Two Customizable Shopping Structures to Fit Your Accounts' Needs
Written by Zach Bruner, Account Manager
There are multiple approaches and strategies to organizing shopping campaigns for PPC. All of these will vary greatly depending on the brand, the number of products, the price of products, the type of products, feed organization, and the list keeps going. However, we have narrowed account-wide shopping structures down to a couple of options that are easy to implement. Again, these aren’t the only two options, however, they promote a universal adaptability while still offering flexibility (depending on the account’s needs). Once establishing a certain structure, you could then fine-tune and customize it to specifically meet your account’s needs.
(Disclaimer: This is not a whitepaper about organization within a shopping campaign, but rather, it is about how you should organize your campaigns to work together within an account.)
Before getting started, it is important to have a good understanding of the backend in shopping campaigns. Even if you feel confident in your knowledge of shopping and how your campaigns are set up, at least glance over The Complete Google AdWords Shopping Campaign Settings Breakdown. While some aspects are covered below, that specific post provides a more thorough explanation.
Option 1: A Tiered PLA Structure
The first of shopping structures is centered around organizing through priority settings and product tiers. In this case, we will have three tiers that align with the three campaign priority settings for shopping: High, Medium, and Low.
Defining The Tiers
The primary concept of this organization strategy is breaking everything out into tiers based on campaign priority. Again, those priorities are High, Medium, and Low. We will use these to group our campaigns by Top Performers, Product Groups, and All Products. Note: substitutes for Top Performers could be high priced products, high profit products, top-converting products, etc. However, for the sake of this example we’ll use Top Performers as the filler.
You can define Top Performers in a variety of ways, but in this case we’ll do so at a product level based on those products that lead in conversions. More often than not, your top-converting products will also be responsible for a higher amount of impressions and spend. If you have a smaller inventory of products, you can have these in one campaign. To do this, simply add them in by their Item ID number when creating the ad group (unless you have a custom label in place that denotes this for you). Then, name your campaign something along the lines of “Google_Shopping_Top Performers”.
However, if you have a larger inventory, you can have multiple Top Performer campaigns. Just be sure to group them with like themes for organizational purposes at the campaign level, with a similar Item ID setup at the ad group level. You can also add campaign-level negatives to prevent cross-pollination across your Top Performers campaign. For example, you could group top performers by product types with naming conventions similar to this.
- Google_Shopping_Top Performers_Boots
- Google_Shopping_Top Performers_Sandals
- Google_Shopping_Top Performers_Sneakers
- Google_Shopping_Top Performers_Dress Shoes
Or, you could group them by Brand, if that makes more sense with your inventory. For example:
- Google_Shopping_Top Performers_Ariat
- Google_Shopping_Top Performers_Birkenstock
- Google_Shopping_Top Performers_Nike
- Google_Shopping_Top Performers_Tommy Hilfiger
In either case, you should give these a high campaign priority, their own budget, and aggressive bidding.
Product categories are your middle-tier campaigns. In most cases, this tier will have the most campaigns/products and each must always have a medium campaign priority. In theory, since you’ve now segmented out the Top Performers, these campaigns will allow those middle tier products to receive more traffic. Remember that your Top Performers may account for the most impressions and spend, so this structure allows you to better allocate budgets and ensure a wider variety of your inventory is receiving traffic. If you have a large inventory or your Top Performers have historically accounted for the majority of your shopping traffic, you may want to also exclude them from the Product Categories campaigns.
Again, depending on the account, feed, and inventory, these categories will vary. However, it is common practice to have a Product Category campaign present for all categories in the inventory. For example, if you sold sports equipment, the campaigns could look similar to this.
