The Information Revolution
Although web publishers and media companies are collecting consumer data at a faster pace and in greater volumes, the focus on putting consumer data to work has been mainly the domain of marketers, ad agencies, and other buy-side players (looking for specific socio-demographic reach, building databases of users that are likely to buy/not buy etc.). In short, the desire for better analysis, more precise consumer reach, and improved campaign ROI has catalysed considerable data usage on the buy-side, to the buyer’s considerable advantage, in a market that is already tilted in the buy-side’s favour (e.g. supply already far outstrips demand).
Meanwhile, on the sell-side of the equation, vast reserves of data are lying unused. This article is a call to arms, encouraging publishers and media companies to take full advantage of the data at their disposal. They must put that data to use in support of their advertising businesses to ensure they survive and thrive in an increasingly audience-aware and data-driven media market. Get ready for the Information Revolution!
A History Lesson
Fig 1. 1990 Jeep Cherokee, White
In the beautiful spring of 2010 I made a stupid decision: I bought a car. I had my eye on a sparkling white Jeep Cherokee, born 1990. The seller was a pleasant and joyfulyoung woman with a screaming toddler. She described the car as one of the only ones painted this special sort of white and assured me everything was still in working order. She demonstrated the electrical windows and the air conditioner, and she pointed out the high-end sound system.
I walked around the car. I poked in the chassis with my screwdriver (some rust, but not at critical spots, this certainly would benefit my position as a negotiator), kicked the tyres knowingly, and intentionally frowned when she started the engine. I then poked randomly – but convincingly – under the bonnet to demonstrate my limitless knowledge of cars and engineering. “Buy it,” I thought, and if I could exaggerate the rust factor, I could get a great price. And I did!
Within a few months, I found that having a 4×4 in the centre of Amsterdam was not ideal, never mind the massive fuel consumption that came along with it. So, I decided to sell. No worries, I told myself: I paid a very good price, and I might even make a small profit on it! Wrong. The first person to look at it was quick to tell me that the gearbox was rubbish, the rust actually was on essential spots, and that the air conditioner probably had not functioned since 1997. In short, I sold it at a loss.
The moral of this story? Information is value. I bought the car because the seller added information to the object that I desired (A unique colour! Everything seemed to work!), and I paid too much. When I wanted to sell, someone brought more information to the negotiation than I had – using that information to drive the price down – and I ended up selling for only half of what I originally paid.
This dynamic is true in most markets. The more (positive) information a seller provides about the product, the better their chance for realising a high price. The more information a buyer has, good or bad, the higher the likelihood that they will have an edge over the seller in price negotiation. In a traditional auction, not disclosing key facts will keep the price low (auctioning a Van Gogh, without shouting it’s a Van Gogh… well eh). And in the stock market, let’s not forget that the kind of information asymmetry we see in digital media is often considered criminal. Again, information is value.
Who has the Information?
As a seller you should understand what makes your product valuable and transmit that to potential buyers. In the digital media marketplace, the seller (media owner) most often owns the first-party consumer relationship via the content and services those consumers enjoy. As such, they know the most about the consumer and are in the strongest position to characterise that consumer in a way that’s appealing to buyers. Unfortunately, recent market dynamics don’t reflect that, and sellers are often at a disadvantage.
Let’s make the current situation more visible by drawing out three balances.
Fig 2.Three balances (the + is the weight)
As you can see there’s plenty of information on impressions, both on the buy- and sell-side. Buyers may know more about their direct history with the particular person that is reached (what did this person search for or view on the marketer website, what was bought, what is the likelihood this person will buy again). The media owner knows more about the context and the page that advertisement may be shown on, the consumer’s interest based on past behaviours, demographic details (age, gender), geographic location, and so on. In short, the media owner has a unique first-party relationship with his consumers that no advertiser will ever have based on the value of the content and services that keep those consumers coming back.
Currently however, it’s mostly the buyer that is using information at scale, particularly in automated channels. While publishers were preaching “content is king,” buyers soon realised that “content is really important, but specific consumer information is truly king.” By finding scalable applications of consumer data via media exchanges, buyers have been able to challenge the publisher’s ‘gatekeeper’ in reaching consumers. They marry premium information with less-premium inventory and end up with a very interesting ad placement – not as good as the premium placements of a publisher, but certainly a very cost effective alternative to publisher direct sales.
Information = Transparency
The simplest method for publishers to convey the product value could be pushing additional audience and placement characteristics into the bid-request – roughly the equivalent of the merchant on the market street shouting what he has to offer. However, many publishers have the (well-founded) fear that this kind of transparency results in cannibalising premium sales. Moreover, there’s concern and fear that buyers will use the information the seller provides for their own benefit after the initial bid/buy (such as retargeting the publisher’s valued users with cheaply bought impressions).
As always, the truth is somewhere in the middle. I would like to turn it upside down. No advertiser will ever buy an impression, for a decent price, that is completely blind; accordingly, most sellers provide at least some information about the impression at placement at the time of sale
Let’s take a look at that good old “yield pyramid” to illustrate – slightly simplified to illustrate my point. On top are the “most interesting” impressions, on the bottom the not so interesting impressions.
Fig 3. Illustrative model for “today”
When not adding (enough) valuable information to the impression, a publisher faces the situation in Figure 3. Some interesting impressions – easily sold by the sales team and through programmatic premium and a whole lot of not interesting stuff, sold through open RTB and remnant channels.
By contrast, while oversimplified, we have Figure 4. By adding additional information to impression pushed through programmatic channels, you can expand your pool of valuable inventory.
Fig 4. Illustrative model for “a better world”
Data – or as I’ve characterized it, information – should be better leveraged by media sellers. Even by adding just a little more information (e.g. something as simple the specific URL in an RTB bid request) to every impression the publisher sells, the volume of “interesting stuff to sell” will grow and make “the pile of not interesting stuff” shrink.
This is not a call by me to start giving away data or wilfully introduce the risk of cannibalising premium. Instead, it’s a call to take a more balanced approach to product definition and inventory packaging. Information is value, the audience is the inventory. Ultimately, it’s a matter of effective inventory segmentation, a dynamic product model, and careful channel management ensuring programmatic and direct sales are never working against one another.
Bottom line: It should be the buyer who declares what characteristics they find interesting and valuable. And it should be the seller intelligently packaging and pricing their audience and content assets, taking those assets to market via the channels that reach the largest number of relevant buyers.
For data/analytics, inventory/pricing management, and product strategy personnel in the publisher world should be incredibly excited. Those three roles are hurtling towards one another, all thanks to consumer data and the addressability of almost all advertising inventory, across almost all digital media. The functions will be one in the same, the very definition of yield management, and it’s on their shoulders the publishers business will struggle or thrive.