Begun, the acquisition wars have.

Some recent news on acquisitions – I have seen more chatter & activity lately in our little slice of the advertising market than I ever have.

While our space is littered with point solutions, this is not Bubble 2.0.  The consolidation that is happening within the space will ultimately give marketers a more holistic view of what their online dollars are doing.  I can get behind that.

Clearspring is building a warchest, see here.

Efficient Frontier buys Context Optional, see here.

A small company, called MicroSoft also got active, see here.

Collective and Oggifinogo (not new), see here.

Enhanced by Zemanta

The Trading Desk Debate

With the advent of the TLA (Three Letter Acronym), I have seen usage of TLA’s in the online media space continue to accelerate (to all of our detriments, RMX, WMI, MNI, etc. excluded).  So it is with great reservation that I point you to the following article which discusses the successes (or lack thereof) of Agency Trading Desks, otherwise known as ATD’s.  Read it here.

Agency Trading Desks, for those not in the know, are corporate entities that have been created primarily at the agency Holding Company level – see the map above.  They have been designed to be a clearinghouse for all (or most) of a holding companies’ media needs (this is also referred to as demand).  Basically (and this was quite a prescient move), the holding companies took a look at where their dollars were being allocated, and saw that most players further down the ecosystem were taking a significant portion of it.

For example, 1 media buy COULD be allocated per the below:

  • Agency & Client set a $5 CPM for a Jaguar campaign
  • $.25 CPM ad serving fee for agency ad server ($4.75 left)
  • $3 CPM to ad network – media cost ($1.75 left)
  • $.25 CPM to pub-side ad server ($1.5 left)
  • $1.50 CPM paid to publisher ($0 left).

From a holding company lens, the split of where that $5 CPM goes is obviously not optimal.  Remember – the agency business has historically been built on razor thin 10-20% margins (estimate).  Did someone say Gross Margin lift?

Secondly – lots of dollars flow through a holding companies’ vast network of agencies.  And just like other business, network effects apply. Meaning – there are positive consequences to aggregating dollars through a single platform, including the buying power, data, etc. that comes as a result of aggregating all that spend through a single platform.  ATD’s can now throw their weight around with publishers and intermediaries based on their spending power alone (although I think the ecosystem has yet to see that take place en masse).

What does it all mean?

From what I understand, (and please correct in the comments where appropriate) the messaging around why these ATDs were created has been, up to this point, ‘spend with us because we think we can do it better than what is out there’.  Which may be true.  I am sure there are examples out there where an ATD-run campaign has vastly outperformed one run in the traditional way (which if anyone can share that data, even if anonymized, email me).

Here’s the thing – the support for that model (supply/demand scale, lots of buying power, etc.) comes on some level, funnily enough, from the ad networks.  Scale, both in supply and demand, leads to better decisions (which we in the ad network side of the business have lived through, many times over).  More specifically – the data that comes from a critical mass of supply and demand allows one (read: an ATD) to make better decisions on targeting, media mix, etc.  And I agree wholeheartedly.  And some ATD’s have seen success with that model – and I hope that more can do it.

The ATDs should be taking back some of that margin, pending they can spend a marketers dollars more efficiently (read: ROI or ROAS) than another vendor.    Reward good performance.  But I have found it helpful to consider that fact in evaluating this new TLA’d RTB’d Agency Trading Desk world.

And ATD/DSP/RTB folks: let’s chat if you’d like.

I, and my company, Undertone support all of the constituencies mentioned above – and we take great pride in that.  Further, the rise of the RTB ecosystem, and more specifically DSPs, has, on some level, torn down the old inefficient scaffolding of yore and is replacing it with a real valuation of a user, impression, a site, or an overall buy. And in that world, we all win.

Tune in next week for a deeper dive into the mechanics of RTB, or real-time bidding.

Disclosure: I am Director, Business Development for Undertone (mentioned above).

Cachebusting in Online Media – what is it?

This was an eye opening document for me, explaining how ‘cachebusting’ works in terms of online advertising, and why we have to do it.

Questions after reading?  Email or IM me – would love to chat.

Introduction

In an attempt to speed up user experience around the web, most browsers implement a technology called cache. This mechanism allows information, such as web pages, images, etc., to be saved on a user’s computer. Therefore, if the user calls for a previously requested web page, the browser will recall the information from the cache and not make another request from the site itself. Once the browser receives a DoubleClick ad image, it will store the image in the cache. Further calls for the ad image will be drawn from the cache – not from DoubleClick — and an impression will not be counted. In order to defect the browser cache mechanism, Cache busting (aka Random Number Generation (RNG)) is implemented.