- Google_Shopping_Product Categories_Basketball
- Google_Shopping_Product Categories_Baseball
- Google_Shopping_Product Categories_Football
- Google_Shopping_Product Categories_Soccer
To reiterate, you’ll (more often than not) want to exclude those Top Performer IDs from their respective Product Category campaigns. For example, if a certain Nike Basketball is a Top Performer, that ID should not be in the Product Categories_Basketball campaign. However, it isn’t the end of the world if you forget this step (often times you may intentionally ignore this step, in fact). If you were to keep the item ID in both, that certain Nike Basketball would not show in the Product Categories campaign until the Top Performer campaign ran out of budget. Again, Google will show an ad for that basketball in the campaign with the highest priority (even if the bid is lower). But, if that high priority campaign is out of budget, it will default to the next lowest available. So, it really comes down to advertiser preference. However, here are some reasons why you should consider excluding.
- It allows better control of the budgets for Top Performers (whether that be increasing or decreasing).
- It prevents those Top Performers from stealing traffic from the middle tier of the inventory (the whole point of this structure is to ensure exposure across all products).
All Products make up the final tier in this type of account structure for shopping. Typically, you will only need one campaign for this. It will be set on a low campaign priority with a much lower bid than your Top Products and Product Categories. In fact, it is both common and ideal to have an All Products campaign in any shopping structure. Here are a few incentives for having this campaign type present.
- It is a catch-all for those items that you haven’t specifically targeted yet or those items other campaigns are not targeting.
- A safety net to show high-traffic products when higher-tiered campaigns run out of budget. In this case, you wouldn’t have to worry about Item ID exclusions since the priority and bids are so low.
For this type of campaign, you can simply target “Everything else in All Products” at one low bid. There is no need to break down by Item IDs or Product Types. Also, this is the only campaign where you should ever be targeting “Everything else in All Products”. It should be excluded in all other campaigns in order to prevent double-bidding, cross-pollination, etc.
There are two final notes for the All Products campaign.
- You should have an All Products campaign in just about any PPC account with Shopping Ads. It doesn’t matter what your structure is, they are a great catch-all to have in place.
- You should also consider breaking the All Products campaign out by device. In other words, three bottom-tier campaigns, each specific to desktop, tablet, and mobile. This can be a good option if you’re noticing poor performance at any device level, the most common being mobile. As advertisers, we know the importance of mobile, even if the return isn’t there. By segmenting these campaigns by device you place a different emphasis and scale bids more efficiently.
Putting It All Together
Since we have determined what each tier consists of, here is a visual to help put everything into perspective.
In summary, this structure allows you to pull out Top Performers and put them into their own campaign. This helps you be more competitive with bidding since the return is there (or should be, ideally). Plus, it helps you keep a more accurate pulse of those products. Then, by pulling out those Top Performers, you can now drive more traffic to the middle tier of your inventory with the Product Categories campaigns. Since the Top Performers aren’t consuming all of the traffic and budget, other products now have the chance to get a higher quantity of traffic. Finally, the All Products campaign will act as a catch-all. Again, it should have a low priority and a much lower bid than the other campaigns. It will cover both untargeted products and act as a secondary option for when other campaigns run out of budget.
Aside from the normal bid changes, search query reports, audience layering, bid adjustments, etc., this type of structure also allows you to have ongoing cyclical maintenance. In other words, you should be moving item IDs and product groups around based on performance. Campaign builds are static, however, your inventory and product performance are not. Here is a visual of this ongoing optimization strategy.
Lower Than ROAS Goal
This type of structure won’t always be perfect. So, when you see certain Top Performers fall below your desired return goal (to the point that slightly lowering the bid won’t fix it), you need to move it back to its respective Product Category campaign. This allows more budget to go to other Top Performers and allows the product you bumped to the middle tier to still show at a lower bid.
Then, if that same item ID/product continues to be below ROAS goal (or even the product group for that matter), you must then exclude it from the middle tier as well. At this point, you are still showing for those products, but at a much lower bid. This allows you to still show for that part of your inventory while making better use of the shopping budget in other areas where there is a higher return.
Above ROAS Goal
On the other end of the spectrum, you can also you use this cyclical optimization pattern on high-performing products or categories. For example, if you pull a report that shows 75% of the conversions in your All Products campaign are coming from one product or a group of products, make sure you are showing them in higher-tiered campaigns. If you are already showing them, maybe it’s time to adjust budgets, bids, and targeting. If you aren’t, then you will obviously want to add them in. You can follow this same strategy when analyzing both All Products and Product Category data.
