Commentary
There Is No Excess of Inventory, There Is A Shortage Of Intent
- by David Koretz , May 7, 2009
Arguing that we have an oversupply of display inventory is like arguing that Google suffers from too many searches.
It is a silly argument that completely misses the point. Search engines rightly view the number of queries they serve each month and the growth of their query volume as an asset. Yet, after more than a decade focused on tweaking the algorithms to drive click-through rate, the majority of searches still do not end in a click. This will be true in perpetuity, because not every search has commercial intent.
Hundreds of millions of searches each month where not a single ad is delivered (try searching for hemorrhagic fever on Google), is not the sign of an inventory problem. It is a signal that neither the search engine nor the advertiser sees a relevant commercial match for those queries.
Similarly, the real issue in display advertising has nothing to do with inventory. The real issue is intent.
The fact that billions of page views go unsold each month is not, alone, an indicator of a supply problem. You have to first understand the quality of the supply.
advertisement
advertisement
Measuring the Quality of Inventory
Most publishers today take an incredibly simplistic view of inventory. Inventory that is easily recognizable by both
advertiser and publisher as having intent, such as a stock portfolio or an allergen count, is quickly sold off at premium advertising rates.
This is where it gets tricky.
The rest of the inventory does not demonstrably show a commercial intent just by a visitor showing up. Rather, you need to understand a lot more about your audience to deliver intent on these pages.
Yet unlike search, most publishers have very little visibility into the success of a display ad campaign. They cannot see the correlation between the audience delivered, and the interaction with the ad unit.
Even worse, most publishers lack both the contractual ability and the technology to optimize the ads. They cannot tweak creative. They cannot test multiple ad formats. They cannot test the ad against multiple audience profiles.
Publishers cannot continue to have the responsibility for a campaign's performance, without the authority to optimize the ad and align it with intent.
The Rules of Intent-Driven Advertising:
In order to correctly identify and satisfy intent in display ads, there is a bevy
of information you need.
Here are the Six Rules of Intent-Driven Advertising:
1. You need to understand the audience.
2. You need to understand the context of the content where the ad is being served.
3. You need to be able to deliver an ad format that integrates with the content.
4. You need creative that resonates with a given audience and in a given context.
5. You need to have reporting that measures success at a finite level.
6. You need an optimization system that will tweak the results to maximize success.
Interestingly, all six of these rules are successfully addressed in search advertising today, and all are addressed by a single company.
Search advertising will continue to capture an inordinately large share of advertising dollars until publishers begin delivering intent-driven advertising.
Unlike search, display advertising is a highly fragmented business with one set of companies that owns the content, another set that owns ad serving, yet another owns creative, and still others own the technology. Layer in an overabundance of industry associations whose members have divergent interests, and you have a recipe for disaster.
We need to deliver an integrated solution.
iPods,
Steel, Operating Systems, and the Benefits of Vertical Integration
History shows us that when companies deliver an integrated product, they gain a huge (sometimes monopolistic) market
advantage.
Apple was able to create a superior user experience with the iPod because it owned the hardware, the OS, the desktop software, and the music store.
As Amazon, Wal-Mart, and others that attempted online music stores can tell you, it is really hard to compete against a piece of an integrated solution.
In display advertising, we have the exact same problem. We cannot compete against search with a piecemeal solution. We need to deliver a deeply integrated solution for intent-driven advertising.
A Great Job Solving the Wrong Problem
Smart guys like Dave Morgan recently argued that display ad CPMs won't recover, but they
are looking at the wrong metrics.
There is a three-year wait list for the new Ferrari, while Hummers collect dust waiting for an insecure guy to walk in. Inventory is not created equal.
When are we going to realize that we are not in the business of supplying page views, we are in the business of delivering intent?
Right on David. I echoed this sentiment earlier today on my blog declaring that the publisher must become the marketer: http://bit.ly/VnLuB
David, couldn't agree more. Targeting the right audience with the right message in the right context shouldn't be so difficult.
However, we're trying to solve this problem in the B2B space, and its complicated enough that it takes all of our effort to get right.
