demand. Eventually sells to AOL for $50 and a cheeseburger.

Final Thoughts

Biz Stone and Evan Williams are already rich (having previously sold Blogger to Google), have funding from VC’s who don’t care about revenue (yet), have a product that’s generated popular, mass appeal and has the potential to change the world. The product is still in its infancy though the concept has been validated by the interest generated by the big players (FB and Google). Just like Google can be called an operating system of the static web, by remaining open and allowing 3rd parties to develop useful, innovative businesses on top of Twitter, they can become the foundation for the real-time web. I say go for it, roll the dice, take the mystery behind door number two and build Twitter into a real value-adding, sustainable business. Then again, there’s a lot you can do with $1 billion dollars…

Why Would Anyone Advertise Online Without Google/Facebook?

Posted on December 25, 2008 by amit

As part of my new job I’m in charge of developing and executing a marketing plan for a new website. I’ve been doing a lot of research into online advertising and trying to figure out the best way to allocate my budget. There are a number of ways to purchase ads online.

CPM (cost per one thousand impressions) is the cost to display your ad one thousand times. For the sites I contacted the range was from $20-$70 per 1000 impressions.
CPC (cost per click), instead of purchasing ads which do not guarantee clicks, Google Adwords and Facebook both allow you to pay only when someone clicks on your ads. Better.
CPA (cost per action), you only pay when a user clicks on your ad, comes to your site and does something (i.e. fills out a form). This is obviously the best for ad buyers, but worse for ad sellers (since it depends on how convincing you/your product are to get someone to take that action). I haven’t found any advertisers willing to sell on CPA basis.

There are also a number of other factors that allow you to specify your audience better.

Facebook allows you to specify where (to the city), what sex, age range, keywords, education (even to the specific college, year and major), relationship status and sexual orientation of your target audience (so that’s how my profile info is used). In my example below, Facebook tells me there are 220 single college women between 18 and 23 in New York who like Pizza (hey ladies!). You only pay per click, set daily limits on how much you want to spend and a maximum you want to pay for click.

Google also allows you to specify location, keywords, daily budgets and how much you will pay for a click. The graph below shows me how many clicks I am likely to receive and how much those clicks will cost me (in Rupees):

In comparison the Google and Facebook models will give me a lot more value for my budget. Here’s why:

Cost (CPC is better then CPM) – The number of clicks I get is much higher per dollar with Facebook and Google. One of the major tech sites I contacted (can’t give their name but you’ve heard of them), offered me 174,000 impressions for 4k over two months ($23 CPM). They estimate a clickthrough rate of .25% which would equal 435 clicks. In comparison I can set a maximum bid on Google and Facebook ads (let’s say .50 cents – $2 per click), giving me 2,000-8,000 clicks.
Targeting - While the tech sites I contacted gave a lot of demographic information about their users (what percentage are IT decision makers, developers etc…), it doesn’t mean the individual viewing my ad actually is in that bucket. With Facebook, I can guarantee that the people who are viewing my ad are in the target audience I want them to be. With Google, it’s up to me to define good keywords that people are searching for in order to reach them. Assuming I can do this (there are a lot of tools out there to help as well), I’ll know my ads are only seen by the right people.
Analytics - There is no way for me tell how many times the ad publisher is actually displaying my ad. They can tell me they showed it 100,000 times but I have no way to validate this. All I can see is the number of people that actually click on that link. Google and Facebook give me tools to see how well my keywords and ads are performing, guaranteeing my investment.

In short I can understand if you are Coca Cola you don’t really care who sees your ad. In fact you don’t even care if people click on your ad. You are huge and all you want is to continue reminding people they are thirsty, your target audience is everyone. But for anyone else why would you advertise with someone who doesn’t offer targeted CPC ad purchasing? Your thoughts appreciated…

The Transition to Real-Time, Social Search

Posted on December 18, 2008 by amit

John Batelle recently wrote a great post about the shift from static to real-time search. To paraphrase he says that Google is/was awesome because for the first time, you could quickly and accurately get static information about any search term. Though this strength is also a flaw. By nature, newer content will have a lower page rank and thus appear farther down in search results. The emerging “liveweb” will be able to answer the question “What are people saying about X right now?” The current best example of this is Twitter Search. But this isn’t enough. The question I really want to ask (and the real moneymaker) is “What are people I trust saying about X right now?”

