Please note that this feature is currently in beta. If you'd like access, please contact support@optmyzr.com.
What is Search Term and Keyword Intent Similarity?
Search Term and Keyword Intent Similarity is an AI-powered attribute in Rule Engine that measures how closely a search term matches a keyword's meaning and intent.
Unlike traditional similarity checks that rely on spelling or shared words, this attribute understands semantic intent. It can identify strong matches even when the wording is different, helping you make smarter decisions about keyword expansion, negative keyword planning, and query hygiene.
When this attribute is used in a rule, each search term and keyword is analyzed for intent-level meaning, and a similarity score is generated. The score can be used in Rule Engine conditions and filters, just like any other numeric field. This allows rules to act on what you meant, not just on how closely two phrases match textually.
Why Use It?
Traditional search term and keyword matching relies on shared words or spelling patterns. That approach often misses intent, especially when users express the same idea using different phrasing.
Search Term and Keyword Intent Similarity helps you move beyond text matching and optimize based on meaning. With this attribute, you can:
Identify true intent matches - Catch high-quality search terms that align strongly with your keywords, even when wording differs.
Reduce wasted spend from irrelevant queries - Detect low-intent or off-topic search terms that slip through text-based rules and add them as negatives with greater confidence.
Improve keyword expansion quality - Add new keywords based on semantic relevance, not just string overlap.Strengthen brand and competitor controls - Better distinguish between brand-aligned intent and competitor or unrelated intent, even when queries look similar on the surface.
Build smarter automation - Combine intent similarity with performance signals (spend, conversions, ROAS) to create rules that are more accurate and easier to maintain at scale.
How to Use It?
When setting up a Rule Engine strategy under the Ad Group Search Terms and Campaign Search Terms scopes, you'll see the 'Search Term and Keyword Similarity Using AI' attribute.
Select the attribute and define the intent similarity threshold you want to use. The score supports simple numeric comparisons, so you can use it like any other Rule Engine metric (for example, ≥ or ≤).
Finally, set up rules to take action based on the thresholds you define. You can include in report, add keyword, create SKAG, or add to negative keyword list.
Recommended Score Thresholds
While thresholds may vary by account, the following ranges work well for most use cases:
Highly relevant (≥ 85)
Strong intent match. Consider adding as a keyword when paired with performance signals (e.g., conversions > 0).Somewhat relevant (60–84)
Partial intent overlap. Best used with additional cost or conversion filters, or flagged for review.Not relevant (≤ 20)
Weak or unrelated intent. Consider adding as a negative when spend exceeds your threshold.
Tips for Large Accounts
When running this attribute on large accounts with high search term volume, following these tips will help ensure reliable execution and more actionable results.
1. Prefilter by campaign, channel, or brand
Narrow down your data set using campaign, channel, brand, or match-type filters before applying intent similarity rules. This reduces processing time and ensures the AI is evaluating only relevant search terms.
2. Combine with spend, conversion, or impression thresholds
Always pair intent-based conditions with spend, conversions, impressions, or cost thresholds. This helps:
Focus analysis on search terms that materially impact performance
Reduce noise from low-volume or low-impact queries
Avoid unnecessary actions on insignificant data
3. Avoid running unfiltered, account-wide rules at a very large scale
Running intent similarity rules across the entire account without filters can increase execution time and reduce clarity. Instead, segment runs by:
Brand vs non-brand
High-spend campaigns
Specific product categories
4. Use intent similarity as a decision layer, not a standalone trigger
Intent similarity works best as a refinement layer. Use it to confirm relevance after narrowing down candidates using performance signals, rather than as the only condition in large-scale automations.
Example Use Cases
You can use the Search Term and Keyword Similarity Using AI attribute for keyword management in multiple ways. Based on the intent score and your defined thresholds, you can decide whether to add new keywords, exclude search terms as negatives, or refine existing automation. Combining intent similarity with performance signals helps reduce false positives and improve decision quality.
1. Keyword Expansion
You can use this attribute to add search terms as keywords. Set up a rule where:
Intent similarity ≥ 85
Conversions > 0
Based on the thresholds explained earlier, if AI assigns an intent score higher than 85 to a Search Term, it has a strong intent match. Along with intent similarity, the Conversions rule ensures that the Search Term does have potential and you can consider adding it as a keyword.
2. Negative Keyword Creation
When the intent match is low, you're at risk of wasting your budget. It's best to add those Search Terms as Negative Keywords. Set up the following rule to exclude search terms:
Intent similarity ≤ 20
Spend exceeds your acceptable limit
3. Brand Protection
In branded campaigns, exclude competitor-intent queries when semantic similarity to brand keywords is low.