Oliver Thylmann’s Thoughts

Good article that nicely sums up what Google is currently looking into. In short, they are looking at having more content, which ends up in the search engine, and can have place for AdSense ads. Why send traffic to Wikipedia who refuse to take ads from Google when you can direct them to Blogger or rather something more Wikipedia like that has. I am not saying that they will hack the algorythm to put their own properties on top, but the problem runs deeper.

As soon as Google has their own properties, people are bound to question whether they are not artificially improving their own ranks. Whether they do or not is not important there. This is really the same reason that I believe Google will have a hard time going deeply into the lead generation business, in the sense that they launch generic information pages on subjects where people can register, selling those leads on, similar to how we do it. Whenever somebody else bids on a per click basis, people will be wondering whether Google does not artificially increase their conversion rates and get their own sites into the ad spots.

Getting a bit tricky here, but there do seem to be limits to what Google can do and they are just testing those limits out. Stuff will remain interesting.

A kid sure occupies you. A friend likened it to watching fire recently. Really not too interesting but you just have to keep watching. I obviously didn’t stop reading though so it is time for clearing the backlog again, dumping my brain into this blog once again, trying to get back on top of things.

• The wonderful people at MOO have launched their MOO Stickers.  I already love their business cards, and this ads to the mix. Looking forward to seeing the company grow into other markets.
• The guys at Townster posted about Vibe on their Blog. Great little video!
• AOL/Time Warner bought TACODA in another deal which moves a network for ads further towards the big players, giving behavioural targeting a further push into the main stream.
• Seth Goldstein launched SocialMedia, which builds Facebook aps, already having over 12 Million Users in total!!! They are looking for a designer by the way. If you know anyone, send me an eMail or leave a comment.
• As for ad networks, check out this article on media week entitled: Marketers Turn to Web Ad Nets. Also a real relevant point for all the people out there running performance based advertising solutions.
• Don Dodge wrote about the Microsoft analyst meeting with some interesting bits. They are serious about advertising, having bought AQuantive and AdECN, having deals with Facebook and Digg. The other big pushes from Redmond are software as a service and consumer services. Big changes going on at Microsoft.
• Clickriver is online, allowing your to advertise your services on Amazon.
• Amazon has started to allow payable AMIs on their Amazon EC2 service. I have the feeling that WeoCEO might be one of them soon.
• Valleywag posted the Facebook Ratecard.
• The NU2M portfolio seems to be transferred to Media Ventures by the looks of it. Watch out for some posters near you in Germany
• Desktop Tower Defense is coming to the desktop as a mutiplayer game. God the time we will loose!
• F5 Networks bought Acopia. A good move if you ask me. They move from traffic management to a more integrated system of managing your entire server system.
• Google is tracking deeper data with AdSense, allowing them to target users specifically based on their habits. With their reach, it’s an amazing behavioural targeting system.
• Plentyoffish.com is sticking with no employees for a while with big changes in the online dating market happening it seems.
• Feld posts about the Social Graph. One of our employees is working on something to allow social graph portability and Brad, ex LiveJournal / Six Apart, now Google, seems to be going in that direction too. This will seriously be a major shift and let’s see if it happens or rather when.

AdJug just raised $2 Million. Congratulations first of all. But this is pushing me further and further into the idea that we need a change in how all these banner servers work. Things like wunderLoop, possibly OpenAds, and others will play a role. Actually, I do believe everybody should play a role.

The problem I am starting to see, especially for the smaller sites, is that they will have to choose one network, like AdJug to serve their ads. Then they are quasi locked in. I personally want the highest paying ad running and I don’t care where it is coming from. I can now go into my installation for OpenAds on my blog and choose to run AdSense ads or AdJug ads or like I set it now, Ormigo Ads (just testing my own stuff here, nothing for your to test, sorry) when people come from Germany. It’s all sub-optimal though. Because for any given page, AdJug might have the better ads, or we might or somebody else might or I might choose to just run AdSense, or even opt to have a CPL AdSense campaign shown when it really fits super well.

What we need is a VISA kind of market for advertising. In general I would love to do this as a start-up, but the thing is that it’s not really a start-up as it is not there to go public or get sold or anything else to make money of as an investor. It’s a separate entity that will make sure that the ground rules are clear. A good starting point to get an idea was what Dee Hock asked when they thought about how to structure VISA:

If anything imaginable was possible, if there were no constraints whatever, what would be the nature of the ideal organization to create the world’s premier system for the exchange of value?

