Sign up for weekly recaps of the ever-changing search marketing landscape.
Unstructured Data Brings Search Effectiveness To Display
Music lovers are able to buy individual songs without buying the whole album. Investors are able to buy individual stocks without buying a mutual fund or ETF. Search marketers are able to bid on individual search terms without buying pre-packaged bundles of terms.
Until recently, however, display marketers have been mostly confined to targeting audience segments, without the ability to identify and bid on the individual components within those audience segments. This is now changing with the advent in display advertising of targeting with unstructured data.
A Perfect Pair: Unstructured Data & Real Time Bidding
Targeting with unstructured data enables advertisers to bid, report, and optimize display campaigns using the underlying data that is too often aggregated and blended together into larger audience segments.
The practice of targeting against lager, opaque audience segments is a relic of the pre-RTB (Real Time Bidding) era, when the available technology and limited access to inventory made it difficult to use the more granular data.
RTB has changed the landscape of display advertising, unlocking access to vast pools of inventory while simultaneously allowing unique bidding on each individual ad impression.
However, advertisers using traditional audience segments are not fully reaping the benefits of the more-precise targeting now available. These include:
- Better performance − Targeting with unstructured data, advertisers can buy just the best-performing parts of the targeted audience, and lower bids or stop bidding on impressions that drag down overall campaign performance.
- Deeper insights − By seeing more precisely which parts of an audience are responding to an online ad and offer, advertisers can expand their targeted audience to similar consumers, and also improve the effectiveness of future campaigns.
- More flexibility − Targeting can be adapted on the fly, both by automated algorithms as well as by human touch. By optimizing more quickly during the course of a campaign, better results are attained faster, and waste is reduced.
Unstructured Data In Use Today
Advertisers today can take advantage of unstructured data targeting in several ways. Here are some examples:
- Keyword Level Search Retargeting − Impressions are targeted to users based on the individual term searched. Just like in search marketing, bids and ads are adjusted for the search terms based on performance. Another key factor, such as recency of the search event, are also taken into account, and ads and landing pages can be dynamically customized to the search term.
- Page or SKU Level Site Retargeting − Impressions are targeted to users based on the individual pages or products viewed, as opposed to lumping all site visitors into a single audience pool. Bids are varied for users dependent upon the value of the product viewed and/or how far down the funnel they have progressed. Dynamic ads and conversion paths are customized to individual products and/or pages.
- CRM Targeting − First party offline or online CRM data is matched to online users to target users who have purchased a particular product, service, or are otherwise engaged with the advertiser. This enables the advertiser to target their existing client base to up-sell, cross-sell, or otherwise build their relationship. By targeting to precise unstructured data, the advertiser improves effectiveness and learns more about users of individual products.
Old habits die hard, and targeting against opaque audience segments has been practiced in display advertising for a long time. Both performance and brand advertisers have much to gain by moving to better-performing and more-flexible targeting with unstructured data. Like in the music industry, technology has advanced and enabled a more efficient way of buying. When that happens, it is just a matter of time before the market follows.
Image used under license, courtesy of Shutterstock.com
Some opinions expressed in this article may be those of a guest author and not necessarily Search Engine Land. Staff authors are listed here.