The back history is that the datafeed providers who fed to the comparison shopping engines (CSEs) also fed to the free Google Product Search program (which subsequently became PLAs). Most of the datafeed companies charged on a revenue-share model, so when investment-free product search sales were combined with CSE sales that had media expense attached, the ROI of the program holistically looked very good. It was likely quite profitable for the datafeed companies, too, who were collecting rev share on a part of the program that had no expense against it. When product search ceased being free and morphed into PLAs, the campaign management was moved from Froogle into the AdWords interface. This became the basis for the land grab since SEMs didn’t want the feed companies sharing access to their AdWords accounts, and they also didn’t want to lose the media spend for these programs. The fear over this latter point escalated when PLAs began siphoning off a large portion of traffic from non-brand paid search keywords.The datafeed companies argued (fairly IMO) that SEM agencies were not adept at datafeed management, and the SEMs stressed (also fairly) that paid search art and science is a different competency than managing structured data.
Ultimately, this situation spawned a new third party which brings us to where we are today. This new category consists purely of structured data providers who inexpensively ensure that the datafeed a) exists; b) is optimized properly for each media provider; and c) provides updated feeds to the media providers in regular frequency. This third party probably doesn’t support any work from the first-generation datafeed companies who have their own internal expertise, but they serve as a great resource for the SEM agencies and/or the marketers themselves. Some offer fully managed service, but most serve purely as backed tech support.
A requirement in many new business RFPs is that the agency be best-in-breed in their core competency. To this point, for most retailers, the management of PLAs squarely belongs with the SEMs considering how similar the PLA strategies are to paid search and the fact that reports from the search interface such as search query and campaign and ad group negatives inform the tactics of each.
Depending on the particular retail vertical, the first-generation feed companies are still best-in-breed for actual datafeed production if the products sold are the same as those available at multiple retailers such as electronics, books, tools, CPG, etc. The first-generation feed providers have large databases containing manufacturer product numbers which is required, standardized data by all the media companies. A retailer can send a rudimentary feed of merchandise, and these feed companies can populate all the missing MPN information. This requirement has often been a thorn in the side of retailers, so this is a useful service.
While the introduction of structured data media programs caused some SEMs to scramble initially, these programs have, in fact, created large efficiencies for agencies who used to have to manually duplicate all the products that are (and always have been) in retailers’ datafeeds and turn them into individual keywords to launch on paid search. Making data quality optimizations to the feed once it’s been uploaded to the search engine interface is far easier for SEMs once getting past the learning curve.