SEM-Like Control For Successful PLA Campaigns

There are notable similarities between keyword and PLA campaigns that, when leveraged properly, will give retailers greater control of their campaign management. Let’s evaluate how retailers can leverage best practices from their keyword campaigns and apply them to their PLA campaigns for maximum efficiency and control over their budgets.

In SEM, search engines decide which ads are shown for which queries based on keywords; but, advertisers don’t have to yield control of their spend to the search engine. Instead, they can institute a proper campaign setup that allows for control at a granular level.

Control Over Your PLA Campaigns

In order to achieve this level of control, it’s best to split your keywords into ad groups by unique match type. Small, tightly knit ad groups allow you to write ads specifically tailored to the match type you are targeting.

If you are targeting a broad query, you can write an ad that applies in a broad sense to the consumer you are targeting. If, on the other hand, you are targeting a very specific exact match query, you will have the ability to specify this level of detail within your ad copy.

Splitting match types into their own ad groups isn’t the only option for maintaining granular control. Another option is to split match types into their own campaigns. If you are an advertiser who requires management of your budget at the match type level, this is the right approach for you.

Marketing your PLA campaigns demands a similar approach to campaign structure. For PLA campaigns, Google maps user queries to individual product targets that retailers set up in their AdWords campaigns.

Setting Up PLA Campaigns

There are several approaches to setting up your PLA campaigns in Google. The easiest is to set up an All Products target, which will market a retailer’s entire inventory in their data feed. This setup is simple and only takes about five minutes to get going, but limits the ability to be relevant at scale.

Because ads are assigned to ad groups, and there’s only one ad group that covers all products, you can’t target different ads to specific products. Furthermore, the same bid is applied to all products because there is only one ad group. If all products are in the same product target, you cannot apply different rules to different performing product types. Therefore, you can’t have different bids for your high-margin or best sellers versus your generic, low-margin targets.

If you apply the same principles of granular keyword targeting (ad groups by match type) to your PLA campaigns, you can overcome the limitations of All Products targeting.

The ideal situation for retailers is to place individual SKUs into unique ad groups. Benefits of this setup include writing specific ad copy for individual products and making extremely powerful bids (details covered a bit later in the post).

It’s important to note that campaigns can only hold 20,000 ad groups, so you’ll need more than one campaign if you’ve got more than 20,000 products in your merchant feed.

Below is an example of how you can split a broad product target out into multiple product targets to make it more granular:

Example 1: Broad Targeting

Ad group – pants

Target a product_type =jeans

Target a product_type=slacks

Target a product_type=trousers

Target a product_type=shorts

Example 2: Granular Targeting

Ad group – levi strauss mens jeans

Target a product_type=jeans AND brand=levi strauss AND gender=mens

Target a product_type=jeans AND brand=levis AND gender=mens

Example 3: SKU-Level Targeting

Ad group – levi strauss mens stonewashed bootcut jeans size 32×24 id1234

Target a product_type=jeans AND brand=levi strauss AND gender=mens AND style=stonewashed AND type=bootcut AND size=32×34 AND condition=new AND material=denim AND demographic=adult AND color=denim blue AND product_ID=1234

 Example 2 has a substantially higher degree of granularity than Example 1 and allows you to write specific, targeted ad copy directly related to the brand, gender and product type. Example 3 actually takes into consideration all attributes and features of the product and gives you the most control of your ad copy development as well as the way in which you bid. At this level, you are able to bid more aggressively for your higher-margin items and best selling items.

Granular Control of Ad Copy Message

When you group keywords by match type, you’re able to offer consumers ads targeted to the broadness or exactness of the query.  When it comes to PLAs, the same principles apply. Therefore, there is no better option than creating ad groups based upon SKU.

In Example 1, you cannot write ads with any degree of specificity past the general idea of “pants.” If your store has jeans, slacks, capris and other kinds of pants, you’ll have to write one ad that covers all different types of pants and you cannot target each individual pant type your store covers.

Since there are no keywords to bid on in PLAs, creating ad copy carries a significant impact in helping determine relevance. Creating ad groups based upon SKU gives retailers the flexibility to target each product in their inventory with a greater degree of specificity. Also, you can leverage your feed to associate real world promotions with their products in a powerful and dynamic manner.

