Is the publisher getting hammered on pricing in the real-time market? The concerns around cherry-picking of inventory are real. The sales channel conflicts are real. Publishers want to put more inventory through RTB - but worry what effect it will have on their bottom line. Pubs are rightly hesitant. But some are developing robust trading strategies to ensure they are not putting downward pressure on eCPMs when selling through automated channels. Data is fast becoming a critical piece in managing pricing, and valuing RTB impressions at their true value. Here Jay Stevens, VP & GM, International at The Rubicon Project, talks about the critical role data plays in RTB pricing, data as a revenue channel, and how publishers can compete with the buy-side.

Data has been the subject of much debate over the past twelve months - particularly in terms of how data can be used to boost CPM pricing. Looking at the issue from a RTB perspective, how is Rubicon enabling publishers to leverage their own data to get higher prices for RTB impressions?

JS: Publishers looking to monetise their inventory through the RTB channel and get the highest prices possible for RTB impressions need to look for a partner with a suitable solution that addresses their needs and concerns in that marketplace, such as: pricing controls, analytics to maximize yield, protection against data leakage, automation to save time, dedication to ongoing development and support. Most importantly, publishers need a platform that can manage all inventory, not just that sold in RTB, so all inventory can benefit from the pricing intelligence gathered through the RTB process.

Real-time bidding offers a wealth of information that can help publishers understand the demand for their inventory. Therefore RTB should not be treated as an independent sales channel, but as a refined, highly-targeted channel, where publishers are selling individual audience units, as opposed to a content channel, across which selling tends to be less efficient and prone to media waste. The ability to offer this new level of targeting allows maximum yield. Therefore, RTB should be incorporated into the broader spectrum of liquidity, as revenue lift is a result of both CPM and fill rate.

Not every impression is going to fetch a good price, or even be sold at all, on the RTB market. Direct sales are still a key channel for most publishers and will remain so, as are ad networks, exchanges, representation firms in markets outside their own. That said, incorporating RTB competition at each level of inventory will drive revenue across the board for publishers by lifting CPMs not only from real-time bidders who buy a percentage of inventory, but from all the different buying channels that together fill 100% of a site’s inventory. While there are risks, these can be mitigated by the right technology, proper sales channel management controls, and the right technology partner to meet the publisher’s strategic revenue goals across all inventory.

Data leakage is a key concern of publishers in terms of opening up their inventory in real-time. How is RP helping publishers understand and address this problem?

JS: REVV Protected RTB addresses the data leakage risk for publishers in two ways: first, technology uniquely encodes REVV non-PII user ID’s and auction ID’s differently for each RTB demand partner in order to prevent the merging of user data across multiple DSPs.

In addition to the technology protection provided by the unique encoding of unique identifiers described above, data leakage must be addressed by contractual controls and measures. Publishers should require that their RTB platform offers complete visibility and control over which demand partners can sell their inventory, at what level of transparency and at what price. In addition, a strong contractual relationship (in this case, between the buyer and the publisher’s monetisation platform) can further restrict the usage of any data that an RTB demand source is exposed to. Contractual protections that restrict the re-use of user and auction ID data are another way the REVV solution addresses the issue of data leakage.

Do you think that publishers are still not compensated properly for the amount of data that is passed to buyer in a real-time buy?

JS: There’s a lot concern in the online ad ecosystem around the disparity of data between publishers and advertisers. Historically advertisers have been incentivised to understand the audience behind the publisher in order to maximise their yield. This has led to a large investment in technology to serve the advertiser, tilting the balance of information to their favour and leading to rampant arbitrage. As a result, most impressions are still sold without adequate protection.

Real-Time Bidding affords the publisher an opportunity to rebalance the terms of the sale of inventory. In fact, in many ways the savvy, well-armed publisher can leverage RTB to gain an advantage in that each impression is evaluated and assigned a value by several buyers. During an auction only one buyer can purchase the impression, but the data gathered from the other buyers is not discarded, rather it is captured by the publisher (or in our customers’ case, by REVV) and stored for analysis and pricing determination on all future bids.

Is this why a lot of premium publishers have been reluctant to offer inventory in real-time?

