Resources - Blog

How to Maximize Amazon Advertising Performance With AI-Driven Keyword Harvesting

Discover the role of AI in keyword automation and learn the benefits of autonomous keyword harvesting for driving discoverability and conversions on Amazon. By Natalie Taylor May 5, 2020
How to Maximize Amazon Advertising Performance With AI-Driven Keyword Harvesting

When searching for a product to buy on Amazon, 44% of consumers scroll through two pages or less of search results.

To ensure your product’s visibility in shoppers’ search results today requires a high-performing Amazon Advertising campaign that targets the right keywords and search terms. For each keyword, you must define the match method — broad, phrase, and exact — and place a bid price for that search phrase.

To do this process manually is not only time-consuming, especially if you have a large catalog of varying types of products, but it also requires access to troves of data to analyze and interpret. As advertising on Amazon becomes increasingly imperative, sellers and brands are turning to technology solutions to automate keyword harvesting, save time and media spend, and optimize campaign performance.

Keyword harvesting is the automated process of identifying the right keywords to target in your ad campaign and autonomously transferring those keywords and search terms from one ad group or campaign to another. The process is implemented in two ways: from broad to narrow keywords and by leveraging automatic Amazon Advertising campaigns. Below, we explain each technique and the role of AI in keyword automation.

Keyword Harvesting From Broad to Narrow Keywords

Keyword match types for Amazon Advertising campaigns allow you to refine which customer search queries your ads are eligible to appear for. Manual targeting for Sponsored Products and Sponsored Brands offers broad, phrase, and exact match types — each which have their own use cases. These PPC ad types can have a variety of keywords on all three match types, with bids based on the anticipated impact. High-traffic, high-impact keywords should have higher bids, while broader keywords should be assigned lower bids.

  • Broad match: Your ad may appear when a customer’s search term contains all of the keyword terms or their close variants, such as plural forms, acronyms, stemming, abbreviations, and accents.
  • Phrase match: your ad can show when someone searches for your exact keyword, or your exact keyword amongst a sequence of words, making it more restrictive than broad match.
  • Exact match: The shopper’s search term must exactly match the keyword for your ad to appear. Unlike broad or phrase match, if a shopper searches for other words before or after your exact keyword, the ad will not appear.

In addition to broad, phrase, and exact match types, Sponsored Products also offer negative phrase and negative exact match types, which prevent ads from being triggered by a certain keyword or search term. When a keyword is selected as negative, the ad will not be shown to consumers searching for that phrase.

Keyword harvesting from broad to narrow keywords requires the advertiser to first provide an initial set of terms, typically broad, that describe the product. The goal here is to cast a wide net around the relevant keywords you want to test for our product, and then move the best-performing keywords into more targeted campaigns.

The initial broad terms can come from either your own knowledge of the product and category, the product description, or experience with keywords used on other channels, such as keywords used in Google Ads. Alternatively, you can also identify keyword research of your products on Amazon to determine the initial set of keywords for your campaign.

Because many of these initial keywords are broad terms, they are typically added with a broad match, which is intended to deliver to a larger base of shoppers and is used to expand your keyword coverage and increase the reach and exposure of your campaigns. Once the keywords are added to the advertising campaign, an Amazon Advertising platform will run ads against these keywords and match them accordingly against certain search terms, given the match type and bid price of the keyword

The results of this match of search terms and their conversion rates will then be available via the search term report that Amazon makes available to the advertiser. From there, the advertiser will typically analyze the report, identify the best-performing search terms, and add them to the list of keywords with exact or phrase matches, which are more targeted and relevant to a shopper’s search.

These more targeted keywords will typically perform better compared to the performance of broad terms and, as a result, will lead to a more optimal ACoS. This is because the cost of bidding on broad terms is higher and conversions are often lower compared to that of more targeted and narrow search terms that are proven to perform better for your products.

This strategy allows advertisers to start harvesting high-performing keywords from the start of the advertising campaign and is typically used for newly created campaigns or newly added products. However, over time, the effectiveness of this method diminishes, unless the advertiser can continuously identify new, highly relevant broad keywords.

The continuous evaluation of the most effective phrases and the process of adding them as keywords with the right match types is considerably time-consuming and prone to human error. Having an automated solution to streamline and optimize this process for you will not only save you time but also save you money by avoiding wasted ad spend on keywords that are not performing well.

Feedvisor’s AI-driven solution for automated keyword harvesting leverages machine-learning algorithms to continuously review the search term reports and the matched search terms’ performance, benchmarks against the performance averages of millions of advertised products and keywords, identifies keywords with the highest potential, and adds them to the appropriate ad group for targeting.

