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.

28 Jan 07 | Business Model Analysis |

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.

22 Jan 07 | Business Model Analysis |