Commentary
Proving The Value Of Impressions
- by Joe Marchese , November 3, 2009
Problem is, there are an infinite number of impressions, which range in value from worthless to valuable, and when a Web site loads on a consumer's computer, they are all counted the same. This is a major issue for advertisers and content owners. But I believe there is a way to prove the value of impressions, a task that should be number one on the "to-do" list of anyone in online publishing or advertising.
It might seem odd that I believe we can prove the value of online impressions after I wrote last week about the problems with buying online media based on impressions. But I think the issue is not that impressions are worthless, but rather that advertisers need an objective way to determine the value of impressions. There were many interesting responses to last week's post, but none of them offered a solution to the real issue: How do you measure the value of online impressions? Or maybe even a better way to phrase the question: How do you measure the relative value of online impressions?
advertisement
advertisement
The answer, of course, is to create a system that can measure the relative value of impressions, a type of "ad unit rating system." Ideally, this rating system would be run by a third party (like the IAB, Nielsen or comScore) that could bring together the largest online publishers (a number of which are very large offline publishers, ala The New York Times, CNN, ESPN), all of which are struggling to support their cost structure due to the inability to demand premium values for their impressions. The ratings group could be a for-profit venture or a not-for-profit trade association. Here's how the rating system would work:
1. The third party would be tasked with creating "tester ad units" with some sort of call to action, like "click here if you are reading this." Ideally, it would be obvious to users that clicking on (or engaging with) the unit would not take them to a new page. The only goal of the tester ad units would be to get people to complete the simplest of actions to indicate they have seen the ad unit. QUALITY OF CREATIVE WOULD NOT MATTER... we will get to that.
2. The third party would then serve the tester ad units in place of online publishers' regular ad units to get a random sampling of people's "attention level" (I'd also call this engagement level). The number of times these units would need to be served to get a good sample shouldn't affect online publishers' inventory for paying advertisers.
3. Every publisher -- and, more important, each ad unit placement a publisher creates -- would get an "attention score." Perhaps of every one hundred people that see tester unit #1, 4 people do the action. For the purposes of this example, we could give that ad unit an attention score of 4.
4. No attention score would be meaningful on its own, but would instead be used to measure the relative value of impressions. For example: ad unit #1 got an attention score of 4, while ad unit #2 got an attention score of 2. Ad unit #1 could be considered twice as impactful for advertisers.
5. Third party creates a market where advertisers can access ratings of ad units. Advertisers could then buy on CPMs, which would make a lot of advertisers very happy, and publishers could sell on CPM. It is incredibly important that advertisers drive the initiative, because online publishers will follow the money.
Here's why this idea would work: it doesn't matter how good the tester ad creative is, because all the creative is the same in the sampling. Advertisers are not paying for clicks, and can therefore focus on developing great creative (that might not even ask to be clicked on), confident that people pay attention to the ad units where they are placing the creative.
There would be a lot of pushback from some publishers, but I believe the framework is pretty straightforward. Google already uses "smart pricing" to discount small publishers who deliver poor-quality clicks, so why not figure out smart pricing for impressions?
What do you think? Leave a note in the comments and drop me a line on Twitter www.twitter.com/joemarchese
Concern I have is that only about 20% of all viewers represent about 80% of the clicks. Not sure this is a representative sample. Many visitors are viewing the ads but are not clicking and may go directly to the site itself or do a goodle search for more information.
I forgot for a moment that this column is called "controversy served fresh daily" but I am glad it came back to me. This idea is so bizarre - so distantly removed from the fundamentals of how advertising works and how media pricing works - that I can only assume Joe is writing with the sole purpose of creating controversy.
Joe,
I definitely like this column better than the other one... however, I still think it misses the mark. You are treating all people like cattle and assuming relative "clicks" are all weighted equally in your view. We represent a network of sites reaching physicians. First off their is a relative value that your model misses - what is the value of an MD on a reputable journal such as JAMA vs. a consumer user on Joe's homepage about health? I'm not sure you proposed model answers that question. Second, I can tell you physicans in particular are bombarded with ads and messages. Many clients don't offer value they offer a brand.com ad (and may cases the physician feels she knows more about their brand than they do). Relatively speaking, physicians are very infrequent "clickers". Does this make the value of the impression less? On the other hand there are only 700,000 of them in the country.... seems to me the value of the impression is far more.