Not only can each ad copy contain detailed features of specific products found within inventory, but your ad copy can reflect real time promotions by enriching the data feed to include promotional details at the product level.

We’ve heard some retailers argue that ad copy for PLAs isn’t as important as it is for keywords because the ad often doesn’t even show unless you hover over the image. However, experienced search marketers will agree that relevance from query to ad is key to any successful PPC campaign.

With keyword campaigns, it is evident that the relevance from query to ad impacts quality score. Currently, the quality score associated with PLA campaigns exists, but it’s not yet visible to advertisers. Is anyone else expecting Google to weigh relevance of query to ad for PLAs when determining quality score? If so, you can be sure that the more granular the structure of your campaign, the more ability you will have to make your ads relevant at scale.

Granular Bidding At The Product Level

For retailers, bidding should be about driving bottom line performance for keywords and PLAs: different lines of products require different bidding strategies.

If you sell prom dresses, you focus your efforts on generating traffic because most consumers interested will try the dress on before buying, so they’ll likely research online and then buy in-store. For cheap watches, you will want to drive performance based on ROAS; but, luxury/expensive watches will demand you drive performance toward gross margin % or CPA.

ROAS may not be an effective measure for high-ticket items due to the proportion differences between the CPCs and AOV, causing too much volatility for your bidding algorithm to optimize against. Only granularly structured campaigns will enable a strategy-based approach that is applicable across lines of business, categories and products.


Optimal Product Target Strategy

Control Your Budget With Negatives

By splitting up campaigns/ad groups into match types, SEMs can better control their ability to show the ads they desire instead of letting the search engine decide on their behalf.

For example, let’s say a retailer is bidding on broad, exact and phrase match keyword [little black dress], and they split up each match type into unique ad groups. In this case, they can add exact match and phrase match negative to the broad match ad group and the exact match negative to the phrase match ad group.

By doing so, the retailer is given complete control of their bidding approach. Without adding match type silos, the search engine could serve the ad for any of the match types for the query [little black dress] and the retailer loses control of how ads are targeted.

Negative keywords for PLAs can behave in a similar way for retailers, ensuring they are retaining the control they need to spend their marketing dollars in the most efficient way possible. Make sure to negative out your SKUS from each other. By taking the various feature differences between your products and adding the opposing features as negatives, you can ensure Google picks the right product for the right user query.

An example of this is by negativing out [large] and [3.2 oz] from a 1.1 oz Burberry cologne. When using these negatives, you are preventing the 3.2 oz version from showing when the 1.1 oz version is the best candidate to be displayed within the SERPs. If you choose not to negative out the 3.2 oz version, you risk this bottle appearing with a significantly higher price near the competing 1.1 oz listings.


PLAs can save SEMs a lot of time — they don’t have to pick keywords and match types, or organize their keywords into ad groups. At the same time, many of the same principles in SEM apply to PLA campaigns.  If you want to have SEM-like control over your PLA campaigns, you need to deploy a highly granular campaign structure.  The more granular your campaign structure, the more control you’ll have over bids, budgets, ads, and keyword negatives — and the more successful your PLA campaign will be.

Opinions expressed in the article are those of the guest author and not necessarily Search Engine Land.

Related Topics: Channel: Retail | Google: Google Shopping | How To: PPC | Paid Search Column | Search Ads: Product Ads | Search Marketing: Shopping Search Marketing


About The Author: is Director of Customer Success at Adchemy® a software development company that helps marketers scale the performance and relevance of their mission-critical paid search campaigns.

Connect with the author via: Email | Twitter | Google+ | LinkedIn


Get all the top search stories emailed daily!  


Other ways to share:

Read before commenting! We welcome constructive comments and allow any that meet our common sense criteria. This means being respectful and polite to others. It means providing helpful information that contributes to a story or discussion. It means leaving links only that substantially add further to a discussion. Comments using foul language, being disrespectful to others or otherwise violating what we believe are common sense standards of discussion will be deleted. Comments may also be removed if they are posted from anonymous accounts. You can read more about our comments policy here.
  • ronniesmustache

    Really nice, Dan- thanks!

    I’m always surprised when I see companies using All Products targeting.

    I’m also surprised to see that they’re often still turning a pretty decent ROI for those campaigns. Go figure.

