Originally published on December 27, 2019, updated January 23, 2020
Dierk Demer of Prestozon shares advanced keyword strategies for Amazon PPC in this guest post.
I often meet with Amazon PPC Advertising managers and hear the same common misunderstanding: using the terminology of keywords and search terms interchangeably.
If you’re looking for a winning Amazon growth strategy, it’s important to understand the differences between search terms and keywords as well as how to optimize each. Many industry professionals in the Amazon PPC space make this mistake in everyday conversation. Whether lackadaisical or generally unaware, this conflation belies a misunderstanding of what each represents and how they should be treated within the Amazon advertising framework.
Search terms are the actual words that end customers are typing into the Amazon search bar in order to find what they are looking for. Search terms are the touch-point with the customer and so, in many ways, are the most important part of your Amazon PPC strategy.
Amazon doesn’t just take what the customer types into the search bar at face value. The Amazon PPC internal algorithms strip out all the extra language and punctuation to get to the core of what customers are searching for. When considering search terms (and keywords), it is important to realize that punctuation and superfluous words are largely ignored.
In order to maximize your revenue potential and efficiency, you need to build and manage your account strategy around search terms, not keywords. But to understand this distinction, you also need to understand what keywords are.
Keywords allow you to target Search Terms in Amazon’s PPC framework. Most importantly, keywords are where you manage your bid. Because bids are managed on this level, it is imperative to understand how keywords interact with search terms so you can make effective bid management decisions.
A keyword’s text is what defines the search terms it will fire on, but not all keywords are created equal. Amazon allows for three types of positive match keywords as well as two types of negative matches. The positive match keyword types are called Broad, Phrase and Exact.
Broad match keywords fire so long as the words in the keyword exist in the search term.
Phrase match keywords fire on search terms that have the same words in the same order (but allow extra words before and after the keyword’s text).
Exact match keywords only fire on search terms that match the keyword exactly. All of these rules have some exceptions. We call these exceptions equivalences.
Equivalences are tricky to pin down. Amazon takes into account certain words or word structures when deciding if a search term should fire for a keyword. Common equivalences are things like plurality or filler words like conjunctions or articles. Here’s an example:
The exact match keyword “magic wand” actually fires on both search terms “magic wand” and “magic wands.” These are equivalent and will always show up on the same keyword. There are other equivalences, as well, but it is important to note that most typos are NOT equivalences. Let’s look at the example search term “magic cape black” and see how different keywords would react:
Understanding the relationship between keywords and search terms is imperative to understanding how to effectively manage bid and ad spend on your Amazon PPC account. For instance, another powerful tool in the Amazon PPC arsenal is the negative match keyword. Negative match keywords work similarly to positive match in terms of search term coverage, but instead of allowing your ads to fire on those search terms, they remove the search terms they cover from consideration. This can be a powerful tool to ensure you aren’t being advertised on expensive search terms that don’t convert for you, but if you don’t understand exactly how they interact with search terms, you can hurt your account rather than help it.
A common mistake is to negate the plural of a search term and inadvertently negating the singular. For example, imagine the search terms “magic wand” and “magic wands” have spent $25 in the last month but“magic wand” has made $100 in sales and “magic wand” has made $0. It’s tempting to negate “magic wands” but doing so would negate “magic wand” as well! Negating the plural search term also negates the singular and vice versa.
As explained above, the only actual touch-point you have with your end customer is the search term, but the only place you can manage your bid and decide which search terms allow your ads to be shown are the different types of keywords. This means that in order to manage bid and your search term exposure effectively, you need to become an expert on the relationship between search terms and keywords.
Let’s look at some examples, first from the keyword level down to search terms:
In the first example, we have a broad match keyword for the phrase “top hat”. Because it is broad, it is firing on a large swath of search terms. You can see those terms in the expansion below the keyword. These different search terms have drastically different cost-per-click (CPC), conversion rate (CVR) and overall performance. Because this keyword is firing for so many search terms, all with different performance and CPCs, it is impossible to find a bid that is going to optimally manage all the search terms and control your cost.
Now let’s flip the typical hierarchy upside down and look at the opposite case, managing from the search term level down to the keyword:
Here we see multiple keywords firing for a single search term. Amazon uses bid to decide whose ad is displayed, but they also use other metrics. Things like relevance, conversion rate and retail readiness are also factored into their proprietary placement metrics. Because Amazon retains a level of control to determine which keyword wins the auction, having multiple keywords firing on the same search term causes interference and an inability to directly control the search term.
In this particular case, the broad match keywords are managing a large group of search terms and so they may not be optimized for this particular one. Notice that the exact match term is performing better in regards to almost every efficiency metric, yet it is not getting the most impressions. By removing this search term from contention for the broad match terms (possibly by using a negative exact match keyword described above), you could shift the impressions to this exact match keyword and drive similar volume but more efficiently and at a lower ACoS. The primary reason to separate out broad, phrase, and exact keywords into distinct ad groups is so that you have the ability to add negative keywords to broad and phrase ad groups to control search term traffic.
Finally, let’s look at a case where a search term and keyword have a one to one relationship with each other, again starting from the search term level down to the keyword:
As you can see here, the metrics for both the keyword and the search term are completely aligned because the search term is only firing for this keyword and this keyword does not have any other search terms it can fire for as it as an exact match (excluding equivalences, which will always have the potential to occur within the Amazon framework at this time). Because there is a one-to-one match between the search term and the keyword you know that bid changes on that keyword directly control your bidding on that search term and only that search term.
This is in stark contrast to the examples above. In the case of managing on a broad or phrase keyword, you may be optimizing bid for one or two of the search terms covered by these keywords, but you may actually be hurting the performance of other search terms under their umbrella. Alternatively, having multiple keywords firing for a single search term can merely cause impressions and clicks to shift between keywords as bids change rather than actually driving meaningful change on the search term’s performance.
Prestozon understands that search terms are your touch-point with the customer. Because of that, both of our keyword rules functionalities use search term data rather than keyword data. For our keyword promotion rules, we are focused on finding new search terms that will be successful as Exact Match Keywords. We scour your auto campaigns as well as your broad or phrase match keywords and look for new search terms that have converted in order to promote them to an exact match.
Our negative rules utilize an algorithm to suggest expensive search terms to negate in your ad groups. These also run on the search term level. This is important because you can add multiple ad groups into these rules and we will aggregate the data across the ad groups. So if you have a single search term firing in multiple ad groups across the rule, we’ll look at the data from all of those to decide if the search term is ultimately worth negating. We also account for search term equivalence to help avoid inadvertently negating search terms that have made a sale.
You cannot manage bid on individual search terms in Amazon, you can only manage bid on keywords. It is for that reason that we feel the strategy of Search Term Isolation is so important to effective bid management in the Amazon PPC space. Because bid is managed on the keyword level, we use keyword level data for our bid management algorithm. This means that the closer you can get to a one to one relationship between each search term and a single keyword, the more optimized your bids can be.
What does all this mean?
Considering keywords and search terms interchangeable, even on the exact match level, oversimplifies the complicated relationship between the two. Amazon PPC bid management and search term management happens on the keyword level, but that doesn’t mean you can’t find ways to manage precisely on the search term level.
Most Amazon PPC strategies begrudgingly deal with the fact that you’ll always be managing your account one step removed from the actual customer touch-point. By using a strategy like search term isolation and a platform like Prestozon, you can directly manage on the customer touch-point: the search term.
Originally published on December 27, 2019, updated January 23, 2020
This post is accurate as of the date of publication. Some features and information may have changed due to product updates or Amazon policy changes.