Web History – the story inside out

Every business has a ‘Statement of Purpose’ from two points – the Marketing view point and the Operation view point. The Marketing purpose is to look out for value to the customer, while the operation objective is to look out for how you achieve that, and what you, as a business are going to achieve.

This whole concept is quite discernable through Google’s latest Web History link that is available on their homepage.

I have tried to present what Google claims to offer you from a marketing view point, and what it plans to get back.

What You Get

By enabling web history, you get to search all previous searches you made. So, you can also track all the previous webpages that you previously had visited through Google. You can perform a search on only those pages that you had visited through Google. However the most important is the personalized search. Say, you are a software programmer and your brother is a tourist operator in Asia. So, you search more on programming syntaxes while your brother searches on tourist attractions and places. The personalization part suggests that while your search on a keyword ‘Java’ might result in more weightage provided to the programming language, the same search performed by your brother might results in links to the Indonesian island.

All the more reason why you should enable web search history on!

What Google will get

Computer Science students might have studied what is called data-mining. Enabling web search history might turn out to be Google’s greatest attempt at gathering data about its users. Now, let us understand this with my own example.

I search on a wide variety of subjects. I search on technology, MBA stuff, India related, etc. Till now, all these are discrete pieces of search which is of no use to Google. Now, I have a gmail account. I use this id to log onto Gmail, orkut, blogger, and also search.

Using my information on orkut, Google knows I am a 23 year old student residing in India. Google also knows that I blog. Also, it now knows that I search for technology and MBA stuff. And well, it knows everything from my eye color to height to what kind of a person I am. So, Google knows almost everything about me that I possibly know about myself.

Not just that, Google also knows who my friends are, and in quite sometime, when YouTube login is integrated with Gmail id, will also start learning my video preferences. What more needs to be studied to offer you the best product available. In fact, with such extensive data mining, I see the future of Adwords not just relying on the keyword relevance. Rather it will be based on more extensive profile analysis. Like the Adwords might depend not just on the keyword, but also on the age, income level, ethnicity, past searches, friend influence, etc.

All this is simply a sign of exciting things to come!

David Vs. Goliath – Does it hold true?

Google recently launched MyMaps, which allows you to create custom maps with your own placemarks. This has sort of made websites such as Wikimapia and other mashups redundant. There is one another site called Picli, which is regarded the Digg for pictures. But, it shall not be too late before Digg introduce a pictures section as well, looking at the demand for a picture section on the Digg website. Then, there is one site I had talked of which helps in downloading YouTube videos. What if YouTube provides that functionality on their website itself?

All these minnow websites have been treading on the thin line unattended by the major websites, and when one of these small sites start to get real numbers in traffic, these major websites crush the little threat that they hold by providing the very facility. It does look that the concept of David getting over Goliath is seemingly impossible in the case of internet.

What’s the strategy then? If you look closely, you will find some sites that have taken better to the idea of complementing big sites, and successful at that.

What’s the strategy here?

Simple. Go through the Terms and Conditions of the site, and you shall know what areas the website does not and will not venture into. For example, YouTube does not feature porn. So, PornoTube makes a successful strategy. Most networking websites require you to be 18 and above. Maybe, a networking site for children should be a nice idea, though regulations and moderatorship are a gray area here. But, the point is that if you would want to replicate the success of major websites, dont clone them on a niche where the very site can be a competitor. Choose where they would not venture into and there is a nice chance that you hit the right spot.

Widget Networking – the next in-thing?

MyBlogLog has had a fair share of criticism in recent times. It was accused of revealing Google Adsense information that was hereto confidential under the Terms of Service. However, MyBlogLog continues to be popular and rightly so.

MyBlogLog has probably been the first service that has rejuvenated bloggers’ interest in other blogs. Till then, communication between bloggers tended to be restricted to commenting on blog posts. But, ever since MBL got popular, bloggers tend to have struck a chord with their readers. For example, I now know many of my blog’s readers by face. Thanks to the MBL widget.

Widget networking seems to be the next in-thing on the net after social networking. Widgets are small pieces of code, predominantly javascript that can be inserted anywhere in your code. MBL in particular lets you know your recent visitors.

