The online advertising industry is witnessing a major confluence between 1) technology and media buying innovations and 2) the availability of robust data, including third-party data that is independent of ad inventory.
The last 24 months have seen the arrival of dozens of new data companies, advances in behavioral targeting and the emergence of demand-side platforms that allow ad agencies to control the bidding and buying of online ad inventory.
Though third-party data has created exciting new possibilities, it has also added a new level of complexity. Its role in online targeting is still taking shape and best practices are not yet well established. Making sense of it all is a challenge for even the savviest advertising professionals.
As you look to assimilate third-party data into your online targeting strategies, you need to fully vet your data partners. You need to know where the data comes from, if it's verified, how it will be measured and more. This brief guide will help you ask the right questions.
Introducing the Era of Independent Data
Five years ago, data that was independent of inventory was largely unavailable. Though some publishers collected audience data, using data to drive campaigns was largely the domain of advertising networks that employed retargeting pixels, tracked referring URLs and deployed pixels to understand engagement. Collectively, these techniques evolved into what we now call behavioral targeting (BT).
The last 24 months have given rise to a number of innovations in behavioral targeting, thanks to the emergence of:
- Technology-driven ad networks
- Real-time bidding
- Publisher yield optimization
- Next generation ad servers
- More sophisticated inventory management and pricing
In early 2009, the largest media buyer in the world – WPP's Group M – altered the Terms & Conditions attached to insertion orders and contracts by adding:
Not withstanding the foregoing or any other provision herein to the contrary, it is expressly agreed that all data generated or collected by Media Company in performing under this Agreement shall be deemed 'Confidential Information' of Agency/Advertiser.1
At the same time, dozens of companies have emerged to provide data independently of inventory. We are now at the beginning stages of a major confluence between 1) technology and media buying innovations and 2) availability of robust data from multiple sources and of many varieties.
Agencies Re-Gain Control
Advertisers and ad agencies alike have voiced serious concern about the collection and aggregation of data related to their campaigns. In response, agencies have taken back control of their campaign-derived data by bringing bidding and buying of inventory in-house. To do this, agencies are leveraging technology platforms, and in some cases, building sophisticated trading desks with teams that combine ad impressions and data; the emergence of third-party data that is independent of ad inventory has enabled agencies to leverage these platforms — increasing agency buying power, enabling dynamic pricing and delivering more detailed audience buying.
Targeting Consumers Where It Counts
Lately, it seems as though a week doesn't go by without a new data source announcing their arrival on the scene. While most media buyers and sellers will publicly state that more data is always welcome, the reality is that making sense of the data is a challenge for even the savviest advertising professionals.
Combining Data for Optimal Results
While marketers have an ever-growing variety of data to leverage throughout the marketing funnel, already-established data types — such as demographics and geographic data — continue to be relevant and should be combined with the data types discussed on the previous page for the best results.
Demographics — such as age, gender, marital status, presence of children, home ownership and income — can act as the foundation of audience models; other data can then be used to refine the audience. One of the chief advantages of demographics is that they are available for large portions of the Internet population where something like in-market data (e.g. individuals looking for a flat screen TV) is often available for only a very small segment of the population. It is also easy to incorporate demographics into audience models because they are the lingua franca of marketers and ad agencies.
Demographics act as useful proxies for life stages and interests. An individual’s life stage and interests are powerful drivers of purchase intent. In fact, demos serve as inputs to the models used to create intender/interest segments (but not in-market status).
Geographic data is also commonly used in ad targeting, and can easily be combined with other data types. In many cases, a marketer is only interested in reaching users in a certain area, such as US-only, a specific DMA, state or even a specific ZIP code. Geographic data can come from a variety of sources. There are companies who match IP addresses to geographic location, as well as companies who obtain user registration data to get more accurate geographic data.
How is Third-Party Data Integrated into Campaigns?
Online ad campaigns can integrate third-party data in a variety of ways, depending on who is doing the integrating. There are three key entities that currently operationalize third-party data.
Publishers and Ad Networks
These organizations incorporate third-party data into their own user profiles in order to a) create unique segments and b) deliver users who match specific data sets requested by ad agencies. Depending on the business model, the data available will either be transparent or opaque (and part of their data offering).
Demand-Side Platforms (DSPs)
The more advanced systems can layer data on top of an ad impression prior to bidding on that impression. The systems are able to ingest data from third parties and associate it with the user. Rather than creating user profiles that combine data from multiple sources as the ad networks do, DSPs map an available impression to a user who has specific characteristics that align with an agency RFP. This real-time association gives media buyers the ability to conduct discrete audience buys. Agencies are more frequently requesting that their campaign impressions be targeted utilizing specific data sets from named data providers.
These entities exist solely to make data available, completely independent of media and inventory. They buy, collect and maintain a rich marketplace of data types, including demographics, geo-data, brand affinity data, life stage attributes, purchase activity and behavioral data. Data exchanges sell these data points in real-time auctions; the winning bidder has the data transferred to them for use in ad targeting and optimization. There is an increasing trend in moving more branded data through these data exchanges, giving agencies and media buyers more power, visibility and transparency.
Emerging Best Practices
When you work with third-party data providers to augment your online targeting methods, make sure they are using the following best practices.
Attribution. The need to understand what drives campaign lift has given rise to "attribution" or the process of quantifying the data's impact on a campaign, as opposed to other factors. If a campaign generated an ROI of 175%, for example, attribution allows you to determine how much of that ROI is due to the utilization of any given data type. Media buyers and agencies are increasingly looking to platforms and data providers that give them insight into data performance.