Still makes me go wow when I read this, each and every time. But this is really what we have to think about. Exchange “value” with “advertising” and a few things become clear. It needs to place those ads that are creating most value for the publisher. Here you need to factor in what a publisher thinks value is. A small publisher might opt to not have image ads, another one might. The most valuable ad will not always be in the same network though, so we need an ubernetwork right there. The advertiser wants their ads to be where they fit best, and that again includes all publishers. The idea is to have an open exchange, where everybody can play, and every member of the system can do whatever they want … as long as they abide to the standard rules/APIs. Think of it a bit like a peer2peer advertising platform.

So what might be an example process. I have structured information about my site. It includes pages, keywords for those pages, available impressions, average clicks, lead values, … I need to be fully transparent. It’s a simple XML file sitting somewhere, available to all, with information about the site and current booking status. You cannot book my site fixed, because it’s all performance based. Of course if you pay 100 EURs CPM you will likely get it fixed as there will be no other performance based deal that will beat that. Next to that there is a cloud of advertisers that tell each other and each publisher what their ads are. This would be simple format of the ads, sizes, content guidelines, targeting ideas (there are no rules, if it works it works), tracking information, … . I can also make a deal with wunderLoop to give further information about my users for possible targeting. They would just either add to my XML file for general information or publisher their own file that can intermix with mine.

Of course there are lots of performance issues to think about, and this might not really be a live system, which often is not even needed, but the general point is that everything is open and the systems will use those ads that are paying best. No more island solutions. An advertiser will be able to access all sites within the network, a publisher will be open for all advertisers. An agency will be able to really optimize bookings for a client as they know which different sources to mix and match.

Of course there are also problems. The wining and dining part of the advertising business might move to the background a bit. The best paying ads will run and boy is there lots of trickery going on there to make some things run more than others because of personal relations. This will have to go as adding this kind of layer will make the system overly complex and will be hard to be codified into a rule. This will really be the biggest problem to get a thing like that to take off.

The end result should be that the best ad from whichever network runs, because the banner server part on my site knows what pays what. It will not simply run the AdJug code and the AdSense code 50/50 but dependent on likely performance.

We are working on different things at Ormigo that are moving into this direction for a small part of the ad world, but it is not something to take lightly. Looking forward to your feedback.

We have a start-up in town and it’s called Shoppero. The general idea of a user driven review system, sharing the revenue with the users, allowing widgets to be placed on web sites with your review, … all that is good and I can see this thing succeed to some extent. The problem is that it is (was?) way over-hyped. We all know that Nico is very optimistic about the future and with Jens Kunath joining in, things are sure to get hot in the current investment climate. Underpromise and Overdeliver guys.

In times such as these, it is no failure to fall short of realizing all that we might dream, - the failure is to fall short of dreaming all that we might realize. We must try. - Dee Hock

They are going in the right direction, even though we already hear of all the problems, first and foremost huge holes in the platform. Werblogger wrote some things together and also pointed to this post on thinkphp about the big XSS hacks. Werblogger also pointed to the problem of the general rules of the platform, but Nico already corrected the point that they will not use your email for spamming. Didn’t expect that and I know that you will write your rules a bit more carefully if you are in a hurry.

But they need to be fast, because the platform is fairly simple to build. Hell, they did it in 6 weeks for a first release. Signing up you will get an Amazon aStore for example, but that is very easy to build and it took me a few minutes (store in german for now). Of course they will connect more affiliate systems but Amazon is really working on having almost everything available, so you are mostly fine there. And if you are big enough to turn any meaningful revenue for other stuff, you can do that too. Of course, this is long tail, but as said, this thing could work.

It just comes to the biggest problem I see with their communication strategy. They are saying that you will make more money with it. Jens had a long post about traffic optimization. The general idea is to give the user what he wants as an ad, and that even Google AdSense is not relevant enough in many case. Why is relevance so important? Because it will lead to higher click through rates. I can even totally attest to that because I have run performance marketing systems for some time now. Now Jens argues that through Shoppero bloggers will be able to get higher revenues than by other means at the moment. Sadly, there it brakes for me. That is total nonsense.

