All posts by Donella

Are You Ready For Google Shopping Campaigns?

Google plans to migrate all remaining legacy Product Listing Ads (PLA) campaigns to Google Shopping beginning in August 2014. Here are some questions to ensure you’re ready.

What are the differences between the old PLA campaigns and Shopping Campaigns?

The Ads – Shopping Ads are still PLAs. The big difference is that now you don’t have to know precisely what is in the Product Feed to create valid groupings. AdWords now connects your Merchant Center feed to your AdWords account in a way that lets you drill down into Google’s category taxonomy to create Product Groups, or you can use other feed attributes such as Condition or Brand, or you can use custom labels to define your target set of products.

Product Groups = Product/ Targets – Each AdGroup in a Shopping Campaign consists of one optional ad (this is the promotional text that will appear with every product ad served from the group) and one or more Product Groups which determine the eligibility of products to be served based on more easily-defined parameters. Don’t let the rebranding of these campaign assets throw you off.

No Keywords – These are not keyword campaigns. They will behave differently and you will want to experiment to find the best way to structure your Product Groups. A simple start should include using Google’s product category taxonomy to group like products. Next break those groups down using additional attributes, such as Condition or Brand, or your own custom feed attributes. You can, however, add negative keywords to your campaigns to avoid unwanted impressions.

Google has provided several learning opportunities for marketers, including basic and advanced webinars:


Are your product feeds optimized for Shopping Campaigns?

Where the old PLAs used adwords_grouping and adwords_label fields for organizing your product targets, Shopping Campaigns introduces the availability to create up to five custom labels that can be used in conjunction with Google product category, brand, item ID, condition and product type to create Product Groups to be targeted.

These attribute filters can be combined and stacked to tailor your feed to meet your needs. Start with the highest level of categorization and break the category down to more granular levels. Your Product Groups are bid at the most granular level specified.


Do you have a plan for measuring the performance of Shopping Campaigns and comparing them to old PLA campaigns?

Gathering benchmarks now will help you see where differences may exist between Search Product Listing Ads campaigns and the new Shopping Campaigns.

Fortunately, Google has introduced some additional impression share and benchmarking metrics that will help you understand your place in the marketplace, too. Benchmark max. CPC and CTRs gives you an idea of the performance of similar product ads in the marketplace.


Is budget set aside for your Shopping efforts, or is it lumped into your Search budget?

If at all possible, segregating your Shopping budget from your overall Search budget can help ensure you have more time and freedom to experiment with the new Campaign type. CPCs have increased as more and more retailers have adopted product listing ads and it will be important to be cognizant of how those changes will affect your overall marketing spend.

Attribution Is Where It’s At

In a session on interpreting multiple data sets at ad:tech NYC, a panel was asked “What data set or point is at the top of your wish list?” The answers were not surprising:

  • Individual psycho-behavioral information
  • TV viewer behavior and conversion
  • Attribution

Attribution was one of the hottest topics at ad:tech. Everyone acknowledges that marketers have made great strides in using attribution to fine-tune their digital advertising strategies. But the Holy Grail of accurate cross-channel, cross-platform, cross-device attribution remains elusive.

Why is attribution so important? The simplest answer is efficiency. Understanding what spurs conversion helps marketers mold their strategies for maximum ROI. There is nothing simple about attribution, however, in a world where the classic purchase funnel has become more of a maze.

Credit where credit is due

Several of the panelists at ad:tech warned of the fallibility of last-click attribution models. Generally, last-click/last-view models favor search. What marketers really need is the ability to give proportional credit to all marketing channels and to understand how cross-channel and intra-channel assists contribute to conversion. While technologies can help marketers gain insight where tracking is possible, no one has been able to address the biggest challenge of all – cross-device attribution. Until someone cracks that code, analyzing behavior across all of the devices consumers are using is more art than science.

So much media …

As the list of devices that consumers use grows, so does the list of publishers and media outlets. Today’s campaigns run across search, social, display and TV, with multiple permutations within each of those channels. So how does a marketer know what really works?

One suggestion from ad:tech panelists was to keep a good marketing calendar and make sure that calendar is shared across media planning and analytics teams. By overlaying time-based data on an up-to-date calendar comprising all elements of a campaign, marketers can get a sense of causality that the numbers alone can’t provide.

Another tip offered was to compare multiple attribution models rather than putting all stock in one methodology. In one case study presented, analysis found that by using a last-click attribution model, only two of the tested creatives were winners. When assists were considered, though, five of the creatives contributed to conversions. Especially in testing creative and placement, considering alternate attribution models can keep marketers from throwing out good creative.

Other presenters encouraged marketers to be bold and creative in their testing: “turn off your branding campaign in order to measure attribution; bring offline data into an anonymous cookie pool to sync with online data; test for causality beyond conversion; explore new data sets,” they said.

Not without challenges

Attribution is a fairly new science, so, of course, it’s still evolving and will need to continue to adapt as the landscape changes. Some of the challenges that technology companies and marketers will be trying to overcome in the near future include:

  • Four-screen attribution – Single users on multiple devices create the biggest challenge in positively identifying “what works.”
  • Audience verification and cookie management – The lack of tracking capabilities for mobile and TV prevent an accurate account of how those platforms contribute to campaign success.
  • Over-abundance of data – It would seem logical that more data is better, but that’s only true if marketers understand which data really matter and don’t get buried in an avalanche of numbers that prevents them from reaching meaningful conclusions.
  • Protecting privacy – Consumers are more wary of tracking tactics than ever, making it all that much more important for marketers to be mindful of protecting their customers’ privacy.
  • Staffing for the new world – In an environment where CMOs are projected to spend more on technology than CIOs, the need for new roles is becoming evident. Agencies and marketers will need to have someone on staff to evaluate the diversified and specialized landscape of tech providers to create the right tech stack for each client and campaign, and make sure that these disparate solution providers work together.

Attribution is guaranteed to be a focus of ad tech and data providers as they strive to create more sophisticated models and overcome technical challenges. Marketers who aren’t paying attention are missing out on one of the best ways to really understand their customers’ behavior.