NBC and YouTube.com Team-up

NBC has agreed to partner up with YouTube.com to reach people who might not watch much television in the summer. The deal would consist of a channel dedicated to the broadcaster that would show clips of new episodes, old episodes and behind-the-scenes footage.
This isn’t the first of the potential wave of television to Internet cross promotion. Warner Bros. has begun selling specific films and television shows on Guba.com, and is in talks with Bittorrent as well (selling about 200 films).

Source: NetRatings 2006

Pictures and the Digital Camera ‘Brandscape’

(by Lucas Roze & Andreas Schiffler, IC-Agency Canada)

Introduction

In recent years, the art of loading film in a 35mm camera has become an archaic practice. With sophisticated high-tech photographic devices, USB or Firewire cables, heaps of flash cards, and a docking stations, the processing studio of today, for most, is their plug and play laptops. Many of the images end up on the Internet. But with all these pictures floating around, what can we do from a statistical perspective to measure brand interest? This article by L. Roze and A. Schiffler of IC-Agency Canada will shed some color into the greyscale world of online picture publishing with a unique analysis about which brands are creating all the images we find on services such as Google or Flickr.com.

Analysis on Technology

While researching on wikipedia.org, we ran across a neat table beside one of the images listing technical details directly related to the image. What we saw was an automatically generated dump of Exchangeable Image File Format (EXIF) data. One of the many data fields contained within the EXIF meta-data is a text referencing the manufacturer of the camera that took the image. This was spotted by Marketing Specialist Lucas Roze and immediately caught his attention as a valuable resource for marketing analysis. It didn’t take long for ICA Senior Software Architect Andreas Schiffler to create a sophisticated spider retrieval program, engineered to pull thousands of pictures from the Internet for web analytical purposes. Combining the spider with some EXIF text analysis, allowed us to retrieve some interesting meta-data information from thousands of pictures and understand their origin. By distributing image sources over a long period of time (Flickr) or many independent keywords (Google), statistical accuracy is maintained in the resulting data set. The data collected in this manner can then be analyzed by what type of brand and model was used, the date and time it was taken, and many other interesting specifications directly related to the reference of the device that took the image or the source of the image file.

Brand Usage

Our unbiased survey spider ran over several days and thousands of images were downloaded and analyzed. Of the images retrieved, over 2000 pictures had usable EXIF meta-data. Using keyword analysis and binning, we were able to establish the top camera brands that are currently in use for Internet publishing.
 

Displayed on the graph above, we see that Canon (40%) is the most popular brand by a long shot. Nikon (20%) and Sony (12%) take comfortable second and third places. Olympus (8%) is still widely used. The rest of the brands are distributed through the remaining 1/5th of the usage. A definite usage looser seems to be big-name Kodak (3%).

We assume online image-posting behaviour by Internet savvy individuals is similar across brands and does not create a bias towards one or the other brand.

Buying Intentions

After establishing the use of digital camera brands, the next step is to compare it to user intentions. Shown in Figure 1.2 are the top 14 brands by intention as compiled from a chart published by Digital Photography Reviews (DPR, January 2005). The DPR table listed purchasing intentions by brand based on site-click measurements.

As we can see, brands have a clear winner by purchasing intentions with Canon (28.5%) models, which attract more clicks than its two nearest competitors Nikon (13.9%) and Sony (11.5%) combined. The following 7 competitors lead by Olympus (8.6%), Panasonic (7.2%), and Fuji (7.0%) focus most of the remaining interest, leaving a few percent for the smaller brands in this industry.

Canon and Nikon are the only two that demonstrate higher usage of the camera, compared to the online interest of users. As such they are the established leaders in the field and simply indicates that many people own the camera, but few are interested in upgrading or making a new purchase. The middle field with Sony, Olympus and Fujitsu seems to be bang-on in correlation with the usage and interest. Kodak, however, and brands on the other end of the graph (Pentax and Panasonic) lies the most interesting component of our comparison. Panasonic, while experiencing a very low usage percentage (1%), shows an interests in the camera brand that is 7 times higher (7.20%). This means that people are very interested in the Panasonic name brand, but are not making the purchase. A similar but less pronounced (50% lower) “gap” between usage and interest can be observed with Pentax. Even Kodak seems to still ride some a smaller brand recognition wave, but cannot translate it into actual use.

How can this gap be explained? While there could be a technical answer - Panasonic might hide more meta-data in its EXIF images – we don’t feel that this is the case, as it would go against common technology practises and EXIF standards. Thus, especially Panasonic, but also Pentax and Kodak, should re-visit their marketing strategies.

Summary

The fairly recent life of the digital camera era seems to be one full of marketing potential. With the meta-data stored in the pictures, it’s a marketers dream. However, there are some road blocks like the use of Photoshop, screenshots, and other editing software that erases the data. Furthermore, it is hard to determine the brands that do not keep the data or stores it under different contexts, and is even harder to pin point the specific models.

A good example is Kodak. After sitting pretty in the driver seat of the paper processing era for many years, the big cheese has yet to excel in the production and marketing of their products compared to its direct competition. Furthermore, with Kodak’s higher prices, this plays a trivial role in the downward effect that we see.