Keyword Harvesting via an Automatic Campaign

Another way to harvest keywords is by leveraging an automatic advertising campaign. This method does not require the advertiser to provide the initial list of keywords but rather the Amazon Advertising platform provides a mechanism in which it automatically explores keywords that might work for the advertised product. The automatic campaign generates a set of reports that provide insight into how the keywords performed, which can then be analyzed and, if relevant and high-performing, can be added to the manual campaign.

However, in order for the automatic campaign to serve as a strong source of keywords for the core manual campaign, the two campaigns must be structurally identical. In other words, the automatic campaign, associated with the core manual campaign, must have the same ad groups and products as the core campaign. Any discrepancies between these two campaigns will likely result in irrelevant keyword harvesting or failure to harvest highly relevant keywords.

Feedvisor’s AI-driven solution ensures this exact structural match between the two automatic and manual campaigns. AI technology autonomously adjusts the ad groups and products in the automatic campaign upon any change made within the core manual campaign, regardless of what caused the change — within or outside of Feedvisor’s platform.

Keyword harvesting from an automatic campaign is a complex strategy that requires ongoing revision and bookkeeping of keywords inventory as well as advertising budget adjustments. Any time the keyword is moved from the automatic campaign to the manual campaign, Feedvisor’s solution takes the following actions:

  • Machine-learning technology identifies the most accurate match type and the initial bid price for the keyword in the manual campaign, based on its analysis of search patterns observed in the automatic campaign.
  • The technology autonomously adds the newly harvested keyword as a negative keyword in the automatic campaign, which prevents the automatic campaign from bidding on it so it can focus on searching for new keywords.
  • The AI-driven solution then automatically adjusts the budget split between the manual and automatic campaign. Learn more about this functionality in the section below.

These actions, powered by AI technology, result in highly accurate and effective keyword harvesting via an automatic Amazon campaign — ensuring campaign structure mirroring and budget balancing to avoid wasted ad spend and maximize the campaign’s performance.

AI-Driven Budget Balancing 

Managing budget between the core and the automatic campaign is one of the most complex AI tasks. Although the automatic campaign is a very powerful harvesting tool, it leaves little control to the advertiser over which keywords the automatic campaign targets — hence the importance of proper budget balancing. It should allow the automatic campaign to effectively harvest new keywords, while also ensuring that the best-performing and proven keywords are allocated the most budget.

Feedvisor’s AI Budget Balancer continuously examines the keyword harvesting process and adjusts the budget split autonomously. The tool assesses the performance potential of keywords inventory of the core campaign and forecasts the performance of the automatic campaign, given the history and the data trends that are available thus far. Based on this assessment, the AI Budget Balancer revisits and appropriately adjusts the budget split between the two campaigns every day, as Amazon campaigns operate on daily budgets.

Leveraging AI technology and the aforementioned techniques, Feedvisor provides a holistic keyword harvesting system. At the start, the system gives higher weight to defining the initial set of broad keywords to get the campaign in motion. It then analyzes the performance of the initial keyword set and harvests new keywords on its own.

At the same time, the system autonomously creates an automatic campaign that supplements the core campaign’s keyword harvesting and ensures that the automatic campaign is structurally identical to the core campaign at all times.

Natalie Taylor
About the Author

Natalie Taylor is the content manager at Feedvisor, where she oversees and executes on the company's content marketing strategy. Prior to her work at Feedvisor, she wrote for a B2B supermarket magazine, focusing on merchandising and marketing trends in the grocery industry.

Final Thoughts

The Amazon Advertising platform typically needs a few weeks of experimentation and data-collecting before it can start to reliably identify high-performing keywords. Once the automatic campaign has had enough time to learn, it begins to generate keywords, which Feedvisor’s AI technology continuously evaluates.

When a highly relevant keyword is detected, the AI-driven solution automatically transfers the keyword from the automatic campaign to the core manual campaign, marks it as a negative keyword in the automatic campaign, and adjusts the budget split between the two campaigns to reflect the transition of the keyword.

The addition of the newly harvested keywords from the automatic campaign, combined with the ongoing harvesting in the manual campaign, gives Feedvisor’s AI technology a powerful field to crunch the data and generate new keywords continuously to keep optimizing campaign performance.

Feedvisor’s End-to-End Platform Helps Sellers and Brands Drive Growth on Amazon

Request Demo

This site uses functional cookies and external scripts to improve your experience. You may change your settings at any time. Your choices will not impact your visit.

I agree to receive cookies

Click here to read our Cookie Policy.