Shahar Solomianik

Since Twitter allows only 140 characters…

Archive for May 2009

The RightMedia Problem, The Microsoft Lesson, And Why OpenX Will Eventually Win

with 54 comments

The RightMedia Problem

The advertising exchange model is very promising and great in theory. Letting publishers put their display ad inventory up for bidding by advertisers really sounds like something that can improve efficiency both on the publishers and advertisers end. In practice however, things are different.

RightMedia is probably the largest advertising exchange today. Yet, publishers still use it mostly for their remnant ad space, allocating their quality ad space for other advertisers and top tier networks. Advertisers – completely aware of this fact – prefer to spend their display budgets with established and reputed ad networks, and leave RightMedia bidding for the smaller advertisers and networks. The result is that the average CPM on the exchange is low.

It seems that the exchange still didn’t reach a critical mass that will bring the breakthrough in its performance. The only way for the exchange’s average CPM to increase is by adding more advertisers and publishers into the game. Only then will publishers allocate quality ad space to the exchange, and large advertisers and networks will come along.

How does an ad exchange grow?

The Microsoft Lesson

According to Yaron Galai, who had suggested Microsoft to offer a 200% rev-share to all publishers, growing an advertising network – and I believe rules apply here for ad exchanges as well – is by growing the publishers base:

While the advertisers are the ones paying for everything, acquiring advertisers is a secondary concern for an ad network. A distant second. The #1 key to making an ad network work is the publisher side. Even though the publishers are being paid, it’s much more difficult to win publishers than it is to win paying advertisers.

Whether or not someone at Microsoft have read Galai’s post, it seems that they had followed his advice. They realized the only way to win market share over Google is by attracting Google’s publishers – not advertisers. So the rumor about an high-paying AdSense alternative spread and many publishers were eager to join once the private beta tag is off. Few weeks later though, when Microsoft opened up pubCenter for the public, payout has dramatically sunk.

RightMedia shouldn’t be learning anything from Microsoft about publishers retention. However, with acquisition in mind, Microsoft’s experience can serve as a good lesson for RightMedia . Tempting publishers works. If RightMedia could only find a way to tempt publishers to join, it will be able to grow its publishers base and advertisers will follow.

But there’s a catch – unlike Microsoft, RightMedia can’t simply double or triple publishers payout. Being just the exchange manager, RightMedia is not involved in the money flow. In the exchange, money flows directly from advertisers to publishers and not via RightMedia as the middle man. RightMedia is not in a position to offer any financial benefits for their publishers.

That’s a problem. And not only for RightMedia, but for any ad exchange out there. How can an ad exchange tempt publishers?

Why OpenX Will Win

If an ad exchange can’t compensate their publishers with money, it may never grow. Only an ad exchange that is able to offer another type of compensation, a service or a added value perhaps, has chances to appeal to publishers. OpenX is such an exchange.

Wait a second. Isn’t OpenX an ad serving solution?

Well, it is. OpenX is an open source ad serving software that is becoming very popular. It disrupts the ad serving business by eliminating a great portion of the ad serving costs. Disrupting is great, but OpenX is a business and needs to make money on its own. Having already a big (and steadily growing) amount of publishers, with a very close affiliation to their advertising space, it was only natural for OpenX to eventually launch an ad network of some sort. Better yet, it decided to take the ad exchange path.

Though still small, I believe OpenX exchange has better chances to make it. It has a large base of publishers enjoying a free or very cheap ad serving solution, whether as an hosted solution or as a stand alone installation. Converting those publishers to participate in the exchange may be much easier for OpenX than it would be for RightMedia to find new publishers out in the cold market.

In a sense, OpenX is not an exchange that offers an added value. It’s an added value that offers an exchange. And that’s why they’ll eventually win.

Written by Shahar Solomianik

May 25, 2009 at 3:13 am

Posted in Online Advertising

Tagged with , ,

When Will We Have Perosnal Targeted Advertising?

with 22 comments

Online-advertising targeting methods are great. As an advertiser, they allow me to narrow the distribution of my campaigns and reach out to my clients more effectively. So if I am after clients from Phoenix, AZ, I’d use geo-targeting. If – on the other hand – I look for people who are interested in dog training, I can keyword-target this term. Newly buzzed Behavioral Targeting can get me even further and enables me to target clients according to their recent tracked online activity.

However, what if I actually know who the people I target are? What are my targeting options in this case?

First of all, you may stress that if I already know who I am targeting, I shouldn’t use advertising as the method for reaching them in the first place. I can try to contact them personally instead. Isn’t that what social networks are all about? Especially Linkedin, which has the introduction mechanism so well implemented in its core.

To this I will answer that:

  1. The amount of people to whom I want to communicate a single message can be too big to justify personal connection. It’s a simple scale issue.
  2. The time it takes to become friends or get introduced to those people may be long. In some cases, I need the message delivered immediately.

Current targeting options do not allow me to personally target advertising campaigns. I can use a mix of other targeting methods to try to narrow the distribution as much as possible, but that’s not really it. Let’s say that the common denominator of all the people that I target is that they’re all video producers from NYC. They’re not just video producers from NYC. They are very specific video producers that I have personally and manually targeted. I can launch a campaign that is geo-targeted to NYC, and keyword targeted around “video productions” very easily. This campaign may as well hit the exact people I am after, but it can also miss. And – it may reach people that I have never intended to reach. If they are exposed to the ad or even respond to it can create useless noise for me or even damage.

What I actually need in this case is a method of Personal Targeting, so I can communicate my message effectively only to those people that I am interested in. I want to be able to put my ads on web pages that those NYC video producers – and only them – are viewing. That’s a truly targeting heaven.

Imagine that while you’re browsing the web, the ads that you see are not ads that are targeted for your “type” but rather, directly to you.

Of course, the only way to practically achieve this is by knowing when its you who visits a page. And in order for this to happen, you must have identified at least once in the path to the page where the ad is. The natural implementers of this targeting method should therefore be social networks. Social networks are the places where you are willingly identify yourself by your real name with the intention of it to be publicly known that you are you.

There are surely some privacy issues here. Although I am not sure those are much of an issue, since the only information the personal targeting vendor would use is the person name. It wouldn’t use any data that this person shared on their social accounts, or any information the person may not to expose to others. Users names are usually completely public on social networks.

Another issue is scalability. It sounds like personal targeting involves very small numbers of ad impressions and future transactions or user actions. Advertising networks are used to work with multiples of tiny figures (costs) with huge numbers of impressions, clicks and actions, and why would any of them develop a targeting method that will allegedly reduce their business?

I think networks business wouldn’t be reduced, rather become more effective. Advertisers targeting personals will be willing to pay a lot in order to reach the exact people they’re after (think Bill Gates), or for their future actions. So the multiples will be of relatively high figures (costs) with relatively small numbers of impressions, clicks and actions. The network will have to carry much less overhead. That’s a benefit.

All in all, having to deal with the pain caused by not having this targeting method, I can’t see many drawbacks of Personal Targeting. Can anyone else see them? I just hope that the major social networks will develop Personal Targeting soon, or are they already developing it?

Written by Shahar Solomianik

May 14, 2009 at 8:03 am