  • daniel morrison

    Thanks for the comment Mr Mustache!

    I’m surprised as well, but the market will start normalizing a bit soon enough and those folks that have granular targets are going to be so thankful they took the time to setup their campaigns properly.

  • ronniesmustache

    Yep- I agree.

    The sophistication of PLA advertisers is increasing very quickly. In February, most search professionals didn’t know what a PLA was.

    Now, you can’t do a product related search w/o seeing a full platter of PLA’s.

    Again, great article!

  • Feed Optimise

    or better yet, you can use – and get all done for you.

  • Lucas von Fürstenberg

    I am always wondering why everybody is promoting one SKU per adgroup. Why not having campaigns for product groups, adgroups = brands and individual product targets within that adgroup?
    Products within an adgroup are similar enough to write one adcopy (if you want to do this down to this granularity).

  • Martin Röttgerding

    Hi Dan,

    I agree that SKU adgroups are certainly the best way to set up a PLA campaign at
    the moment, especially in combination with automated bidding. Control
    over the ad copy is also a big advantage, but I have to disagree with
    the part about query-ad relevance. Quality score is about click-through
    probability (think: predicted click-through rate). Including parts of
    the query in the promotion text won’t help your CTR as much as something
    like “Free Shipping”. So write whatever makes most sense to attract
    customers and ignore keywords when writing ad copy.

    By the way,
    here’s how everyone can create their own SKU campaign without paying for

  • Martin Röttgerding

    Mainly because products that share both brand and product type are still not always similar in terms of price, margin, conversion rate, order value. etc. For example, you might have this year’s expensive Nike running shoes right next to last season’s shoes sold at a discount. Some merchants even put accessories into the same category (let’s say an extra set of laces for your Nike running shoes).

    There’s also the part about messaging. In the example above there’s a general promotion about free shipping on orders over $75. Now you might have some products within a category, let’s say Nike running shoes, that cost $69 and others at $79. A granular structure lets you promote free shipping on items above $75 right away.

  • daniel morrison

    +1 for Martin’s comment below. However, Lucas brings up a good point that I neglected to address.

    Leveraging campaign structure based upon your marketing / budgeting practices can be instrumental in helping achieve your business goals. Creating silos of product types and putting those into campaigns with similar products can help your reporting efforts and budget planning. However, if your budget is determined by brand spend, it might be beneficial to silo your campaigns based upon brand.

  • daniel morrison

    Hey Martin, thanks for the comments, very interesting dialog.

    I agree, in part, with your point re: quality score. Google determines quality score based on a number of factors, but pClick is only one of them. Other factors include historical performance of the ad group and landing page relevancy.

    For PLAs, Google accesses your product data feed and matches the user queries of consumers to relevant products within your data feed, there is arguably a high likelihood that Google will incorporate the relevance promotional ad copy as a weighting (or perhaps query to title / description within the feed?). To take it a step further, do you see a world where Google might consider Google Wallet, Google Trusted Stores or Google Seller ratings as having an impact on QS?

    Aside from quality score, advertisers should consider the entire eco-system of paid search when planning and executing their ad copy strategies. Google’s interests don’t necessarily align with the interests of retailers. Google gets paid per click, retailers get paid per sale. Writing ad copy that maximizes CTR might maximize clicks (Google’s interests), but may actually hurt a retailers CPA. Therefore, writing ad copy that incorporates the query/product consideration may help to qualify clicks and drive down CPA, improving bottom line performance.

Lastly, I’d like to commend you on your write-up on Certified Knowledge. This is a great how-to creating on State of the Art PLA campaigns; extremely helpful guide for any retailers that want to get started right away without the use of paid tools. That being said, there are some great paid tools on the market that offer additional features to help overcome obstacles retailers face, but it’s about finding which tools contain the functionality that will help meet your business goals.

  • daniel morrison

    Hey Julie, yes, promotional text is what I’m referring to.

  • Martin Röttgerding

    Hey Dan,

    No argument on automating PLA creation and management – we don’t do it manually either.

    As far as Quality Score goes, I’ve done a lot of research on the topic in general. I know a comment isn’t the best way to get this accross, but here goes…

    Google has this thing called Quality Score that they show next to keywords on a scale of 1-10. And there is another thing called Quality Score that’s used in the ad auction that actually determines ad position and click prices. Those two are very different, although Google (and most search marketers) like to mix up those two.