The correlation between the interest and usage by brand allows us to clearly compare apples with apples (the camera brands) and yields some very interesting results. The graphs are a measuring device to compare intentions with demand, something impossible to measure in the past with the film stock models.

Using key performance indicators with this information could prove vital in the hands of digital camera marketing departments that want to establish a competitive advantage over the competition. Trends might be spotted, and the secondary industries such as digital-image processing and tourism services could use this information to optimally place valuable advertising space within their businesses.

Currently, digital cameras are relatively cheap to produce and are being mounted on everything from cell phones to Barbie dolls. The trend will continue and the digital revolution in the camera world is not about to slowdown any time soon. Furthermore, with the Internet simplifying picture exchange with sites such as flikr.com, photo.net, and picasa from Google to name a few, online publication will only get more common and more accessible. These factors combined, presented in this analysis, prove to be a very valuable tool now, and in the future.

Cross Marketing for Viewers Made Easy

(by L. Roze, co-author A. Schiffler)

As we all know, marketing has evolved from the late trials of billboard ads to banner ads on Internet sites. It is speculated that a typical person is exposed, on average, to 3’000 different types of advertising each and everyday. Marketing gurus understand that a person is desensitized to the everyday bombardment of advertising, and thus acts accordingly to innovate and grab their attention. Most of the time, we are unaware and pass them off as nothing new, but we do have times of focus, and we do pay attention to certain things, right? Well, this is the prime real estate that marketers are always trying to reach. Today, the focus isn’t on the commercials in the previews before the movie, but the product placements during the show. The ultimate test is to find that kind of opportunity which can provide added value to the viewing audience by not alienating them from the action of viewing.

I’m not what you would call a TV addict, but during the long winter months in Canada, I do my fare share of viewing. And I, unlike most, pay close attention to the commercials and the product placements in network slots and in movies. After some digging around, I discovered this industry has some ingenious marketing ploys brought forth by the top dogs in the US media industry, which I’ve coined “Slot Shopping” (patent pending). Slot Shopping is defined as the art of making products available during or after the act of viewing: the ability to view and purchase items seen on the episode or the movie! What a concept! However, currently there are some major flaws with what they are doing, and I plan to suggest a solution that could change the way we watch AND shop.

Case Study: Fox
The Fox Broadcasting Company, established by Rupert Murdoch in 1986, has launched many actors into stardom through its television series and movies. With the current pop culture focused on consumption and collection, shows like The Simpsons and movie such as Mr. & Mrs. Smith generate lots of buzz in these sectors. So, Fox offers some merchandise on their site, like DVD’s and other paraphernalia, but it fails to take it that one step further for the extra added value.

Problem: Fox
The products in the shop are very centric to the shows and the characters. This is understandable for shows and movies that feature animation, but for the others, they seem to have missed the mark. The focus of the merchandise is strong branding it with logos of the show, characters faces and expressions, or the movie title, but the organisation and variety of products is lacklustre.

Case Study: NBC
Known for its infamous peacock logo, NBC launched its television debut in 1941. The popularity in its shows has created much attention in the media and on the street. Take for example the controversial sitcom Will & Grace, one of the hottest shows. This situation comedy is broadcasted in over 32 different countries and deals with mature content. Revered for its style On-Air, the NBC understood this and responded to viewer request by setting up a shop on their site to sell articles of clothing and home accessories seen in specific episodes. You have the option to hand pick the episode, location, or character to see the featured products that was showcased: Will & Grace returns 190 products ranging from espresso machines to boots. The Tonight Show with Jay Leno is another popular show that invites many rising stars: 237 products most of which are music CDs for performing guests. They too have the opportunity to push sales through the NBC.com site concurrently with the televised appearance. What better way to watch and shop from for home, this is a revolution in home shopping. Or is it really?

Problem: NBC
The site is well made, however, it lacks a connection from the show to the site. This breakdown in communication likely leads to low sales and loss of potential revenue. The solution lies with assimilation of the show through technology such as streaming media directly on the screen.

So is Slot Shopping a dud?

With current implementation, the major problem in most cases is the lack of a recognizable connection between the shows or movies and the websites that push the products. This is understandable, because TV and the Internet are two different media types and integration is hard. Very hard! The answer might be with the current developments of Internet technology itself: P2P, streaming, on-demand video. As we know, the Internet is taking a stronghold on the viewing audience. People like a more interactive experience that allows them to be in complete control. What better way to give them this freedom of watching an episode again by offering a free streaming video service over the Internet coupled with the ability to acquire products or articles seen on their screen? What better way to shop for products sprinkled with endorsements by your favourite TV actors in the form of short clips?