Optimization. Data should be tested and measured within the construct of the campaign itself so that other variables — such as media placement and creative — are taken into account. Optimization techniques allow advertisers to do this by exposing the data on all ad impressions at the beginning of the campaign. After collecting a statistically significant sample, certain population segments will emerge as being more responsive than others. The ad server should then be instructed to target the high-performing segments and ignore the underperformers. The simplest way to optimize is to exclude underperforming inventory or identify segments that are non-responsive.
Combining Data Sets. In most cases, it may appear that different data types are competitive. In reality, advertisers have found that one plus one really does equal three. For example, a car manufacturer may want to market a new luxury SUV. It can use in- market data to target people who search for the new luxury SUV but the volume will be very low. It can target "Auto Intenders" but that may be too broad. Instead, it should target the in-market category with the most reach (i.e. auto intender) and overlay third-party data that categorizes users based on the type of vehicle they typically drive using an audience data overlay strategy. Using this methodology, the manufacturer can now target auto intenders who are likely luxury SUV drivers.
The 10 Questions You Should Ask When Evaluating Data Providers
Third-party data is still in its early stages of development. While it has the potential to give your campaign results significant lift or ensure your message is reaching the right audience, you need to fully vet your data partners. Below is a list of key questions you need to ask yourself and your partners before taking the plunge.
|Question to Ask||Implications|
|1. Where does the data come from?||Ample data about your audience is worthless unless it is reliable and accurate. Find out if the data is collected online or offline. Does the provider collect its own data or obtain it from other parties? What type of other parties are involved? Is there a long-term agreement with other parties and does the partner have exclusive access? Does offline data come from sources that have direct relationships with the consumer? Is the data user submitted or inferred?|
|2. Is the data verified?||You need to have a certain degree of confidence in your third-party data. Find out if it is verified at some point or if it is used regardless of the level of inference. If it is verified, ask how, by whom and how often.|
|3. How is the data delivered?||Be sure to ask how the data is connected to either cookies or a delivery platform. Is there a cookie-matching methodology that utilizes pixels and/or java script in the browser? Does the provider use unique IDs to match user's offline? Is there server-side delivery?|
|4. How regularly is the data refreshed?||Depending on the variety, data can go out of date almost before it goes into use. What's the point in targeting your campaign based on out-of- date data? Be sure to ask how often data is refreshed.|
|5. Are you dealing with stable data?||The integrity of the data is extremely important whether you are driving intent or building a brand. The ability to leverage the same data in subsequent campaigns strengthens the value of the information. Avoid data that is transitory so that you can continue to use it over a period of time. Demographics, lifestyle segments, brand preferences and attitudinal data tend to be stable. On the other hand, in-market data is very short-lived, i.e. transitory. If a user is in market for a TV, they are likely to make a decision within a relatively short time frame.|
|6. Is the data easy to use?||Your data partner may have the most accurate, most scalable, most predictive data set available but if it is not easy to ingest, then it cannot add value. Make sure your data partner has a clear understanding of your processes and platform so that you get the maximum benefit.|
|7. Is your partner built to last?||If you are going to build your strategy on third-party data, you should be sure that the provider is going to be a long-term trusted partner. Will the company be around next year? Does the provider have your best interests at heart? Does it show signs of growth? Is it well funded? Does it have reputable partners?|
|8. What privacy protections are in place during the collection and delivery process?||With the spotlight on Personally Identifiable Information (PII) — from Congress, the FTC and privacy advocates alike — it is imperative that everyone involved understands the methods of data collection. While full disclosure of data collection and distribution is now standard, there are additional questions you should ask. Is the provider a member of NAI and/or IAB? What are the public perceptions of the company? Is the company currently under public scrutiny by privacy advocates or federal regulators? Can the provider formally state "our data cannot be tied back to PII or associated with PII"?|
|9. Can you continually test results across a range of campaigns?||Data attributes may test well in a lab or on a single campaign, but consistent performance is what matters. Performance can be defined simply as reaching the desired audience for an awareness campaign or as conversions for advertisers looking for customer acquisition. Continual use of your audience over a spectrum of campaigns is where true results, or lack thereof, are exposed.|
|10. What is the data provider's attribution methodology?||With so many factors going into driving campaign performance or non- performance, it is important to be able to pinpoint the attribute that is driving the greatest lift. Be sure to ask how your data provider isolates its data when measuring its impact.|
The influx of third-party data providers and data types will only continue. As with any dynamic marketplace in the early stages of its lifecycle, new entrants will bring new ideas, technology and capabilities to the marketplace. The result will be better performance from all data providers, which in turn, will increase online advertising effectiveness.
Regardless of your role in the display advertising ecosystem, you can derive benefits from incorporating third-party data into your business model. For publishers and ad networks, key benefits will include the ability to confidently respond to RFPs and deliver audiences at a higher CPM. Demand-side platforms will enjoy the ability to intelligently bid on inventory. Advertisers and ad agencies will benefit from the ability to clearly define desired audiences and measure delivered audiences.
Before entering into any partnerships with third-party data providers, you must do your research. Be sure to avoid any pitfalls by asking the right questions and following the recommendations outlined in this brief guide.
1. "GroupM Revises Terms For All Online Ad Buys, Claims Data is 'Confidential'", by Joe Mandese, MediaPost, February 2, 2009