Recently I was approached by another website who introduced to me their new widget called AutoRoll. Their product fit well into this topic, so I decided to introduce the widget on my blog. AutoRoll is a network of blogs like MBL, except that the widget for AR displays links that my site visitors would most likely want to visit. In effect, it is a self-learning widget that tracks the link which my visitors click on the widget and learns to display the most clicked ones on my blog. Similarly with my blog on other AR user blogs.

Widgets as these are getting common. Every blog has quite a few widgets running on them, which do much more than simply increase the page-loading time. For example, both the widgets that I have discussed here help the blogger get closer to his readers by knowing their preferences and tastes.

However the question remains if this widget networking phenomena shall last. There cannot be just too many widgets on a blog. The page would simply stop loading then. In my opinion, given this constraint of only a few widgets that can be put on a blog, there is going to be a fight for this blog space. And in this, only the widget that fulfils the needs of the blogger shall last.

MBL has indeed taken a leap ahead. Not only has it established a social networking between the blogger and the reader, it has also become property of one of the Internet giants, Yahoo. However the other widgets like AR need to keep innovating to make sure that they don’t die a natural death. For example, while reviewing the AR widget, I felt that the ‘site statistics’ link not being optional can be a great negative aspect. Not all bloggers might want to reveal statistic. Another point is that over a period of time, some blogs become popular and others might not. That would mean that these less-popular ones pull out the widget because their statistics are bad compared to the links that are shown on the widget. In the course of time, the widget might come to a stage where it cannot grow further.

AR is no doubt in beta, and hopefully the product can only improve much more from here, but I am only trying to highlight the constant-competitiveness that these widgets must ensure in order to remain popular amongst users.

Next: Google Ads for your Video Games!

TechCrunch recently announced that Google has made a $23 million acquisition of AdScape Media. Adscape is a San Francisco based company that is credited with developing the Real World/Virtual Word Gateway (shortly known as RVG). This acquisition means something significant to Google since RVG helps in taking dynamic ads to a new platform so far not ventured into by Google; Video Games!

In fact, this is not innovative thinking from Google anyway. Microsoft, in May last year, had announced acquisition of Massive, a well established company whose technology has already been taken up by popular gaming sites like Miniclip and SOE. Google had however hinted at such an acquisition a long time back.

Now that the two rivals have taken on the video gaming platform, it is interesting to see the advantages and disadvantages that the two companies shall hold in having made such an acquisition.

Microsoft and Massive

Microsoft’s acquisition was very much a necessity owing to XBox Live and the MSN games. This is in the sense that Microsoft already had the necessary infrastructure in place for them to start reaping the harvests of the acquisition rightaway. This apart, Massive already has a huge list of partners of their
own. This has helped in proliferation of their advertising network straightaway.

That being said, the advertising model used on Massive is different from the Contextual advertising model used on MSN. The current advertising model on Xbox is quite the stereotypical ads that are served on the virtual real estates elsewhere, for example, on Billboards on the roads(inside the games, that is..),etc. that is accepted to taking the games close to real.

Google and Adscape

Adscape is intended to help developers of even freely downloadable games make money by serving ads. This also includes dynamic placement and control of the ads that appear in the games. However, I feel Google has much deeper intentions in having acquired Adscape. As put on TechCrunch, Google might take forward the Adscape technology for its SketchUp software.

SketchUp is a layman’s tool to develop your own 3D design of homes and your neighborhood. This can be used in the virtual world developed on Google WareHouse. Simply put, WareHouse is a user-made world that can be otherwise be seen on Google Earth. With this acquisition Google has got hold of a technology which can be used to make even SketchUp a money making venture. But what is interesting to note is that for the first time, Google shall be venturing into advertising technology that is outside the conventional contextual-targetting. This is even more significant since unlike Massive, Adscape does not currently hold a client-base. This means that Google will have to take its advertisers’ network to the new medium. The problem however is that the new form of advertising on video games, which, most probably might not be contextual-targetting, might not find favor with the majority of its existing
advertisers.

Only a statement from Google explaining how it plans to proceed further in this particular segment shall clear doubts in minds of advertisers as well as speculators like me alike.

