Display + Search

Increased revenue and higher prices. That’s the promise, right? As marketers are given the tools to find greater value in display advertising, more dollars should be allocated that way. Greater spend should drive higher prices. In theory, anyway.

When I evangelize biddable media with publishers I talk about the promise of higher prices. I stress that I actually want to pay more, because it means that the inventory is working for me. Once I see something that works, I want to buy as much of it as I can. I encourage publishers to push their unsold inventory onto an advertising exchange so that I can use the advanced tools available to identify, bid and pay what it’s really worth.

This is an important point, and one I still don’t think most publishers appreciate. As a buyer, I absolutely want to see cpm prices increase in a dynamic marketplace – because it means that there is value in the impressions, that buyers are finding that value, and that the determination of the price is fair and open. What I don’t want to see is cpm prices that are dictated by fiat with little or no transparency as to the value of the inventory being sold and no clue if anybody else is finding value there. Implicit in this are two assumptions.  First, that the campaign or campaigns I am running can afford the inventory (I don’t like to be outbid), and second, that the inventory for sale actually is valuable, or at least that some of the inventory is more valuable than the prices it was being sold at through direct or less dynamic channels. I’ll leave those discussions for another day.

Therefore, a sign that display really is converging with search would be increases in non-premium display eCPM’s. This is critical for the long-term success of the channel. Without higher eCPM’s in non-premium display, many publishers will simply refuse to sell impressions through the indirect channels that are most able to provide search-like functionality for display marketers.

On Friday, Pubmatic issued a Pricing Brief showing that CPM prices for inventory passing through the Pubmatic platform have been increasing every month since January. Eureka! It’s starting.

Or is it? Let’s look at the several data points we do have.

• Pubmatic Pricing Brief – reports a 35% price increase since January 2009 with increases every month starting in February culminating in a 15% price increase from May to June 2009.
• Rubicon Project Q1 Market Report – declining CPM each month in Q1 2009 after a 2008 that saw month-over-month increases. However, revenue growth was seen both across the platform and on individual sites.
• Exchange data – confidential data I’ve seen from several exchanges through April 2009 shows cpm prices lower than December 2009 but generally on an upswing from January through April.
• Search PPC data – Google released Q209 results and average price per click increased in Q2 vs Q1 for the first time in three quarters.

There is a lot of noise in this type of data, particularly right now. We’re in a recession, which could mean that the positive impact of convergence is overwhelmed by the negative economic environment. It’s also hard to control for seasonality when looking over just a few months. Finally, these are just three data points, although given that these three platforms see billions of daily impressions, looking at all three together should give you a pretty good idea of what is going on.

With only one definitive public report, it’s too early to definitively say that display CPM prices are trending upwards industry-wide.  And it’s definitely too early to conclude that we’re seeing the start of a long-term trend.  But based on the various reports and what I’ve been seeing in the market, I think the following happened thus far in 2009:

• Q1 pricing was flat or down vs prior quarters – driven primarily by seasonality after the normal Q4 holiday spike and, possibly, the delayed effects of the recession as 2008 budgets ended and smaller 2009 budgets were allocated.
• Q2 pricing was up – Pubmatic is the only display report which goes beyond April, and is very bullish through Q2. Rubicon has not issued its Q2 Report yet and the exchange data I have access to stops with April. The Google search PPC increase indicates that the economic haze may be lifting (although I should note that Google PPC historically is highest in Q2 of every year, though analysts do not know why).  I expect to see Rubicon and the exchanges show an increase in cpm’s in Q2.

Assuming I am right, why would display prices be increasing in Q2?  One conclusion making the rounds based on the Pubmatic report (from an Econsultancy blog and Sarah Baehr, Razorfish’s VP Media in a ClickZ article) is that the increase is due to publishers dumping higher quality inventory onto networks, thus allowing networks to pay more. Another conclusion is that technology is enabling smarter buyers to be able to pay more. ClickZ quotes Adam Kasper, SVP at Media Contacts US, “There’s a lot more targeting technology available that makes me as a buyer want to pay more…We’re adding more technology onto our network buys.”

It’s likely a combination of the two. Better inventory moving to non-premium and better technology to take advantage of it. Without better technology, I do not think the market would react to better inventory as quickly as these numbers indicate. Similarly, the impact of better technology alone would take a long time to see in the pricing numbers as innovations are introduced and copied over time.

Lastly, ClickZ notes that neither Baehr nor Kasper see much evidence for price increases in network buys. I would discount buyer viewpoints in favor of the platforms that get a more holistic view of the market. A buyer’s view of price is impacted by too many variables – the portfolio of advertisers active for that buyer at the moment, variability in the goals of those campaigns, specific targeting used,budgets allocated, etc. I’m a buyer and I generally refuse to answer questions about what I am personally seeing happen to pricing. Exchanges and Pub-side Optimizers have a much more consistent and uniform view of the market from which to evaluate real pricing dynamics.

I will be following these pricing reports going forward as I follow the convergence of display and search advertising. Over time, tracking changes in pricing of non-premium display will will help keep us grounded and force us to really understand and think about the dynamics of the market.