    To maximize their own revenue, Google has to rank ads according to their potential revenue. The potential revenue of an ad (from Google’s perspective) is determined by two factors: the bid per click and the probability that the click occurs.

    If you take Google’s official formula (Ad rank = Bid x QS) and compare it with the revenue-maximizing formula (Ad rank = Bid x click-through probability), you get the idea that QS has to be click-through probability. Only if you believe that Google is interested in maximizing revenue, of course.

    Now, click-through probability is not a factor that just falls from the sky. Determining this factor is all but trivial, especially when you have no directly related data to go on. This is why Google needs things like historical CTR of related elements throughout an account, or relevancy of query to ad.

    Anyway, the point is that QS is not some arbitrary value to score ads according to a subjective notion of quality. Quality Score is a means to an end (making money) and in order to achieve this end it has to be an estimate of click-through probability. Knowing this we can assess things we believe to factor into quality score.

    A part of the query mirrored and therefore bolded in the headline of a text ad increases the likelyhood of a click – no doubt a strong indicator of click-through probability. The same part mirrored in the promotion line of a PLA – why would that increase the likelyhood of a click? And if it doesn’t, why should Google reward it?

    As far as click-through probability on PLAs goes, I believe there are two main factors that influence whethere a user clicks on your ad: price and picture. It’s nice to have bolded product names, but the picture draws the eye much more than that.

    As far as Google Wallet, Google Trusted Stores, and Google Seller ratings go, we can apply the same process and check whether these impact click-through probability. Since there are shiny badges and stars I’d definitely say they have to be considered in quality score.

  • Lucas von Fürstenberg

    I do see both of your points. However as Martin mentioned further down, historical ctr within an adgroup also affects your quality score.
    When you include the size of those Nike shoes mentioned before you have 10 adgroups with essentially the same products. Why split the clicking (and the history) in 10 adgroups? I believe it is always a trade off between being super granular and joining those products that belong together. I believe for us the advantages of having a click history faster outweigh the disadvantage of having a potential sale item in a group with regular priced ones.
    Other than that: since the shipping cost is in the ad now right below the picture anyways, the significance of mentioning that in the adcopy has gone down quite a bit.

  • Jeremy J Brown

    Hi Dan, do you have any data to show that having targeted promotional copy (or any promotional copy for that matter) makes much of a difference?

    That should be something you could A/B test. In limited testing, we have not seen much of a difference in performance.

    I ask because as Lucas points out there are benefits to having 40 ad groups instead of 4,000.

    I agree that’s it’s often beneficial to bid at the Product Target level and we do that for a number of clients, but you can have multiple Product Targets in 1 ad group and each with separate bids. As long as your ad group structure makes sense, you should be able to drive strong results and bid appropriately.

    Overall, more than a few people are preaching the value of 1 SKU per ad group and it reminds me of the people who used to preach 1 keyword per ad group. Both can be detrimental to sensible management. Granularity can be taken too far when it interferes with management and actionable data analysis.

  • daniel morrison

    Hey Jeremy, sorry for the delayed response.

    We’re currently testing running different promotional text messages to evaluate differences, but don’t yet have conclusive results to share.

    You raise valid concerns around taking granularity too far, but technology can help with improving efficiencies. If you’re not leveraging automation, manual campaign management can be near impossible with 1 SKU per ad group, so you’ll need to develop a strategy that is functional and optimal based upon your resources / campaign management tools.


Get Our News, Everywhere!

Daily Email:

Follow Search Engine Land on Twitter @sengineland Like Search Engine Land on Facebook Follow Search Engine Land on Google+ Get the Search Engine Land Feed Connect with Search Engine Land on LinkedIn Check out our Tumblr! See us on Pinterest


Click to watch SMX conference video

Join us at one of our SMX or MarTech events:

United States


Australia & China

Learn more about: SMX | MarTech

Free Daily Search News Recap!

SearchCap is a once-per-day newsletter update - sign up below and get the news delivered to you!



Search Engine Land Periodic Table of SEO Success Factors

Get Your Copy
Read The Full SEO Guide