Display ads and CPM are broken

This is the idea that the CPM model is an inadequate measures of performance and therefore their pricing needs to change.

Problem (s)

1) CPM.  We still don't know that when a user sees a display-ad if she is more likely to buy that product. We can run focus groups, we can measure clicks, we can run surveys.  But we don't have any directly correlated and measurability.   This is one of the reasons why I am sure we will see a proportional drop-off in display ads (recession or no recession).

2) Forget consumers, advertise to influencers. 
It is not really display that influences buyers at all.  It is key community influencers. 

Solutions

How are people solving the above problems?

1) Targeting 
When you’re concerned about the veracity of your performance metrics, the obvious way to get around them is to advertise only to be people who you know want to buy.

A range of methods exist:
•    Targeting certain online behavior like Blue Lithium.
•    Targeting specific geographies like Yahoo Smart Ads
•    Targeting specific interest groups
•    Targeting certain communities only at certain times “timing”. E.g. look at Freeads.co.uk weekdays 11.30 – 2pm; it will be dominated by McDonalds ads.

2) Influencing the influencers

Good customers cannot be identified soley by their purchases (see Leveraging user generated content from forrester ).

Some advertisers are beginning to believe that it doesn’t matter if you’re CPMs are wrong, so long as you’re targeting the right people. 

“Now it’s about activating people of influence. They will then promote your brand continually because it protects and extends their personal reputation. People activate based on uniqueness of information, trust and credibility of source. CNET claims to be zeroing in on the 50% of advertising that does work.”

Neil Ashe, CEO CNET . See also "Understanding Influence and Making it Work for You,".

CNET claims to get higher CPM not because their audience are High Net Worth Individuals (like FT.com & Economist does) but because their audience is chock full of influencers.

Forrester have gone one step further by claiming that metrics will focus on the acquisition of  evangelists rather than purchasers.  They predict that a completely new metric will emerge called the cost per acquired advocate (CPPA).

3) WOM agencies

A few specialist word of mouth agencies are emerging specifically to target the influencers.  They are telling brands to allocate their cash to specific users and attempting to knock the bottom out of the CPM market.

Conclusions

There seems to be lots of suspicion today over online performance metrics. A number of developments are trying to solve this including media mix optimization and new targeting techniques.

The concept of effectiveness is core to bridging the online offline shift but arriving at a single measure of effectiveness is a big challenge for the industry.

The commonality for all advertisers is that they want all want a reduction in risk. They want to get better bang for their buck.  They know the web is changing things but it’s still too early to know how and exactly what they should do.

Comments

[PaulSweeney]

Really tight neat piece.

[David Cruickshank]

Hi Paul,

Very thought-provoking.

I'm not sure CPM is so much a measure of performance as a pricing mechanism. Arguably, it's the application of CTR (click-through rate) to CPM that provides the performance measure.

Ultimately, Cost Per Action seems to be optimal for advertisers. It suits publishers less, because the intangible value of brand building is not converted into value for the publisher (amongst other reasons). In the online advertising market, the best model for buyer is not necessarily the best for seller too.

I see a bit of an ever-flux between CPM and CPA (with other options such as CPC and cost per lead inbetween) depending on market conditions. Advertiser-friendly conditions will cause a move towards CPA and Publisher-friendly conditions, CPM. My sense is that we are moving towards a greater abundance of inventory and eye-balls online which is why CPM will gradually give way to a more tangible arrangement in the medium-term, IMHO.

I like the idea of influencing influencers which is similar to one of the core messages in Gladwell's Tipping Point.

[Don Jones]

There is another belief system that says to NOT focus on influencers, focus on the "sneezers" at the grassroots level. They're the customers who then tell their friends...

[James Cherkoff]

The idea of opinion-formers has been around in marketing for a long time. However, it may be the case that each marketeer will need to develop his own metric model to find this elusive bunch. Unlike mass media, whereby the world was divided roughly into chunks that appealed to many different product categories, today's kaleidoscope offers the possibility of creating media portfolios that fit specific brands like a tailored suit.

[Vitaly]

I have a strange idea about targeting influencers/opinion-formers. If they are the guys who should spread the word why not let them leave product reviews on the advertised product site and then post those reviews on ads? Let's presume we have rectangle ad formats suitable for text. The ad would have advertiser logo, domain name, link to the site somewhere in the header, then comes the latest review (chosen randomly), and then comes 'Read more...', time stamp, user name, linked to email if the influencer preferred to do that.
Of course someone would need to moderate reviews and (!) which I like best: negative reviews should get into the pool on a par with positive and middle. Personally I prefer to learn about the shortcomings first and then read about the advantages. So the optimization mechanism would learn soon that negative reviews get more click-throughs and more conversions and more acquired advocates.
I know it sounds crazy, all of it. But many great ideas sounded crazy in the beginning. What do you think? Have you heard of anything like that? Will it work? If not why?