'Most Valuable Sites' for the Web 2.0 era

Web 2.0 has been around for more than two years now. Around June of 2006, two Israeli programmers got together to create a website in their sparetime that will carry the logos of all the new and upcoming web 2.0 sites. It has hardly been half a year since, yet the website now carries more than 833 logos. That in itself explains the proliferation in the number of websites that have been propping up in the post-dotcom bust period. Google has emerged as one of the, if not the most coveted website from the previous century. Now, with such a huge number of websites that have come up in the short time that web 2.0 has come to be, it makes sense to read these new websites that have come up and guesstimate which among these will be the ‘newsmakers’ amongst this new bunch.

Looking from another perspective, we need to see why only one or two of these websites should actually become a ‘newsmaker’. It does not take much time to reason out that sites that can actually find a solution to the existing chaos will emerge the winner eventually. That was the very same reason why Google became hot property at the turn of the century.The new medium of web was churning out so many websites that it became virtually impossible to look for one particular data from the millions of pages. Google solved that very problem.

Problem of web 2.0

Similar to what we saw for Google, if only we can find the problem that we might face with this huge onslaught of websites, picking the most valuable website among the new lot should not be a difficult task. Now, here is the situation: 833 websites that’s been on Go2Web20 alone. That makes the actual number of websites much higher than this. Now, zeroing on the website shall not be a difficult task, thanks to Google. But managing them can be. Imagine this: I love reading, making friends, surfing news, watching TV, etc. Now, there is a website for me to do each of this. For instance, there is Shelfari for the avid reader, Digg for the news buff, MySpace for networking, and YouTube for videos. Now, the more the number of sites, the number of places I frequent increase too. Here comes the problem: I not only have to remember all these numerous sites, but also have to remember my login information for all these websites.

Having one login for all the website might not be sensible for some, and del.icio.us might in itself get cluttered when you try to bookmark websites and individual pages alike. For that very reason, I thought it would be really important for us to know why the following two websites shall be considered the most valuable ones in the long term.

NetVibes

NetVibes is a website that lets you add all your bookmarked websites on one platform. So, in other words, if you visit Digg, orkut, youtube, flickr and a dozen other sites, you can add link to all of them here. This apart, you may also add RSS feeds, perform image, and blog searches, check mail; basically do everything that you visit the internet for. This is not the only place you may do this. Even Google offers its own personalized home page feature, but as with any website, NetVibes has this function as its USP and that should help it in the long run in branding. There is one very critical aspect that could be to NetVibes financial favors. Since the very purpose of NetVibes is to serve as a platform for you to jump to another website, it naturally serves to be the default webpage for most users on their browsers. The enormous number of visits that this shall generate makes it one of the most prolific earners on the internet.

OpenID

This is a not-for-profit website that is aimed at achieving a mission of ‘One User Id for all’. Though this is in the initial stages, once OpenID is made available for all major websites, unique login ids for every website may become a thing of the past. Interestingly, though I feel that this is one of the most valuable websites in the coming days, it does not feature in the 833 sites listed, probably because it does not strictly fall into the category of Web 2.0

Like you see, these two websites, when combined offer you a strength that is most desperately needed in the age of web 2.0 when dozen sites prop up every other day. Though these websites might not achieve the greatness that Google achieved in its time, from a user point of view, a combo of these two sites is the best thing to ask for.

Yahoo makes a BrickHouse opposite Google's Labs!

This week, Yahoo announced the launch of its new super-product called Pipes. Pipes is not just a great product in terms of innovation, but it is also that after a long time, it is actually one hyped product from Yahoo which has been built in-house. Most of the other Yahoo products, including Flickr, MyBlogLog, etc were acquired from startups.

Before, I move further, I will take a few lines to explain what Pipes does. Yahoo Pipes basically lets you, a user, create a mash-up of your own. Taking one example from the site itself, you can create a mash-up that lets the user know the latest products in a certain price range, that is out on eBay. This is created using the appropriate APIs available from the eBay website. Other easier examples will be to make a webpage that contains the latest articles on say, “Internet Business Strategies” from various sources of the internet. You basically use the feeds available from the different news sources and simply enter them on the Pipes creation page to get this done. You may go through the nice tutorial available on the Yahoo Pipes page.