Image Tags

For advertisers using DART for Advertisers (DFA), ad images and clickthrough URLs are stored on DoubleClick servers. These images and URLs can be called by image redirect tags that are implemented in ad slots on web sites. The image tags are sent from the web sites to users’ browsers in lieu of images and clickthrough URLs. When these image tags are processed by the browser, a request is made to the DoubleClick servers to retrieve the ad image and clickthrough URL, which are then returned to the browser. When an ad image is sent, DFA records an impression; when a click is processed, DFA records a click. If the ad image is retrieved from cache and not from the DoubleClick servers, the impression will not be recorded. This can happen if cache busting (RNG) is not utilized.

Cache Busting with a Random Number (RNG):

Browser caching can be defeated by dynamically creating unique tags for each ad image served. This ensures that when a user navigates from page to page the ad image is not delivered from the browser’s cache, but rather a new image is delivered from the DoubleClick servers for each request. To accomplish this unique tagging, a different random number is added to the image tags that reference the ad image. Typically, when a browser sees another image call with the same name, it pulls the ad from cache. To defeat this, an ord= attribute is added to the HTML tag, and when implemented correctly, prevents the images from being pulled from the cache. Tags with different “ord=” values prevent this caching problem because the different value forces the browser to retrieve a new banner from the ad server. If not implemented correctly, though, browser caching will not be defeated.

DART Tags and RNG

DART tags generally come in two forms: Standard Image Tags and IFrame/JavaScript tags. Both contain multiple instances of an attribute for cache busting: ord=[timestamp]?.  Here is an example of a Standard Image Tag:

<A HREF=”http://ad.doubleclick.net/jump/N409.somesite/B470;sz=468×60;ord=[timestamp]?”>

<IMG SRC=”http://ad.doubleclick.net/ad/N409.somesite/B470;sz=468×60;ord=[timestamp]?” BORDER=0 WIDTH=468 HEIGHT=60 ALT=”Click Here!”> Click Here!</A>

Random Number Example

Once the [timestamp] value has been replaced with generating code by the web site, the tags are sent to the user’s browser. Let’s assume the code generates the random number 86753091. The resulting HTML tags would look like this:

<A HREF=”http://ad.doubleclick.net/jump/N409.somesite/B470;sz=468×60;ord=86753091?”>

<IMG SRC=”http://ad.doubleclick.net/ad/N409.somesite/B470;sz=468×60;ord=86753091?” BORDER=0 WIDTH=468 HEIGHT=60 ALT=”Click Here!”> Click Here!</A>

Upon receiving these tags, the browser would retrieve an ad image from the DoubleClick servers, DoubleClick would count an impression, and the resulting ad image call would be stored in the cache. The next time these same tags were sent to the same browser, a DIFFERENT and RANDOM number should be generated. If the second set of tags sent to the browser generated the random number 13457890, the resulting tag would look like this:

<A HREF=”http://ad.doubleclick.net/jump/N409.somesite/B470;sz=468×60;ord=13457890?”>

<IMG SRC=”http://ad.doubleclick.net/ad/N409.somesite/B470;sz=468×60;ord=13457890?” BORDER=0 WIDTH=468 HEIGHT=60 ALT=”Click Here!”> Click Here!</A>

The browser will not attempt to retrieve the ad image with the ord=13457890? attribute from the cache because it interprets the tag as a reference to a different ad image. This causes the browser to call the DART ad servers; consequently, a second impression is counted.  If the same ord= value had been used for both sets of tags, the browser would have retrieved the second ad image from the cache and the second request would not have been recorded by DART.  If the ord= is implemented correctly in image tags, DoubleClick will count an impression for each request from a browser. If the cache busting string (RNG) is NOT implemented in image tags, DoubleClick will undercount the number of ad impressions as reported in the Advertiser’s/Client’s view.

Also – if anyone knows how this document originated, please let me know so that I can link and credit.

Regarding Frames and Framing. In 2009.

There was quite an uproar when Digg released a new feature called the DiggBar.  The basic premise is that the DiggBar would allow a Digg user to rate articles as they were surfing the net, through a process called “framing”.  Framing is basically a process wherein an originating site (Digg.com in this case) shows another site (i.e. joshkaner.com) in a “window” in it’s own site.

The DiggBar would be applied to all site visits originating from a Digg “short URL” – (see bit.ly, or tinyurl.com). These short URL’s have become popular, primarily because of Twitter (which has a 140 character limit on posts).

Back to the story.