The art behind this is an organized timeline that would coordinate clusters of products specifically timed during key events of a movie or show. This data, coupled with the proper presentation is a winner. A constant stream of products would be distracting and annoying, furthermore, not all movies and shows could support this technology (e.g. imagine the shopping experience during Ridley Scott’s “Alien”). But a sensible and highly interactive approach, using dynamic content technologies such as AJAX would enable the potential consumers to link themselves to the product directly while continuing to enjoy the stream, thus stimulating an impulse purchase made easy. The potential of being a click away could generate a huge potential in traffic and in sales. Clips of highlights of the show could also be offered together with the product, to demonstrate that the article or item was in fact worn by that star. Sort of a virtual value-add and endorsement of the product. Moreover, the web analytics potential could answer some very important questions regarding customer behaviour while viewing. These types of study could provide unheard of competitive advantages with the use of leading edge data mining tools.

Think about it, how cool would it have been to be able to get the exact same leather jacket and glasses as Neo in The Matrix, or order that bikini worn by Hally Berry in the James Bond movie Die Another Day – and get these show as free stream On-Demand from the Internet This stylistic dream could soon become a reality – maybe not with blockbusters like The Matrix – but certainly with more traditional network fare. There is definitely a huge potential for product sales when making television shows and movies available over the Internet.

Power and the Internet

(by A. Schiffler)

It is amazing how easy it is to forget and ignore a simple fact: the Internet with all its "free" communication and information  is a big energy wasting, power sucking HOG of a construction. Once you've read the numbers below, there can be no doubt why there is a digital divide and Africans don't need a donation of our old PCs: only the affluent can afford to "plug-in" and operate that kind of juice-sucking machinery in the first place.

Take Google for a start: Google operates one of the largest computer-clusters on the planet, to provide us with a sub-second search result (and the ads that go with it). They run upwards of 250K servers, collectively consuming a staggering 20 Megawatts of power for a nice electricity bill: 175 Gigawatt-hours per year - almost a Million Dollars a month. And that's just their server farm, never mind the offices and equipment that connects it to the rest of the world.
-- http://en.wikipedia.org/wiki/Google_platform

At the time when "Energy-Star" labels went onto most Dell PCs in 1999, the experts were discussing the total power consumption of office equipment and network infrastructure. An estimate of 74 TeraWatt-hours (TWh) per year is estimated up. The Internet barely makes a blip in the total (copiers and laser-printers are just way better "consumers" than modems), with telecommunications equipment taking about 5-10% of this total or about 7Twh/y ... note that this is in 1999 and for the US only.
-- http://www.lbl.gov/Science-Articles/Archive/net-energy-studies.html

A few year later in 2002-2003, the reports were refined and give a number of scenarios with interesting sounding names like "Zaibatsu", "Cybertopia" and "Net Insecurity" which solidify a new conservative estimate to around 3.5% of TOTAL power consumption.
-- http://www.rand.org/pubs/monograph_reports/MR1617/MR1617.sum.pdf

Today usage is further up (as usage grows), so we can assume a conservative level of 4% of total power consumption, factor in the 4TW of US usage in 2006 and get a nice amount of 1401 TWh/year (including all the office equipment again). Thus our 5% Internet portion, is now at a whopping 70 TWh/year - perfectly in line with the 10-fold grows in hostnames from 1999 to today as reported by Netcraft.
-- http://news.netcraft.com/archives/web_server_survey.html

Will the trend continue. Sure it will! With more always-on-devices and entertainment platforms like the PS3 that would be considered a supercomputer just 10 years ago, we will definitely continue to use more power for our IT needs. Even though companies such as Apple like to put a positive spin on it -- http://www.apple.com/environment/energyefficiency/ -- the fact is, that current hardware (CPUs, video cards and even networking equipment) will use more power. And new operating systems like Microsoft's Vista, which will require a 3D accelerator card (the second larges power consumer in a PC) - a software move that will push for more transistors running at higher speeds on Millions of desks. Thermal design is really the primary limiting factor in microchip design today and current processors burn up as much as 100W of power when in use.
-- http://www.tomshardware.com/2005/11/21/the_mother_of_all_cpu_charts_2005/page4.html

But hey, I forgot one more thing: The production costs for all the equipment that runs the show! Its a bit like the hybrid-car-connundrum: Forget hybrids, but give me a car that lasts 5 more years - that's green. Because on a whole, the longer lasting car will probably save more energy than driving one of the latest battery-powered gizmos would - because a lot of energy is spend on making the car in the first place. The same applies to the tech equipment: Five new computers over a period of 10 years (28GJ) is about two-thirds of a car (47GJ) in terms of energy consumption for production.
-- http://www.newsfactor.com/perl/story/19992.html

So this leaves my environmental conscience - as tech worker, full-time programmer, dare I say internet-addict - with quite a bad feeling about the whole thing and one can only hope that technological advances will turn the trend around in the near future. As an individual, I think one can try to do more with less,  keep the old stuff and live with simpler cooler-running technology as long as possible. But in the end it leaves me still a searching for practical answers ...

Update (2-Jun-06):

Some related links -  Scott McNealy's eco-friendly challenge - ecoComputing Update Center

Update (14-Jun-06)

NYT Article - Hiding in Plain Sight, Google Seeks More Power ... its more like 450K Servers for Google soon. Heck, they build besides a river like a power plant. Also, Microsoft seems to be aiming to be the king of the hill with 800K server by 2011!