Each audience and each intent is different, so the solutions, while vertically integrated, will likely be vertically specific as well.
Best,
Russ
http://twitter/follow_bizo
David,
This was one of the best, most concise articles on the subject and I could not agree with you more.
The problem is that a lot of agencies have very young planners who handle their metrics and they are taught "quantity, not quality". this does a disservice to their clients. Many Publishers seem to have the same mentality and that is why so many sights fail. I think it comes from a print mentality.
Slowly but surely intent-driven advertising will take hold once advertisers get a clue and start to demand it. It's like advertising in a down economy, those that do always come out way ahead. The prices are lower and visibility is higher because their competitors are cowering in the corner.
Well done. Additionally, I loved the Ferrari/Hummer reference.
The thrust of your article is right on, David. Publishers who want to attract and retain brand advertisers need to internalize the advertiser's objective and partner to deliver that objective whether it be intent or any other brand building objective, e.g. awareness, preference, favorability or attribute association.
It is time publishers stopped trying to foist measures that are convenient relics of the direct response game like click-through-rates onto brand advertisers.
Vizu works with numerous publishers and ad networks to help them sell more brand advertising by communicating their brand building capabilities to advertisers in a language that advertisers understand. This is a direct quote from one of them who forwarded this article to me:
"Great example for you from our recent campaign with Brand X. They were only measuring CTR, which is good on our site, and thinking it was a big success. We were able to show them they had 20% brand lift and 67% purchase intent interest as well, which was huge. What was the coolest, however, is that the data showed that purchase intent went up for the entire category, which showed that their ads were good but not doing enough to differentiate their brand as the leader. Basically, as the category leader, their ad messaging was all wrong. Blew them away.
Not only does brand advertising work on the web, you can measure it much better than in print.
I’m on a mission to educate as many brand managers as I can find!"
Do you think this Publisher is going to get a disproportionate share of Brand X's next media buy and be able to command premium CPM's? You Bet! Why? Because the publisher is showing a commitment to brand building and bringing valuable information to the table.
Note that we also seen numerous publishers build extremely strong relationships with advertisers by sharing information about why a campaign did not work and what could be done to improve it next time. Doing so is a fantastic way to build trust.
@Dan, I think you mis-understood the point of my argument. I am not arguing that intent is a brand metric. Far from it.
I am arguing that intent is a PERFORMANCE metric.
Brand and performance do not have to be so at odds with one another.
Misguided publishers and ad networks keep arguing they are separate, but brand impact is just one metric of measuring performance.
Huh? So you are saying that the laws of economics don't apply to advertising?
We have gone from a world of 10 media companies with a cartel like stranglehold on the supply of consumer time and attention to a world of hundreds of thousands of fragmented media companies trying to generate and hold waning consumer attention while trying to secure money from advertisers who have more choices and more ways to effectively judge your performance.
if you think understand your audience better and better reporting is the answer...well, the laws of supply and demand beg to differ with you.
The horse has left the barn. There is now, and forever will be, more supply of inventory than there is demand for that inventory. At all price points. At all qualities. At all points of the supply and demand curve. In other words, a massive excess of low value inventory results in a reduction of rates up and down the ad economy.
I've been blogging about the economic tsunami hitting the ad industry and the solution that fixes it all...www.OurSeatAtTheTable.com
-Dave
@David, did you actually read the article?
First, the top 100 publishers earn 91% of every ad dollar spent, so you can talk about "hundreds of thousands of companies" but the majority of those are HOBBIES, not businesses. They are feasting on scraps.
Second, the search industry has seen an increase in searches for the last 10 years straight, yet click prices went UP not down. That is because the demand increased based on the RESULTS they offered. Neither demand nor supply are static.
Third, supply and demand laws presume you are MEASURING A SINGLE PRODUCT. A publisher delivering users with an intent to purchase a product vs. just displaying a banner ad is not a difference in "quality." It is a different product altogether.
The real problem is that you think of every page view as a single product. Good thing companies like Google didn't share your perspective.
