My presentation is on Slideshare and here:

I had a great time, especially enjoying the presentation by Fru Hazlitt, MD of GCap London. Presenting was fun too. I recorded the audio, so I’ll replace the Slideshare with a movie of the slides plus audio soon.

My presentation explored the origins and possible solutions (including the one mooted by IASH) to the problem of how to safely place ads next to user generated content. Essentially, I think a combination of behavioural targeting, automated content analysis and manual publisher classification is probably needed, and that a code of practise based solution like IASH’s won’t solve the problems to the satisfaction of advertisers who need innovation and results.

The answer is almost certainly that the ad agencies bought some ad space on a network. Networks are companies that sell the advertising space on large numbers of websites. Some have their own technology delivering ads, some are simply sales organisations that use other companies for the delivery of the ads.

What’s been happening is that these networks have been selling space to each other, so an agency’s purchase of ad space on a big high-profile network might have ended up being delivered by someone else entirely: and before you know it, your expensive brand is all over a streetfighting website. The practise is called “blind chain buying” and a lot of people in the ad industry want to stop it.

So this morning I was at a meeting where a presentation was given by some people who represent the Internet Advertising Sales House, which is a body of concerned ad networks, who want to introduce some good practice into the ad network world.

They’ve done some very good work, I think, defining a code of practice, recruiting leading industry members, putting auditing in place (using ABCe), and next month the first tranche of audited members will be announced. The code is all about ad networks not allowing questionable sites into the world where reputable brands advertise, and also about identifying general advertising opportunities separately from niche ones. For example, anyone might advertise at Reuters, but only certain brands would advertise at Loaded, so they have a “standard” and an “optional” classification for websites.

This is a very worthwhile initiative. There are issues, which I’ll summarise below, but it’s a step in the right direction, and I was impressed by the thought that had gone into getting things right.

The issues? Here are the main ones I saw:

• The classification system relies on ad networks and sales houses vetting and classifying their own publishers. Different people might classify different sites differently, and the process for normalising this doesn’t seem to be there yet.
• Some websites have areas which might be standard and others which might be optional. Wanadoo (the old Freeserve portal) has gambling content, as do MSN and Yahoo in the UK. So the idea is that sections of a site would be classified differently. This sounded like it could easily go wrong…
• … and just because part of a website is OK, that doesn’t necessarily mean that an advertiser will be happy being in that part, if the adult content is only 2 clicks away.
• It didn’t seem to address the excellent solution people like Flickr have developed (explained here) for having adult content, but keeping it behind a secure user preference.
• The system is based on having a vetting process in place, and auditing that process. It isn’t based on checking what sites actual ads appeared on. This is probably an issue of concern for most agencies and advertisers.

The other thing I was left wondering was if this will help to bring about a website classification system, so eventually pages that carry advertising will have a classification on every page, like computer games and movies have today. Will we end up classifying everything?