Online ad models breaking?

By February 8, 2008 4 Comments

Advertising networks

There is a good bit of chatter at the moment discussing the efficacy of online ad models. The major point coming from an e-consultancy piece and then embellished by Mike Butcher on Techcrunch UK is that ad-inventory is at the same time rising in volume and falling in quality – resulting in plummeting CPMs. Plummeting CPMs means plummeting revenues for publishers and community sites leading to questions about sustainability of the web industry.

This is similar to an argument that has been around for a while, and often espoused by Alan Patrick – namely that the ever rising number of web2.0 companies can’t all hope to survive because their simply won’t be enough advertising dollars to support them.

There are a few things going on here. I’m going to try and separate them out.

Firstly – not all web2.0 startups will survive. In fact, a few great ones will, but most of them won’t. That is the cold hard logic of the capitalist system. It is also what makes it great – to have successful innovation you need to have a lot of failed innovation. So the notion of an ever increasing number of companies getting an ever decreasing share of the same advertising pie doesn’t hold true. Most of the startups won’t ever take much of it.

Secondly – as what we do on the web is evolving the notion of CPM advertising is becoming problematic. The issue starts with the humble page view. When the dominant activity on the web was reading web pages – e.g. news – then advertisers knew roughly what they were getting per thousand page views. They could make adjustments depending on the demographic of a site and the type of content – but one page view wasn’t radically different from another.

Not so anymore.

Surfing habits on social networks are vastly different to those on old-school content sites and users spend much less time per page.  The situation has been further complicated by differing use of AJAX such that what is one page view on one might be two or more on another.

Thirdly – the debate about CPM, CPC and CPA is straightforward if you are thinking about advertising with a pure below-the-line direct marketing mindset, but not if you also have brand objectives.   The three different systems are fungible if the only goal is to drive sales – by monitoring click through and conversions CPC and CPA can easily be converted into effective CPM (eCPM).  But some campaigns are only brand based, in which case click through is not necessarily a revealing measure, and CPA is by definition irrelevant.  Better would be survey based measurement of post impression inclination to buy, but that is difficult to come buy.  If brand awareness is the goal measurability on the web may not be much better than for TV.

For me the other important thing to remember is that people often don’t see the ads anyway, although as Alan pointed out to me, that might not be much different to newspapers or magazines.

Despite these problems with the current system I don’t see a crisis.  For sure some companies will disappear, just as other new ones will come up, but the total amount of advertising spend on the web will continue to grow.  Of that I’m sure.  All the forecasters predict it and for me it is a natural consequence of media consumption shifting from TV and other places to the web.

New forms of measurement are needed though, and the obvious thing to base them on is time.  The notion of a page view ignores completely whether someone sees the ad for seconds or minutes, and this is at the heart of the problem.  The longer you are exposed to an ad, the more the advertiser should be willing to pay for it.  Following this logic through it might make sense to stop thinking about frequency caps and start thinking about length of impression.  You might want to show the same ad on ten pages on Myspace where the user is hopping from page to page, but only twice on a site like the FT where the user spends say five times as long per page.

Click through rates and CPA calculations are then appropriate for campaigns designed to drive sales rather than brand awareness.

At the end of the day the fundamentals of new media are the same as for old media.  If people spend a long time in front of your content then you have a valuable property, and if not, not.