Ari Paparo, Senior Vice President of Product for real-time ad platform AppNexus, gave an engaging talk at last month's AdMonsters conference in London about real-time bidding and 'Yield Management 3.0'. After his presentation, we asked the former product leader for DoubleClick and Google some more questions about the impact of RTB and plans for AppNexus expansion into Europe.
What impact is real-time bidding (RTB) having on publishers, advertisers and agencies? Who are the winners and losers?
RTB is having a seismic effect on the entire display ecosystem as it is both removing inefficiencies and exposing new opportunities. It's hard to characterise one group of "winners" and "losers" since really it's forcing all parties to evolve and become more efficient.
Publishers are now getting the tools to address their channel conflict concerns head-on rather than make gross decisions on who gets access to inventory.
Advertisers are able to use valuable data to reach their audience at the exact right time, in the exact right place, making the value of the impression that much greater.
In addition, agencies are likely to play a bigger role in the evolution of RTB over time. Their role as media buyers will probably change, and those who can bring innovative buying techniques to their client will benefit, as shown by some of the more advanced agency trading desks and buy-side performance companies.
At the AdMonsters OPS: Markets conference in London last month you talked about Yield Management 3.0. Can you explain what you mean?
Publishers have been looking to optimize yield since the first ad networks emerged in the mid-1990s. In the first stage of yield management, 1.0, this was a manual process using tags. Publishers would push and pull tags into their ad server based on daily reporting from ad networks and attempt crude optimization.
In the second stage, the supply side platforms ("SSPs") emerged and enabled the use of rules to increase yield. These parties added value by pulling in ad network data more rapidly and in an automated fashion, then creating rules on which sources of demand correlated with which types of inventory, thus creating better results.
This world has evolved over the past year to include RTB, but most of the decisions within the RTB framework (floor pricing, private exchanges, etc) are built within sophisticated optimization systems.
The third stage of yield management is emerging now and is based around using data to inform yield decisions. The RTB environment provides publishers with a complete view of their demand curve - this is unprecedented visibility for a supplier in any industry, and an asset that publishers need to take advantage of.
Using the data on the demand curve, publishers are setting intelligent floor prices and bid biases to manage channel conflict at a more granular level; optimize revenue while limiting user experience impact; and capture demand money where there is a willingness to pay.
Do you believe there has been a shift in power to the buy side?
Right now, it is fair to say that the buy-side, utilizing DSP (demand-side platform) technologies, has an advantage in technology and technique against the sell-side. The metaphor I've heard publishers use is that the current situation is like "someone digging for gold on my property, and they know where it is while I do not."
Over time this situation is likely to reverse since the supply-side in RTB environments gets a complete view of the demand curve for each impression, whereas the demand side will only see the curve for those impressions actually won. With improved tools and a greater attention to the nuances of the auction environment publishers, should be able to move beyond simple price floors and private exchanges to more complex and higher yield strategies.
Some people talk about a 'race to the bottom' for publishers in terms of pricing. Do you think this is true?
Premium publishers can already restrict the availability of their inventory in order to preserve its value. For some publishers that means not participating in RTB or ad network environments. However, for the majority of publishers that means making intelligent and informed decisions about how to manage their channels.
For example, some publishers are allowing entry into their private exchange pools only for those buyers who spend a certain minimum on reserved inventory. This arrangement maintains the direct relationship while giving the buyer the advantage of algorithmic, real-time buying.
For undifferentiated supply, it will be more challenging. With a general imbalance of supply and demand (more supply), pricing for commodity inventory will, by microeconomics, be driven down.
Can you explain the reason for private networks and private exchanges? And what the impact is.
The primary impetus for private networks/exchanges is to maintain a direct relationship between sellers and buyers. The private exchange allows the publisher to leverage the direct relationship in the RTB environment, without replacing it entirely. In addition, the private exchange allows for negotiated floor prices and other special accommodations, further enhancing the traditional buyer-seller relationship.
Some argue that real-time bidding isn't always real-time. Is that true?
I've heard this criticism and it strikes me as odd. While traditional ad servers update their bidding rules and reports once or twice per hour, AppNexus is nearly instantaneous for changes to bidding and budgeting and the actual transactions are always done in real-time. We have seen that some buyers seem to work on predictable hourly budget updates, which is unfortunate because it allows for intra-hour arbitrage. Customers should ask their vendor whether this is the case.
How would you compare the US and European markets in terms of how quickly RTB is being embraced in the marketplace? Is there a problem with scale outside the US?
RTB works when there are liquid pools of demand and supply. In the US market this is already the case. In Europe, of course, there isn't a single market, but rather a series of local markets. And within those local markets the buying and selling pools tend to be concentrated within a small network of companies.
As a result, right now, there is less liquidity throughout Europe, and much of the trading consists of remnant and performance campaigns. I believe this situation will change within the next 6-12 months as the obvious technical and operational advantages of RTB take hold and some of the more advanced publisher techniques I've been talking about make local premium publishers more comfortable with making their inventory available.
What plans do AppNexus have for expansion into Europe?
AppNexus was founded in New York in 2007 and is led by the pioneers of the original ad exchanges at Yahoo!'s Right Media and Google's DoubleClick. The company has experienced dramatic growth since its founding with more than 120 ad networks in 12 countries using AppNexus' real-time ad platform.
Last year, AppNexus complemented its two US-based data centres by opening a data centre in Europe, to bring real-time bidding to the EU and the Middle East. In addition, we have just opened our first office in Europe and made our first hire in the UK, senior account director, Ajay Daved.
Daved will work with new and existing clients, including online advertising networks, agency trading desks and demand-side platforms to spearhead the growth of AppNexus in Europe. He is a great first hire for our European operations, which is a market we are really committed to growing. London will be our hub for Europe and we are recruiting for the rest of the European team.
How has Microsoft's investment in AppNexus impacted its strategy? What is the latest on its exchange?
Microsoft is a great partner and investor in AppNexus. The Microsoft Exchange is open for business in the UK and the Netherlands, with more countries to launch throughout 2011.