Here is an example for joshkaner.com (click the image to expand to a larger version):

diggbar-example

Notice something very important: the URL (the stuff after “http://” in the address bar).  It shows digg.com.  But the site is joshkaner.com.  How is that happening you ask?  FRAMES.

They are an ugly product of a by-gone internet era (see:  Geocities, etc.).  And they serve no purpose other than to obscure.  One of the implications of the DiggBar is that now all the traffic that accesses joshkaner.com through the DiggBar gets counted towards Digg (not joshkaner.com as it should).  This effects site traffic, visitors, ratings services (like quantcast.com) and advertising that I place on the site.  Not cool of Digg.  For more heated arguments against the DiggBar, check out Gruber’s search results page for “DiggBar”.  He was clearly not happy.  Lot’s of people werent.  So Digg changed how the DiggBar worked.  It gets a little complicated, with shortened URL’s and the like, but suffice it to say Digg fixed it.

Now, to my point.  I was cruising through Facebook today, and came across an article I wanted to read that was posted by a friend of mine.  So I clicked the link.  And look what I see:

fb-bar-example

Look familiar?  It is.  It looks like the same old DiggBar.  Now I could be totally wrong, but how is this different than the DiggBar?  If you have any idea, or want to prove me totally wrong, please do sound off in the comments.

Amazon buys Stanza iPhone App

Image representing Amazon EC2 as depicted in C...
Image via CrunchBase

From my comment on Brad Stone‘s NYTimes article here:

I was talking about this with some other iPhone owners – see comments here:

Also – as I argued there – even though Amazon’s book reader app was lacking in terms of features and customization, in won a spot on my home screen due to the content. It always has and always will be about the content.

The UI and app is important, but means nothing without the content. Good strategic acquisition on Amazon’s part, and hopefully they will integrate Kindle books into Stanza’s front end.

The one thing that is missing (and I don’t believe they are allowed to do this under OS 2.0 Ts and Cs but WILL be allowed to under OS 3.0) is the ability to purchase books in the actual app.

Look for it this summer in iPhone OS 3.0

I hope that there is integration quickly between the 2 apps, and all associated content stores.  Because Amazon has the content, Stanza has the app and the UI, and we the custuomer can benefit greatly.

Web Ego Map

Interesting – this ‘map’ came out last week, and apparently is done every year. Interestingly, very few people will know most of the connections on here. But the visualization of web influencers vis-a-vis a subway map is kinda crazy.

 
 

via Gawker: Valleywag by Ryan Tate on 4/11/09


Sure, the fourth “Web Trend Map” from branding firm Information Architects is a nifty piece of graphic design. But that’s not what makes it viral.

It’s the human impulse to try and find one’s brand, or one’s employer’s, among the “most influential” that turns the bastardized Tokyo subway map into self-promulgating piece of marketing.

The map is a reminder — as if we needed one — that, these days, you don't need to be a print magazine publisher to use an arbitrary ranking system to get people to look at your content.

Speaking of which: You can find the full-sized map here.

[via Daring Fireball]

Posted via email from JoshKaner.com

Boxee Update

boxee_logo

For anyone who has not heard of Boxee and you own either a Mac or an Apple TV, I highly recommend you download and install it.  Here is how Boxee puts it:

on a laptop or connected to an HDTV, boxee gives you a true entertainment experience to enjoy your movies, TV shows, music and photos, as well as streaming content from websites like Netflix, CBS,    Comedy Central, Last.fm, and flickr.

It is basically a software application that you can use on a Mac or an Apple TV to view all sorts of online content including:

The company (Boxee) recently made some news after Hulu cut off it’s access – basically blaming the content providers who provide their shows to Hulu.  I totally hear Hulu – and truly believe they have been fighting for their users accessing the site thru whatever means they want.

But it had led to a cat and mouse game of Boxee figuring out a workaround, Hulu blocking that, and so on.  And we think that Boxee has finally figured out a way to keep Hulu on their software for good.

You can read all about the details of what has been added and why Hulu is likely to stay at Lifehacker.com.

Regardless – Boxee is a very powerful tool for consuming online video content from whatever source you want. Highly recommended.

via Lifehacker

Admin

I just got done with adding functionality that will allow iPhone and iPod touch (i.e. MobileSafari) users to see an optimized version of the site.

I am not sure how I feel about having a “second” internet, and developing versions for the various different platforms out there. But, the benefits that this new functionality provides for MobileSafari users is really great – it makes the site much more navigable.

So – if you own an iPhone or iPod touch – go to joshkaner.com on it. Pretty cool right?

Let me know what you think in the comments.