Option 2: Branded Segmentation
When considering the dynamic and seemingly uncontrollable nature of shopping, it makes sense that one would be a little skeptical. In general, shopping is more often than not out of our complete control. However, there are definitely ways to sculpt shopping traffic, especially from a branded vs non-branded perspective. With this type of structure, you can use priority settings, bids, and negative keywords to decipher the branded intent of your users, and then segment accordingly.
Does It Make Sense for Your Account?
As a preface, it is critical to note that this type of structure can be a fairly time-consuming initiative, especially if your account is working with a large inventory and multiple campaigns. So, before you begin, make sure that it makes sense for your account. The good news is this check is a pretty simple process.
To do this, you’ll need to pull a search query report for the shopping campaigns in your account. I recommend a date range of last 30 or last 60 depending on the traffic volume (or even the last year if you are seasonal or have low volume). Once in Excel, create a new column that denotes “Brand” or “Non-Brand” for each query. (Tip: Streamline this by adding a filter for queries containing branded terms, including misspelled variations. This will allow you to denote “Brand” or “Non-Brand” much quicker.)
After each query has a brand/non-brand association, you’ll need to create a pivot table with the data. Here’s an example of how to organize your pivot.
This will allow you to view the brand vs non-brand query breakdowns at the campaign level. From there, if the data doesn’t differ much between your branded and non-branded queries, then it probably isn’t worth implementing in your account. However, if you see a clear difference then keep reading!
Tip: PPC Hero has multiple posts on SQRs and pivot tables, so check them out if you need more help with this part.
Before we get to the actual builds, let’s cover how the setup actually works. The segmentation essentially comes down to negative keywords, priority settings, and bidding strategies. Here is a visual representation:
As we know, you cannot target keywords within a shopping campaign, but you can use negative keywords to prevent certain queries. To begin this process for the non-branded campaigns, first create a list of branded terms and misspellings of those terms. You can check search query reports for common misspellings of your brand. Typically, you should create these negative keywords as phrase match in order to achieve broader coverage. You can also include specific product names if it makes sense for your account. A negative keyword list of general terms isn’t necessary for your branded campaigns. However, you can include an exact match list of general queries as negatives in the branded campaigns for additional security. You can retrieve these by doing a basic search query report for the campaign. Again, not necessary, but an option if you want to feel more in control.
Campaign priority settings should be medium for brand and high for non-brand. Your first thought may be to place a higher priority on branded campaigns, however, it is in fact the opposite. By placing a higher priority on non-branded campaigns, it allows general queries to filter through the non-brand campaign first (as non-branded queries are more difficult to control). Then, if the query is branded, it will bypass the non-branded campaign and filter to your branded campaign. This is why a negative keyword list for the branded campaign is less important.
More often than not you’ll want a more aggressive bidding strategy for the branded campaigns and a more conservative strategy for the non-branded campaigns. However, this will vary depending on performance and account goals. Here are some likely scenarios to consider.
- Maybe brand has strong performance and non-branded spends without converting. (Push brand)
- Maybe it’s a slow point in the year and you just want to focus on branded efforts. (Push brand)
- Maybe it’s the holiday season and you want to focus on customer acquisition and account growth. (Push non-brand)
Make sure you align your bidding strategy with the current account objectives.
You can have as many or as little campaigns as necessary depending on the size of your account. Here is an example of how your shopping campaigns would look once everything is fully segmented.
In this example, “Campaign 1_Branded” and “Campaign 1_Non-Branded” would house identical products since the segmentation is based on branded intent. Any additional campaigns would follow this same principle.
Top Performer Campaign
Since you’re placing a high priority on the non-branded breakouts, it is common practice to abstain from a top-performer campaign. By creating branded/non-branded segmentation for your shopping campaigns, it offers enough space for continuous optimizations and optimal account performance. However, you could still include a top-performer campaign if it is important in your account. Just be sure to place a high priority, bid aggressively, and exclude those products from their other respective campaigns. You can exclude them by using negative keywords as well as manually excluding the product IDs within the product groups. If you do go with this route, however, you should be extra vigilant of campaign performance and ensure you’re sculpting traffic as intended.