A few examples where static search is insufficient:

Breaking news/current events (we’ve seen many examples where Twitter breaks news stories first – i.e. San Diego Fires, Mumbai Terrorist Attacks, etc..,) Many times Twitter is able to provide (arguably) more accurate real-time reports that the news that traditional news outlets later pick up on. While Google is incorporating News Results at the top of it’s list:

It’s still not enough if I want to get a grasp on the current situation/perspective/events occurring now (though interestingly enough the NYTimes ad on the right is incorporating real-time content).

Finding the amount of times people are searching/mentioning a term – Google Trends allows you to verify the pretty obvious fact that searches for Santa tend to spike at the end of the year, but did you know Canadians were the most actively searching for Santa and that progressively each year less people are asking Google about him?

The equivalent of this is Facebook Lexicon (interestingly no label on the y-axis, not useful for an actual research tool) and Twitter Twist (been down for quite a while). This is huge for PR and Product companies who want to be able to monitor how much buzz a person/event/product are receiving.

Product Reviews / Social Shopping – When most people want to buy a product online, they head to Google. This is why advertising on search is a big moneymaker (as opposed to advertising on social networks which is seriously struggling – people go to Facebook to waste time, flirt, look at pictures of their friends drunk and write each other stupid messages, not buy products). There have been a bunch of entrants into this market but none so far has provided a compelling product and achieved critical mass. Hopefully now with Facebook Connect/Google Friend Connect, someone will make this possible. If I’m looking for a camera, I want to know which camera my friends use and love, then find the cheapest place to buy it. The first person who makes this easy to do (using my Facebook social data) is going to make a killing.

Google dominates static search, Twitter knows what you are doing, Facebook controls your social graph and Apple knows where you are. The company able to blend all of these services (static, real-time, social, and location), while maintaining the strictest privacy controls, which will dominate the next era of the web

Billion Dollar Business Idea

Posted on October 16, 2008 by amit

I have an idea, it’s a good one too. This is gonna be big, we’re talkin’ bigger then YouTube, Facebook and Cute Overload… combined. I’m goin’ to let you in on the secret, but there’s a catch – I need your help. I just need you to answer to the following question: What is the killer service that combines people’s location (g/iPhone) and existing social data (Facebook, OpenSocial) to drive real life interactions and sales? Give up? Me too…

The convergence of the mobile and social web (although over-hyped for years) presents an enormous opportunity. For a long time we’ve heard rumors of a mystical app where you specify you specify you want some new size 13, white, Nike Zoom LeBron VI’s. You walk by FootLocker and all of a sudden you receive a message letting you know they have it in stock, in your size, and if you buy it now, you get 15% off. Or the app that let’s you know the cute chick across the bar is single, and shares your love of Stan Getz and Quentin Tarentino. Perfect? Not really, privacy concerns and unwelcome solicitation can quickly ruin this product. Success with location based services (LBS) can only be achieved by giving users complete control over their privacy, and continuously building trust by providing more useful and accurate services.

I’ve spent some time playing around with the current LBS offerings Loopt and Yelp, but I think there’s a lot more that can be done. This summer I came across an article in the NYTimes describing Sense Networks. I downloaded the CitySense app on my blackberry and was floored, absolutely incredible. It allows you to see in real-time where people are, if the population density is higher or lower then normal in certain areas and recommend places you would like to go based on where you’ve been (currently only for SF).

But this is just a for-fun demonstration of their technology. Some of the tough questions that Macrosense is trying to answer include:

Where do people make the decision to go to one place versus another? Which places in the city are “influence points” versus “stops en route”?
From where do the most people come from before arriving at a particular location? Where do most people go afterwards, and how does this change throughout the day?
What is the demand and elasticity of demand for places, activities and services by income level? What are the most dramatic increases and decreases?

This is dope and extremely profitable. The information could be used for marketers and business trying to determine the right location for their store. The company is backed by hedge funds and I bet are making some scary decisions based on their findings (i.e. economy goes down 10% no one is going into the Apple Store, short the hell out of the stock)

Another great, new product is the Sekai Camera. The concept is not only awesome, but the video is highly entertaining:

As we continue to make our phones smarter with faster and ubiquitous connections and GPS, we can enhance our real world experience by utilizing the thoughts, opinions and opportunities presented by friends and strangers alike. Ideas anyone?

Ways to improve Twitter – Special Interest and Contact Groups

Posted on September 26, 2008 by amit

The first part of this article described why people microblog and the current problems with Twitter. There were two questions:

How can you quickly find and engage in conversations around topics you are interested in?
How can you identify which notifications are important and send/receive them in the appropriate method and frequency?

