I was musing again today on the cost of distribution having shopped for a replacement cell phone battery. I found it both online and at the local Best Buy electrical store, but the price disparity was surprising. Online video camera and cell phone batteries quote a list price of $35 to $50 online but are discounted online down to $2.50 to $9.50. This is to accommodate the higher price that local stores want to charge as the recommended price.
Best Buy were of course charging $38 for the same battery that I just ordered one from Amazon for $1.38 and which will be delivered for below $4.50. We can fool ourselves with all the justifications used to explain the premium to us. Brand, Service etc. but in reality it was the availability and convenience at the end of the street that provided justification and the cost of running and stocking that store that needed to be paid for. Manhattan rent and staff to find things for us, as oppose to the online search and automated pick and ship warehouse in a cheap location with cheaper staff.
Twenty years ago I designed under contract a simple single chip circuit which a games vendor sold as a plugin component to a games console. The electronics trader in Hong Kong charged 50c per unit. The Chinese manufacturer probably received around 30c. The games vendor sold the units packaged at around $10 to a US distributor who sold them to Game Stop at below $20 who retailed them at $39.95 to the consumer. One could consider it to be a markup of ridiculous proportions but in reality it is the distribution and productization (Branding, selection, packaging, marketing, advertising, placement) that is being paid for. I did make a mental note to myself to try to be the middle man next time around! Such a next time alas didn’t occur.
In rearranging my apartment recently I found myself selling books, DVDs and furniture that I had purchased at retail, or auction in the case of the furniture. The reality is that the value of these possessions when redistributed by me is negligible. Even worse the furniture has a negative value if I have to pay for it to be stored or removed as identifying a ready buyer is no mean trick even in a world of online markets. I have given up on letting Craigslist people visit the doorman.
All of these experiences remind me yet again that so much of the value is about where something is and identifying somebody who wants it. It is one thing for the shopper to be able to find the object on Amazon but quite another for the object to find a buyer. That is the side of the equation that most of the sales, advertising, brokers and agents work on. This is also the area that GroupOn attacks — being partly marketing to new customers and partly a method of clearing perishable supply. There is still plenty of money to be made in improving this life-cycle. Distribution to the right buyer, not the product itself, is what we are really buying. The distribution and not the product.
Newspaper readership is continuing to decline. We continue to read messages directly from companies and governments while still getting news of incidents from a traditional news source for free online. Layoffs at newspapers continue a pace.
The amount of non paid add space is noticably declining in a number of publications and even on television despite it being an election year.
Jeff Jarvis has posted a presentation on the current state of the news industry here.
The internet isn’t free. If you have a significant amount of traffic your hosting provider will start charging you or constrain your account in some way. When it comes to video content most amateur producers are using the likes of YouTube or BlipTV to distribute their videos. They are getting approximately half the revenue where there is any. The adds are typically a single placement before or after (‘preroll’ and ‘postroll’) or small adds popping up during the content (‘overlay’). The problem is that the infrastructure has to be paid for. In the case of YouTube it is being subsidised by Google as well as its original venture capital like the other players. Major players such as TimeWarner are receiving revenue from preroll and postroll advertising and probably paying fairly directly for their infrastructure and production overhead which is a large part of the slow provision of these services.
What that infrastructure costs isn’t clear but as a guide we know that at the retail level traffic costs are as low as 25c/Gb. A typical recurring video might be anything from 0.1Gb to 1Gb depending on the duration and resolution. That gives a cost of 2.5c to 25c per viewer. Google has bought a lot of fiber and will be able to serve video at a significantly lower cost by providing their own network. Others will be using brightcove etc. in order to get economies of network scale at some unknown cost. The retail figures give a range of cost range per thousand viewers of $25 to $250 for traffic charges.
Television advertising typically realises $30 per thousand viewers (CPM) but can vary from $5 to $50 depending on the region, time of the week and program. Internet video advertisements typically realize less than one dollar but fall within the range of $0.25 /CPM and $5.00. Television programs in the US typically include 12 minutes per hour or approximately 8 advertisements of 30 seconds in twenty minutes of air time. So television might realise $240/CPM for twenty minutes. Internet video might realize for preroll,postroll and overlay as much as $1.50 to $6.00 /CPM across a much smaller audiance.
So infrastructure costs need to be significantly lower or somebody else needs to be paying them to allow the current advertising revenue model to be profitable. There are legitimate reasons why internet advertising realizes a much lower CPM due largely to the low barriers to entry and resulting low premium.
There is an assumption that advertising will support internet video because television is supported by advertising but this doesn’t necessarily follow. The constrained access to television made it a premium product and added credibility to the content and advertising. The economics of broadcast television, which is still what provides significant theoretical viewer numbers to television, was based on a cheap distribution method which could be paid for with commercials. This isn’t necessarily true of an infrastructure which requires physical installation to the consumer’s premises and servers which consume over 1% of the electricity in the US. The energy costs aren’t decisive. It is the infrastructure cost that needs to be reduced, both for the servers and the distribution, if advertising is to be the sole basis of the model.