JS: Historically, as publishers worked with ad networks to monetise their unsold inventory, there was a slight degree of data leakage, as ad networks use machine learning algorithms to determine which audiences on what sites perform for which advertiser. However, with the rise of real-time bidding, the opportunity for data leakage becomes a much more substantial threat to publishers. This trading mechanism is inherently more transparent and instead of black-boxed within the ad network, the data is closer to the source of budget. This is why we at the Rubicon Project require our DSP partners to adhere to contractual protections that restrict the re-use of user and auction data.

What role is analytics playing in guiding real-time inventory prices? How is RP delivering this kind of business information to publishers?

JS: One way we offer this kind of business intelligence is through auction simulation. Using a Hadoop cluster and proprietary algorithms, the REVV platform simulates the auction at different price floors, usually at penny increments between five cents and 20 dollars. During the simulation the system records the revenue gained for each impression whether it goes to the highest bidder at the second price, at the floor price or even if the floor prices out all bids and the impression goes to a non-RTB demand source. After the simulation another algorithm runs through the revenue at each floor and determines at which floor price the revenue is maximised. These simulations are run at different levels of audience granularity and the resulting, boiled data is stored.

REVV pricing intelligence algorithms draw on this data whenever a publisher has configured a segment of their inventory to use a dynamic floor. The publisher has the option to choose a price range for that inventory and REVV responds by providing the optimal floor within that range. If the publisher doesn’t choose to set their own price range, REVV will make a price range recommendation based on the bid price and volume distribution. Either way, new floors are constantly set and re-set for the inventory in response to changes in demand.

These Dynamic Price Floors are one of the ways that the REVV platform empowers the publisher to claim higher yield from their inventory, and reclaim balance in the marketplace.

Rubicon is looking to offer a holistic approach to inventory management and pricing. How can this business intelligence ensure that the best price is being paid for guaranteed, non-guaranteed or real-time impressions?

JS: Through granular pricing controls, publishers can set the price floors, on an advertiser by advertiser basis. This feature, often called private ad slots, enables the publishers to mitigate any erosion of their rate card, and opens up the opportunity for more sophisticated management of third party channels. However, RTB is not a silver bullet, and while CPMs are certainly higher than the yield seen from ad networks, often at 2-3x the price, DSPs are taking a thin slice of the inventory.

Our philosophy is that real-time bids for inventory needs to be analysed against all other sources of demand, both guaranteed and non-guaranteed as there are cases where an advertiser through RTB will yield a higher price for certain impressions that have been sold direct, and likewise, there will be cases when an ad network is willing to pay more for a user than the yield from a DSP.

Are we likely to see a third party data market emerging in Europe? What’s Rubicon doing in this area? Will buyers be able to buy third party segments for Rubicon for real-time buys?

JS: Third party data providers are already emerging in Europe, but largely taking a wait and see philosophy before they commit fully to regional expansion. The potential implications of the ePrivacy Directive which will go into effect in May could raise serious blocks to that market’s growth and development. And while we are extending our existing third party data partnerships and integrations from the US into the EU, we, like almost everyone in the industry, are trying to understand the full implications of this legislation on a market by market basis.

What kind of trends are we likely to see in terms of how publishers work with data buyers and how it will inform their sales strategies?

JS: One of the key features of bid landscape reports are that they provide the publishers with the intelligence they need to understand what an advertiser is willing to pay, on a user by user, impression by impression basis. As REVV is integrated with the same data partners that the leading demand side platforms are, publishers on REVV have access to the same intelligence as the advertiser, able to carve out high value segments of inventory, and raise price floors accordingly for those particular audiences, maximising overall yield.

For years publishers were strip-mined of data without being properly compensated by their ad network “partners”. Do you think publishers are now in better a position today in terms of how they manage and leverage their own data to unlock true monetary value?

JS: Absolutely. As advertisers submit their bids for each impression, the level of intelligence that can be gleaned and utilised by the publishers to understand the value of each audience pocket is something never before seen. This actionable insight gives them unparalleled knowledge of their audiences the capability to set corresponding floor prices.