The answers: special interest and contact groups are discussed below.

Special Interest Groups

There’s a lot of people constantly engaging in dialogues I want to be a part of, but just don’t know about. I love that the advanced search is able to filter tweets based on words, people, places, dates, and attitudes, it allows me to, for instance, find all people within 15 miles of lower Manhattan talking positively about the Mets. The problem is once I have the results, what can I do with them? Sure I can subscribe to an RSS feed but that takes me out of the whole call and response flow that makes Twitter so engaging. Hashtags is a 3rd party service which allows you to create and track tags.

A couple of problems with hashtags:

It’s a third party service and not integrated into Twitter.
No way to subscribe to a specific hashtag and see it in my stream.
There is a lot of redundancy (i.e. Barack, BarackObama, Obama, Obama08, Obama-President)
There is no taxonomy (#family:genus:species)

Recently Twitter released a special strem for the elections. This is awesome and I hope a sign for whats to come. Here’s my wishlist for special interest groups on Twitter:

Be able to subscribe to any advanced search result.
Give a short name to the result and have a list appear underneath your followers.
Be able to specify if tweets from a specific group appear in your stream.
Be able to click on any of the groups and have a real-time, updated page of the results (like the Election Page):

Allow for the custom insertion of dynamic filters within a group. (For instance if election is the main keyword you are searching, be able to add/delete dynamic filters like Obama, Biden)

Special interest groups would allow new users to quickly ramp up their involvement and allow long time users to filter all their messages into clearly discernible buckets by specifying a combination of people, topics and location. Some examples of what this could to do:

Allow Obama to identify and address all people who were positively supporting Hillary in Virginia during the primaries.
Allow Apple to follow the launch of their newest product. Combine this with trending topics within the group and identify most active complaints. Send a notifications to all those afflicted with a message that the most recent update solves their issue.
Allow me to find all people who have discussed Twitter improvements in the last 2 weeks, and ask for their suggestions.

Contact Groups

The premise here is that some users I really care about, and some I only care about if they have something really, really important to say (to me). Currently Twitter allows you to receive all messages, or messages directed to you via SMS. They also let you turn it off during certain hours:

This is nice but there’s a lot more that can be done. Here’s how I would structure contact groups:

Allow users to create unlimited contact groups and easily add and remove users and special interest groups. These groups are private and cannot be seen by other users.
For each contract group allow the following receive options:

Delivery Method: Modify the settings for how you would like to receive different types of messages including: all messages, messages directed at you, and high priority messages (discussed below). The following options are: Twitter Stream, SMS, Email, IM (allows multiple select)

Frequency : Modify how often you would like to receive messages from this group: Real-time, Daily, Weekly (daily and weekly would send aggregated list of all messages)

For each contract group allow the following sending options:

Private Messaging – Select which contact groups this message is sent to. Specify if messages to this group appear in your public stream (twhisper).

Priority Alerts – Allow users/services to send messages they deem as critical (breaking news, emergency messages). There would need to be a system in place for people who abuse this (similar to flagging comments?), but this could be immensely powerful.

There are also (at least) 2 new revenue streams here:

Give personal accounts a small number of priority messages, while offering more for pay
Create customized reports of keyword trending for your particular market/product/campaign. (update: stumbled on this site: http://twist.flaptor.com/freq?gram=twitter%2C%20friendfeed%2C%20facebook)

In summary, special interest groups would allow anyone to identify and engage with user’s who are discussing a particular topic. Contact groups would allow full control (privacy and frequency) of how messages are sent and received. In combination, Twitter would much more attractive to business by being able to analyze the abundance of opinions being shared and individuals by customizing their Twitter experience to fit their schedules and habits.

Why Twitter? and Problems with Microblogging

Posted on September 24, 2008 by amit

I was asked recently to take a look at microblogging and suggest improvements to Twitter. First off…

Why Twitter?

It’s a difficult question to answer, even for those who do it regularly. You start off tepidly following your in-the-know co-worker and soon you are fanatically telling your 513 followers that you are:

Recently the NYTimes Magazine had a great article Brave New World of Digital Intimacy, which describes the growth, stigmas and motivations of microbloggers:

“Social scientists have a name for this sort of incessant online contact. They call it “ambient awareness.” It is, they say, very much like being physically near someone and picking up on his mood through the little things he does — body language, sighs, stray comments — out of the corner of your eye…

The growth of ambient intimacy can seem like modern narcissism taken to a new, supermetabolic extreme — the ultimate expression of a generation of celebrity-addled youths who believe their every utterance is fascinating and ought to be shared with the world…

This is the paradox of ambient awareness. Each little update — each individual bit of social information — is insignificant on its own, even supremely mundane. But taken together, over time, the little snippets coalesce into a surprisingly sophisticated portrait of your friends’ and family members’ lives, like thousands of dots making a pointillist painting.”