I like your thought process, and I think publishers are more than happy to help measure audience quality and even share "average performance per ad unit" types of metrics (which somewhat does what you are proposing with ratings). But at the end of the day, "engagement" and responsiveness depends on many factors (much of them outside of the publisher's control) such as quality of product and offer, quality impact and engagement of creative, etc... . It also request data sharing. Some clients won't even share metrics back with publishers... How are publishers to help improve performance when they are asked to drive with a blindfold on? This is a team effort. Good media requires good partnerships between publisher and advertiser.
@Mark glad you liked it? I do mean to get people talking/thinking, but I wouldn't say it's my sole purpose. There needs to be some sort of scalable solutions to objectively measure the value of impressions. Got any good ideas?
@RJ I am not saying that the "attention score" is the sole measure of CPM price, each network, and really any web property, still sells it's differentiating features (i.e. quality and type of surrounding content, like medical), all this does is provide a baseline for how prominent the ad unit is.
Perfectly unstandable that not everyone clicks, and the creative of the tester unit would have to be different in that all it is asking for is acknowledgment of message received (not a click to a new web page). Think "like" on a facebook feed. I saw it, liked it, but am not willing to comment, or navigate to a new destination.
Apples to Oranges. How do you measure the value of one over the other or even consider them the same. Advertising is not uniform or consistent or basically a standard commodity. The industry has not yet really defined “impression”. Instead of involving a third party to determine a standard evaluation, I think it is better to let the market decide. Did I get value for my advertising spend. CTR is usually under 1% so that is not really a measurement for value in comparison to billboard, print, broadcast or a determinant for interactive value.
Joe,
It is interesting that you have wrote this today. We've been working on the similar excersises, but instead of valuing each click, we are developing values for different formats for ads. Like many commenters stated here, different publishers have different 'values', which most people are aware of, however, ad formats, be it a top ad, side ad, video, rich media, have different 'attention' values as you so rightfully put.
The key is creating client and consumer value through branding points, recognition, top of mind awareness and engagement tied to measurable sales. The marketing mix is varied and there is not only one solution the key is what works and does that solution satisfy client ROI, goals with measurable objectives tied to brick and mortar and online sales. Where we are going to see increasing value and market competitive advantages is in the branded application "apps" especially if mobile and web platforms are integrated. This will make web impressions an old measurement.
Proving...what?
Maybe not being engrossed with the agency or sales side, I am missing something, but at its core it still relies on a click, so it somewhat comes to click vs. impression, regardless of how you slice it.
This would be a good idea except that there is already an "attention score." It's called a CPM and differentiates quite well how advertisers evaluate sites.
Just like in the offline world some media impressions are valued higher than others. The market aggregates that data and drives demand.
The value of display advertising as 100% relative to the context its served in, not the medium its served from... A tacky mortgage rate display ad on a high CPM or high "attention score" site (aka big offline media publisher) generates just as few clicks on a low CPM site (aka video game blog). Now, that same tacky mortgage rate ad served up on a real estate site or next to a mortgage financing article is incredibly valuable. Unfortunately, most publishers really aren't equipped to effectively place context targeted ads and as such can't monetize effectively.
Google keyword targeted display performs really good things for the B2B lead gen clients I have- they get a decent CPA that is very close to search. If only publishers had better tools to offer this level of targeting to the advertiser and relevant advertising to the consumer.
I don't think that selling CPM impressions like airtime or inches will ever EVER carry publishers- I'm convinced the way forward is they fire their ad sales teams and hire good biz dev people who place highly content targeted CPA offers through Commission Junction and direct CPA arrangements with advertisers. AKA the publish picks the offer to match to the content and profits from their effectiveness in delivering relevancy- not the publish pushes impressions for the highest price on whomever is willing to pay for eyeballs.
Lots of great theories, but what do you do when you CANNOT evaluate on a site-by-site basis? There are too many sites and too many types of ad units to evaluate. And you can't isolate the value of certain impressions versus others, because you are running many marketing programs at the same time.
It's not perfect, but I have yet to have anyone propose a better solution, only make case about imperfection. Based on that type of arguments, radio and tv wouldn't have a system to sell on ;-)
Also, basing value of impressions on the end effectiveness of campaigns for brands seems to put the publisher at the mercy of advertising creative more than ever, right? What if I deliver great impressions to a brand, but the creative doesn't change anyone's perception of the product or service?
And so...valuation insight, please?
see my comment on re visibility...might be a start?
http://www.mediapost.com/publications/?fa=Articles.showArticle&art_aid=116633