Finally, be sure to have an All Products campaign present. Again, if there is one thing that should be consistent in every PPC account with shopping, in any structure, it is an All Products campaign. This campaign should have a low bid and low priority setting. Its purpose is to act as a catch-all for those products you aren’t specifically targeting, as well as a safety net for when other shopping campaigns run out of budget.
To begin building this structure, be sure you meet the following requirements:
- You already have shopping campaigns present in your account.
- You understand how to use AdWords Editor (If you don’t, there is a ton of great content on PPC Hero about it)
One of the best ways to begin the process is to start in AdWords Editor. Once in Editor, create two identical copies of the campaign you want to segment. You can do this by copying and pasting at the campaign level in Editor. This will save a ton of time by replicating all existing product groups, bids, locations, audiences, bid adjustments, etc. The goal here is to have two identical shopping campaigns that differ only in the branded presence (or lack thereof) in the initial query. Now, add “Branded” to one of the new duplicate campaign’s name and “Non-Branded” to the other. Then, make sure they are paused and post to AdWords.
(Creating two duplicates is really advertiser preference. Technically, you only need one; but creating two allows your original campaign to still exist. The original will obviously be paused once the branded and non-branded versions are ready to go. However, it will be there should you ever decide to revert back from the branded/non-branded segmentation.)
The final steps in AdWords revolve around negative keywords, priority settings, and bidding strategies.
As previously discussed, here is how you should approach each.
- Branded Campaign
- – Add negative keyword list of exact match general queries (optional)
- – Set priority setting as medium
- – Set bids according to account strategy and goals
- Non-Branded Campaign
- Add negative keyword list of branded terms and misspellings
- Set priority setting as high
- Set bids according to account strategy and goals
Once you are to this point, you are ready to begin implementing branded/non-branded segmentation in your shopping campaigns.
Tip: Instead of manually adding negative keywords, create negative keyword lists (especially for your branded variations) so you can easily reuse them on other campaigns.
Should you choose to test this type of shopping structure, it will naturally add more campaigns to your management list. However, don’t let that alone prevent you from testing this structure. You can still carry on with your regular shopping updates and optimizations. However, be extra vigilant of the queries coming through for these campaigns, especially in the non-branded versions. As stated, this isn’t a perfect system. You may notice a new branded misspelling in a non-branded campaign, a high-volume general query in a branded campaign, etc. But, by continually keeping a pulse on these queries and excluding accordingly, you will ultimately gain more and more control. Also, more time spent on the frontend will only create a more solidified branded segmentation.
Brand/Non-Brand Final Thoughts
This type of shopping structure is much less cut and dry than the tiered approach we covered in Part 1. There is more ambiguity and due to the dynamic nature of shopping, it isn’t 100% perfect. However, through testing and ongoing campaign maintenance, branded/non-branded shopping segmentation can show great returns. Again, it can be time-consuming depending on your account sizes so make sure it is a right fit.
Here are a couple of examples of when you should consider this structure:
- You have vendors that show for your brand and include your brand name within their ad. This will allow you to bid more aggressively on branded queries without overspending on traffic that is higher up in the funnel.
- You are in a very competitive industry and want to associate a specific spend and return on non-branded top-of-funnel shopping efforts. This will allow you to target and show metrics for those specific PLA initiatives.
There are obviously more uses, but these are common situations where this structure would make sense. So, if you think fits your account, then test it out and see how you can become more competitive with branded PLAs, more efficient with non-branded PLAs, and beyond.
Again, there are a plethora of ways to organize shopping campaigns within an account. However, these strategies are great for allocating budgets, competitive bidding, mitigating cross-pollination, and continually optimizing your PLAs. Furthermore, they are easy to implement and universal to nearly any e-commerce account with shopping campaigns. Once you establish one of these two options, feel free to customize and tailor the structure to best meet your accounts inventory, needs, and goals. They structures act as a foundation for a shopping with the flexibility morph into a very unique shopping strategy.