It’s a pretty simple calculation really. You can get a 1% CTR and more on AdSense if you place them right. Now let’s presume you are making $1 CPM on AdSense through this. That means that 1000 impressions lead to 10 clicks worth $0.1 each. That is not a lot and you can probably go higher, but lets be conservative. Now let’s look at my last 10 posts to see how relevant my ads are. I have router management and dsl ads on my Fonera post, the unddu.de post has social networking in there, my first post about mymuesli.com has VC stuff and the VOIP via Mobile post has lot of fitting VoIP links. All in all very fitting and nothing in there that would lead to a higher click through rate in my mind if I had linked my Canon EOS 350D or Nokia E70 or some Books or Wine or any other product. I mean Fon is linked anyway, mymuesli too and stuff like Jajah and Cellity are linked to. Nothing for Shoppero to earn there unless they change my links which I won’t let them do.

But let’s presume I do post about the Nokia E70 and how much I like it. Let’s presume that 1000 people read it (probably $1 earned via AdSense, possibly more as there would be other high paying E70 links). As the direct link to an E70 it might have a better CTR and let’s presume that it is 5%. Most of those would probably not even want to read a review (they just did) but get more info on the E70. The many people that don’t know me, are very unlikely to do an impulse buy of a phone (or any product) and will probably continue to think about it. Those that do know me, will probably send an email to get more views, pretty similar to how Jens bought his E61. In any case, many will not buy in a trackable manner but somewhere else. Knowing conversion rates of special landing pages for free services, I do not think that they will get conversions anywhere near or above 1% from after the click. And if I then get $10 for the sale, I would get 60% of the $5 made, meaning $3.

This obviously is more (5% is a high CTR! 2% and you are lower) but the problem is that it presumes that every post I do has a fitting review to it. That is surely not the case and if it is 1 to 10 than it is a high ratio. Suddenly I am only making 30 cents. Of course I am also making money on the stuff that Shoppero lists on their site (20%) but Squidoo tried that too and failed. Above that, suddenly it is no longer about making money via my blog but almost free for Shoppero. Above that I have work as I need to create the widget and it’s not ala AdSense where you put the code on your site and you are done. Sure, review bloggers can make money, but they also can with linking to Amazon directly or joining other Affiliate Systems. And yes, you are also making money on other stuff Shoppero makes money with, like AdSense, but I am pretty sure that you are making more integrating AdSense on your blog directly than by loosing people through click through rates and through an additional revenue share.

Jens actually suggests adding an adget (how they call their ads) on the right side of your blog which means that CTRs are way lower than at the end of the article (I know). It’s also no longer context sensitive, which is the entire point. As a short background, I have run AdSense similar ads for services on some of the biggest publishers and am doing that again, and we did beat out Google AdSense (and are doing it again with Ormigo for relevant traffic). The more interesting thing is though that one publishers opted to put links from a big affiliate system at the same spot (good for them, you need to try stuff out!) but stopped doing it after a very short time even though they probably got way bigger commissions than normally or rather good CPM without consideration for conversion (which is why the affiliate system stopped doing it :)).

The thing is that sales are just barely trackable on the net. Too much moves outside the tracking system. I would be very happy to be proved wrong, because we need something more for blogs. We are moving the same direction with Ormigo, going more long tail on the advertising system, but first need to increase the breadth of offerings where we can create customer contacts (which are instantly trackable). It’s hard. Blogs are very diverse, and relevance is a must.

As a summary. I think for special niche blogs or for pages with reviews, Shoppero might make sense. Then it might be too much work for the hassle though. I’ll probably add an Amazon aStore to the right side of the blog soon to try it out (too much US traffic that Shoppero can’t handle and OpenAds can do the targeting for me for Germany and other Traffic.)

Good luck with your start-up, it’s interesting times.

Update: There are other things than only monetary reward for using a system. This is something that Shoppero can play, because they are the nice guys. One of their biggest benefits that is.

I have been wondering how they do it for some time. The thing is that I can easily use a group of hacked PCs to randomly click ads through a normal IE. I am not really sure how Google would be able to find out which clicking PC is a bot and which is a real person.

There is one other solution though, which seems to gain credibility through this post on Webby Thoughts.

It makes me wonder whether Google has algorithms in place to notice when certain posts get traffic spikes and consider a higher percentage of those clicks as invalid clicks.

That is exactly the thing and the only thing how this can work. They will just monitor traffic over the entire network and they for one know that distribution of ads and distribution of content will remain relatively stable, so if something spikes in the distribution curve then they will presume it to be fraud and remove the spike. That seems what has happened in the above case.

This also means that if you suddenly have a short spike in your traffic this will probably not result in too much additional revenue, which will only come if the spike is not a spike but a real continuous growth in traffic.