David.
Thanks for the thought provoking post. I agree that it is to the interest of publishers in the long term to own their destinies when it comes to delivering accountability to marketers. It is obvious that the more performance shaping variables that they have directly under their wing, the more confidently they can step onto the batting plate.
But this risk transference also means much more effort on their end - which immediately begs the question of how much more they can charge. Let alone the effort needed to QC out brand play parameters from unclear, or wrong, briefs; or their ability to connect the dots with what happens to the traffic when it hits the other end of the journey.
Its an excruciatingly painful trip to take.
Also one could argue that while "intent" or "interest" is a good way to look at contextual fit, this also demands more granularity in establishing what exactly that is.
Need-based or want-based motivations?
Personal or social motivations?
Beliefs?
Passions?
For most publishers who will never enjoy the long tail data that Google has on each individual, I would suspect that the best they can do is to resonate advertiser intent, be it a belief, or passion, or behavioral trait, they want their brand to hard wire with, and differentiate against.
This is an extremely dynamic code that should be left to trained communication professionals and technologists to algo off. Simplifying this code into a single currency or focus, is akin to pouring water into the petrol tank?
hwooter @twitter
David, Thanks for your insight on this. I have so much to say on this subject. Forgive me if a go a little off topic.
Rules 1 through 4 should be easy:
Here is what we know: Advertisers will still pay a premium for display ads with the most credible publishers in any vertical that is editorially relevant to them. If you are Gucci, you choose Vogue over People - it's good for your brand, and you'll pay a premium for it. So part of the effectiveness of the ad, from a brand perspective, is decided by the brand value of the publisher. If you are Old Spice, then maybe People has more value than Vogue.
Rule 5 and 6 are the difficult part.
If we are not looking at direct response advertising, then the click does not tell the story. With great creative, maybe with a content driven widget, we can measure engagement, to tell part of the story. But that still doesn't measure the value of the impression itself (the impression that does not get clicked on, or interacted with in some way). How do we measure whether our impression actually left an impression? Some people believe they have the answer, as Dan from Vizu points out. And that may be part of it. But, short of attaching an ocular measuring device, and a EKG to everyone surfing the internet, I'm not sure we will ever have that perfect measurement.
In the meantime, one thing is for sure: if you place an ad with the wrong publisher, you will get poor results. Supply and demand set's the price - and there are still plenty of advertisers willing to pay $10CPM's and-up for the right inventory.
There will always be a waiting list for the next Ferrari... 'cause driving one makes you look so damn good!
David - great article. A lot of companies we work with understand and use search much more intensely than they do display advertising, and so the comparison and the contrast between the two is always useful to look at and discuss.
The "too much inventory" argument is definitely one where it is easy to miss the point. In display and search alike, if there were one easy standard interface that allows for placing advertising (e.g. Google), zero transaction costs to putting up an ad (same creative, no API costs!), and no minimum bid (and/or no accounting for the user's costs in having to wade through irrelevant ads - quality score), then of course there would be ads for most searches and most display space. In the same way that the blackjack table has a minimum and a maximum bet, the latter to prevent you from just doubling your bet every hand, whether the impediments are publisher-made or not, the reality is that the testing costs to trying out so much inventory to see if it works is just impractical. The natural and unnatural impediments (like floors) that exist even in the ad exchange world exacerbate this.
So more dollars will flow to more "known" quantities in terms of sites - large publishers, brand names, big networks with a track record and some data. And for the most part, this is rational. Some ad networks have done a good job of aggregating a lot of small site supply, but there is too much difference in things like how sites are grouped, where and how publishers put ads on their sites, impression fraud etc. for this to work as well any more.
Overall, as an industry we need to bring to bear whatever tools we can to measure, grab, map, correlate, track and re-use intent - and publishers, vendors, advertisers will have to come together to help make it happen and start to standardize how this stuff is created, traded and used. The publishers who (with help) can create constructive and open solutions for advertisers to better run, test and optimize campaigns are going to win, whether the inventory pie increases 10% or 10,000%.