Pipes is not the first time somebody has created such a product. There is already atleast one another product that serves a similar purpose. The website, called Ning is already a huge hit among web users. So, the question we need to ask is why did Yahoo choose to create Pipes when it could have acquired Ning, as it did with Flickr or MyBlogLog. The answer to this lies much more deep inside the company’s strategies.

It might not to be wrong to link the in-house development of Pipes with Yahoo’s competition with Google. Google, right from its inception days has been quite an attractive destination for innovators. As a result of their 20 percent rule; which allows the in-house programmers let one day of their work for their own personal projects, they have fostered an environment of innovation. Not only that, this has also helped Google Labs come up with innovative products every now and then, including products like Google News, Orkut, etc.

But the question still remains if the question of paying more to buy a product than developing it inhouse is the only reason for Yahoo to launch Pipes. Interestingly, it is not the only reason. A much bigger stake exists in the form of employee retention which has kept Yahoo more worried than Google, since Google has earned the reputation of being a great place to work in for innovators. This(focus on innovation) is one step which will not only help retain employees, but also serve Yahoo to project as a destination for innovation.

Incidentally, Pipes is just the first in the series of many more innovations to come. Yahoo is yet to formally launch BrickHouse, which is going to be the innovation center, and is quite similar to what Google Labs is for Google. The focus of BrickHouse is going to be on innovation and new creative products to the user.Though it is only going to launch in March, Yahoo BrickHouse is already working on several employee product initiatives. For example, one Yahoo employee’s ground work on a project that will help gather a web visitor’s ‘fingerprint’ in the form of his image or profile when he visited a web page is already underway. This is a great tool for publishers since they can now develop websites that can be closely customized to the visitor profile. It already looks to be one of the in-things in the months to come.

Yahoo's Panama model – For good or bad?

In a significant move in the advertising segment of the Search Engine industry, Yahoo has announced that it is now completely doing away with the bid-to-position model of ads to give way for a more democratic “Marketplace Design” model, more similar to Google Adwords. It is a mixed signal of things to come, and as I shall opine here might require that Yahoo now starts marketing its Search Engine better if they should continue to sustain their model.

What’s the difference:

In the bid-to-position model, the ads alongside search results are based on an auction system where the advertiser who bids more for the spot gets to bag the spot. In effect, it did not consider the relevance of the search keyword to the ads too much. This model was based on the presumption that though this system shall fetch lesser clicks on ads than the Google Adwords model, the revenue generated shall be higher considering that Yahoo generates more money per click.

Where’s Yahoo missing out:

There are a couple of things that I feel Yahoo is missing out. First is the advertiser competitiveness. If you look at the Search Engine market share, Google has a huge market share compared to any of the other search engines. For Yahoo, their business model made sense because they focussed more on the bid amount (and not on the keyword relevance which is important for hige click-through-rates), so that they could discount the fact that
Google has a higher market share.

Now with the Panama model, things shall not remain as it is now. The model being closer to Google Adwords, we can safely assume that the ad Click-thru-rates are similar to what exists for Google (that’s roughly 1-2%). Now, that being the same, Yahoo lags behind Google on two fronts: (1)The number of searches made, and (2) The amount bid for a click. Owing to the fact that Yahoo advertisers are catering to a smaller audience than those for Google, this would mean that the newer model has pretty much sealed the leader between Google and Yahoo with respect to the revenue garnered from the search engine segment.

There is one more aspect to it. Yahoo’s Panama model, in my opinion has one basic flaw. I could explain it better with the following quote taken from the source.

Historical clickthrough rates (CTRs) are one part of how ad quality scores are determined. To get this information, Yahoo will pull data (relative to other ads displayed at the same time) from both the old system and the new Panama system. The new ranking algorithm emphasizes data “freshness” and will use the most current information available.

