We had a recent discussion at 3Q Digital about PLAs – specifically, whether it’s better to segment products at the SKU or category level. The discussion was short and one-sided: we all agreed that SKU-level segmentation is the way to go.

plas

Maybe you’re segmenting by category because management is easier, and we get it…but it’s definitely not as effective. This post is going to break down the pros and cons of each type of management, but I’ll be honest: if you’re not convinced by the end that you should go the SKU-level route, I haven’t done you much of a service.

Here goes…

Segmenting products at the SKU level

Pros:

-Insight into individual product performance rather than a group of products. If you want to easily break out product category level as well, you can add relevant SKUs to an ad group and make that your product category (e.g. shirts, pants, shoes, etc.) or even include product types/categories in the feed.
-Reviewing individual product performance allows you to optimize and test at the most granular level:

-Overall product performance: Increase/decrease bids according to how well the product is performing against your goals.

-CTR: If CTR is struggling, consider editing your headlines, description lines, and images.

-CVR: If CVR is struggling, you may want to test a different LP (of course, the product still has to be the main focal point).

-Price: What’s the impact of decreasing or increasing price? Rather than include a $10-off promotion (which the user needs to hover over the PLA to view), test tasking $10 off the base price to get the upper hand on competitors.  Measure performance, and determine if lowering prices is a viable option to increase return.

-If ad groups are broken out at the SKU level as well, you can include individual product-specific promotions in PLAs (although this can get messy with a large number of products). Keep in mind that promotions are added at an ad group level, so if the products are grouped in an ad group, the promotion must be relevant to all products within the ad group.
-Note: This is going to change in the new Google Shopping campaigns, where promotions can only be added at the campaign level. If you’d like to advertise more than one promotion in PLAs, you will have to create separate campaigns and set up campaign priority levels.
-Setting up SKU-level targeting is just as simple as product category targeting. Just create a bulk upload sheet and use AWE to make all of the changes

Cons:

-Harder to manage. Segmenting products at the SKU level means there you’ll need to do more time-consuming bid adjustments/analysis.
-Funneling traffic is much more difficult in PLAs than regular search. A search query can trigger a range of products even if broken out by SKU. Negatives can be added at the ad group level in an attempt to help funnel traffic, but there is no guarantee in SKU to query mapping. This applies to product categories as well, but is much more complex at the SKU level.
-Related to the mapping issue above and with the lack of data available in the Google UI, it’s possible that increasing bids for a certain product will start stealing traffic away from closely related products.

Segmenting Products at the Category Level

Pros:

-Easier to manage. You can review data at a higher level and report on the bigger picture. Rather than worrying about individual products, you’ll find it easier to report on brands, product types, etc. (Keep in mind that this can be done at the SKU level too if ad groups and product types/categories are set up properly).

Cons:

-Lose insight into individual product performance. Since products are grouped together, this can be a big issue because:

-A product category may be performing poorly. This will likely cause the advertiser to decrease the bid.  In some cases, the poor performance may be driven by one or two products while the other products in the same group are performing very well. By segmenting at the category level, adjusting bids will affect all products in that category, regardless of performance. Segmenting at the SKU level allows advertisers to review the performance individually, making the necessary adjustments to optimize the campaign more effectively.

-Segmenting at the product category level makes it much more difficult to test headlines, descriptions, LPs and images. Being able to measure the impact of certain changes is crucial when testing; advertisers lose that insight when grouping products together.

A common misconception would be the grouping products by category would also mean less work to set up the feed. That’s not the case at all, since all products must be included in the product feed individually, regardless if you are using SKU or product category segmentation. Honestly, the most time consuming aspect of PLAs is setting up and optimizing the feed, which will need to be done no matter how the products are being targeted and reported on.

So how’d we do? Ready to go SKU-level on your PLAs? It may seem like a lot of work, but you’ll love the numbers you’ll have to show for it.

This entry was posted in Featured. Bookmark the permalink.

3 thoughts on “Always segment PLAs at the SKU level!

  1. You hit a nerve with this one. Boss says sku level. I say category because the feed is over 1 million sku’s ( I manage the day to day). I hope he’s not reading this as you fall into his camp. .

  2. Why not at the ROI level? Obviously you have to scale your time – use a tool like: Adchemy for PLA or Marin Marketer. With Adchemy you can scale your sku level, descriptive-kw with your adcopy and get insights into your search query’s that undermine performance. Marin offers ability to automate some of the tasks for search query optimization (assuiming you can leverage these in your PLA campaigns) – we often set automated negations for search queries that add up in cost but not performance.

    Be sure to sell your client on the value of paying for PPC services so then can afford the optimizations necessary to manage their campaign effectively!

  3. I often find it ultimately comes down to the amount of data at your disposal. If click volume for individual SKUs is relatively low, it can often be beneficial to jump up a step and segment at the category level.

    Also, just because an individual SKU has performed well in the past, does not necessarily mean that it will perform well in the future. Working at a category level will naturally control for product fluctuations and can allow more reliable analysis.

Leave a Reply

Your email address will not be published.


*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>