The article continues to describe how these short updates allow you strengthen your relationship with loose acquaintances. When that cute friend of a friend you once met at a party posts pictures of their trip to Buenos Aires, or that guy you met at a conference starts ripping apart the Mets bullpen, they are essentially reminding you they exist and offering you an opportunity to connect. Microblogging also provides an opportunity for following breaking news, a targeted audience for marketing your product, and (as opposed to blogging) requires little to no commitment.

Problems with Microblogging

Though Facebook still dominates (38.2 million unique visitors in August vs. 2.3 million for Twitter) Twitter is blowing up. I will never, ever, forget, the reason Twitter is so successful is because of it’s open API. Twitter’s popularity has grown because it has allowed 3rd party developers to build a whole world of enhancements to its core services. At first, finding ways to improve seemed like an impossible task. My initial thoughts about geographical, social and keyword visualization was already done. So was any enhancements to messaging attachments including, events, pictures and videos (a la Facebook Wall).

Rather then try to identify new features that Twitter could add, I took a step back and tried to identify the problems with their core service. When you first join Twitter, it’s not fun. You can follow the 5 people you actually know and search for the people you want to know (i.e. Henry Rollins, Kevin Rose), but you still aren’t involved/committed in the Twitterverse. Then there’s the reverse: Barack Obama is following 83k people. Given his busy schedule, how can he slice and dice all the updates he receives to, for example, reach out to all the Hillary supporters in NY who recently pledged allegience to Palin? Problem #1: How can you quickly find and engage in conversations around topics you are interested in?

Ok, I’ll be honest, some of my “virtual” friends I don’t really like. I’m not sure why I agreed to your Facebook friend request but I don’t want to receive any more emails inviting me to your poetry reading and I’m too lazy to defriend you. On the other hand, when my friend gets arrested I want a text message asap. Problem #2: How can you identify which notifications are important and send/receive them in the appropriate method and frequency?

Part 2 of this article will discuss the solutions to the problems above: special interest and contact groups.

Pluggd – Speach recognition technology for video

Posted on June 11, 2008 by amit

This is pretty awesome:

Publishers have the option of attaching a heat map of sorts to their videos. The map shows up below videos as a variously colored bar, which ranges from blue to red and activates when the user types in a particular topic. For example, if you’re watching a clip about a golf tournament, you can enter “Tiger Woods” and the bar will show you where the commentator spends time discussing and showing footage of that famous golf player. The topics are automatically detected by a combination of speech and contextual analysis, so publishers don’t have to break down their videos manually.

New Google APIs (Language/Charts/Visualization)

Posted on March 20, 2008 by amit

Google releaesd three new API’s this week.

The first is for automatically translating content . Additionally there is language detection which guesses what language text is in (and with what certainty) .
The second is an update to the existing Charts API. Charts API allows you to really simply embed PNG of different types of charts and graphs. For example, and example 2. Try manipulating the variables in the URL.
Last is the Visualization API. Allows you to create embeddable widgets which display XML is different ways. All sorts of widgets are already available in their gallery.

For more google APIs checkout: http://code.google.com/more/

Nu Analytics

Posted on February 11, 2008 by amit

Using Javascript/Ajax we can build significantly better UI’s and create a much better user experience. Many companies still can’t effectively translate a better user experience into value ($$$). The biggest reason for this is that the statistics captured and reported by many are still page views. Enter the Web 2.0 Anaytics offerings:

French company Alenty is able to track what area of the screen you are looking at (duration): Alenty Demo

Nuconomy goes one step beyond, allows you to track not only Ajaxy (duration) stats, but interesting correlations as: friend request to purchase ratio and number of reviews written vs. clicks on ad. Pretty awesome…

There is a huge opportunity to offer free (or cheap) Web 2.0 analytics before the big players (Google, Omniture) do. Maybe we can then develop photo galleries that don’t reload between each photo…

The next step is to develop targeted ads that can seamlessly and unobtrusively change based on the your ever changing content…

http://www.alenty.com/xwiki/bin/view/Demo/AudienceDemoMultiContent

Amit Klein

Figurin' it out…

Category Archives: Business