One parameter of the Panama model is the historical clickthrough rate. What is surprising is that Yahoo shall weigh the ad quality based on the CTRs that the particular ad had garnered through a combination of both the old and new system. The question is why should it take into account the old system? As we know, the old system had one ad unfairly placed over the other owing to the bid amount. This meant that those ads at the top of the table in the old system would have had an unfair advantage of being clicked more than the ones at the bottom. That means their CTRs were higher. If you take this CTR as one of the parameter in the new system, then it would mean that those advertisers who had bid more in the old system will still gain an advantage in the new system. This, in effect would cause their CTR in the new model also to be
higher, thus biasing the whole model. It is simply a cascading effect in place.

What it means for the advertisers:

Advertisers should in general be happy about the change. Because now it is more about keyword relevance and less about money power. They can now expect many more clicks on their ads; much more targetted ones at that, which in general should help them in increasing the conversion ratio of customers over mere curious visitors. But then, with minor flaws as I have already mentioned regarding the historical CTR parameter, the model would still serve the top bidders of the previous model extend their lead.

But one persisting question is why should advertisers remain with Yahoo in the first place. In the previous model one can expect advertisers with not so relevant products, but still enough money to get a return through bidding higher to be interested in the program. But now with Yahoo’s model resembling Adwords more or less, why should advertisers come to Yahoo? With a larger search base, they can expect to get many more customers from an ad on Google than on Yahoo. It is still a debatable question, which can only answered over a period of time.

Offline Businesses Online: Comparison of different Business models

It makes a lot of sense for small businesses to enter online. The extremely cheap setup cost, coupled with a wordwide audience and a 24X7 open shop makes it an attractive proposition. There has hence been a spurt in the number of online stores ever since the WordWideWeb started.

But the business model in which such websites operate makes a great deal of difference. I shall take up three examples to show how effective or not effective these different models can be.

FridgeDoor:

This is a model that most businesses follow. Fridgedoor is basically an online shop for fridge magnets. The model is quite simple. Display the catalog online with the price, allow a secure transaction gateway (like Paypal) and sell. You may as well use affiliates to sell the product for you, whereby you pay commissions to those who sell them for you. You may use CommissionJunction for such purposes.

Though the model looks quite robust, it can go wrong in its implementation. It pushes you from your core competency into territories that you don’t know. For instance, though FridgeDoor’s core competency is meeting the supply-demand gap of FridgeDoor magnets, this model requires a host of other things, like increasing the page popularity using SEO techniques, working out a proper affiliate commission model so that the commission paid is neither too less nor too high. It is not that the other business models I shall discuss, do not require them. They require them too. But, the return on investment is higher there. Also, FridgeDoor is a hit because it was one of the firsts in its niche, so that it has been able to retain its leadership ever since. However, a company starting now may find it difficult simply because there are too many existing players in every possible niche that gaining customers through this model may be a tough ask.

CafePress:

CafePress is a unique example which operates outside the conventional model. “If you have to get over the existing big players, get new players in the fray“. CafePress is a website where you can shop for tshirts, caps, mugs, and other gift items. You choose the item, and they make them and ship it to you for a fee. This is very similar to FridgeDoor. But their way to increase business is through getting people with good design but lack of resources (read tshirt printing expertise) to get into the business by simply making the design for the tshirt or cap and promoting them on their website. This way, both cafepress and the designer make money for what they have done.

So, where does the company benefit? Here, CafePress has not spent on areas outside their expertise like SEO. They have not spent on new design research, which the seller base has taken care of. They only fufil the orders, and money spent is on promotional activities, which is an inevitable need for any firm. This besides, CafePress has also taken to the affiliate mode of business (which has the roadblocks that I mentioned earlier). But the fact that the website has already garnered a user base that is ready to design the products makes the publicity effort much more easier.

Moo:

This is the most exciting business model that is exploiting the environment of web 2.0. Moo is again in the printing business, and they print cards. Their business model is to use the current online social networking scene to gain a customer base that would have never used their products otherwise. Moo calls for users to take their online relationship offline, and to do that prints ‘minicards’ for distribution. Now you can use these cards to provide your Skype or simply your email address to your contacts as an alternative to business cards.

To find the customer base, Moo has taken to Social networking websites like Flickr and Bebo. Since these websites provide an exponential referral list, targetting the web users in this model would mean that if the website can get one user to print his Flickr photo on the card, then his contact list could be the next target for a Moo purchase. This model has so far been extremely effective, since the ‘modus operandi’ is fresh and interesting.

A mixed model?

Now, the purpose of me discussing the three is not to judge the best model. This is because, not all kind of products can be marketed using one of the above models. But, there is definitely a lot of potential in a mixed model. For example, FridgeDoor can print Flickr photos on their fridge magnets, and hence a tieup with Flickr shall not be a bad idea. Fridge Magnets are a great gift-material and hence the company could try out some newer models to increase customer base, rather than trying the conventional model. No matter FridgeDoor has been an extremely popular site, but, gaining newer customers this way is not unwanted either.

Though Moo’s model is exciting, it is something not all businesses can copy, simply because not everyone prints business cards. However a brainstorming can lead to many other similar exciting business models. But, a little competition to Moo’s model is not bad after all.

Next Generation Image Search technology

What is the next generation image search going to be like? StockPhotoTalk, in an article on a similar topic hints at image search engines like YotoPhoto and Pixsy that help provide copyleft images and images from those places on the internet which are usually not spidered by the traditional search engines from Google and Yahoo. But is this the next generation we are talking about?

In my opinion, Next generation Image search engines mean much more than that. Consider this: You are taking part in an online quiz competition and you are given a picture of a celebrity to be identified. Now, here is an ideal situation that is not addressed by the current search engines, where images are displayed only for keyowords that you provide. In other words, the above situation will require a search engine that addresses a situation opposite to the current scenario, where all you need is to right-click on the image, copy the image location on the website server, and paste it in the search bar of the next-gen image search engine. The search engine should then be able to compare this image with those in its index and provide relevant images or text from the web.

How to go about doing this:

Now, the question remains how you can go about creating such a search engine. The answer lies in a software whose algorithm is capable of identifying the varying hues and shades that appear on a photo, compare it with the hues and shades that appear on the other millions of photos that it has indexed from across the web, and provide those that match with the queried image. Moving ahead, the software should also be able to match texts that appear predominantly in the texts accompanying each of the web pages containing the matching image, and provide a text based result. For example, if I query an image of Mahatma Gandhi, the search engine should be able to deliver text results of the Indian freedom movement.

Present Technology:

Though the ideas presented seem kind of far fetched, it is not too far away as well. Current researches on this field prove so. For example, sometime back, I had blogged about MyHeritage, a ‘Face recognition Software’ that helps you find out which are the celebrities you resemble most with. This software works very similar to the algorithm I have talked about in this page. Whereas MyHeritage is a software ‘trained’ only to identify human faces, more extensive research on the same technology can help us move more closely to the next generation image search that we are talking about.

And the task is already been taken up. The Ohio State University is already researching on such a technology. Called WISE (acronym for Web Image Search Engine), the search engine allows users to upload images from their desktop to the university server and then compare it with the existing images. This is the closest that we have yet been to the next generation of image search.

Roadblocks:

So, why are the present stalwarts in the field not taken this project yet. The problems, in my opinion, quite justify the same. For one, it is the time taken for a search. Unlike the Google Web search, you cannot expect the search to take place in those ‘0.12 seconds’. Checking color patterns through the millions of images takes much more time than a mere text search. And the next problem is that of monetization. Unlike the text search which reaps in millions to the Search Engine in the form of sponsored links, image search technology does not justify the costs. It costs much more to develop technology that will understand images to realize the underlying text and then provide the most appropriate ad, than the return on investment that Google or any other search engine can expect to make from it.

Future indications:

The best bet is now on the WISE technology to take shape. For now, this search engine indexes only around 112,382 images, which makes most of the results provided irrelevant to say the least. But, the technology will only be proved when the indexed image list grows, which shall happen sooner or later. Another possible direction is an acquisition of the MyHeritage technology by the existing bigshots in the industry. For one, MyHeritage is not just about Face Recognition Technology. It is more about building family network online, which is yet another form of social networking on the web. This model will align with existing business models that Google has been operating in, as in Orkut(which is one of the top social networking sites). Hence, an acquisition of the MyHeritage technology by Google shall prove to be a phenomenal shift in search technology.