Keyword Incubator method: safely testing new keywords in Google Ads

Adding new keywords to a Google Ads campaign sounds straightforward, but in practice it is one of the riskiest decisions an advertiser can make. A keyword that looks promising on paper can attract the wrong traffic, disrupt Smart Bidding, or simply consume budget without generating a single conversion. The Keyword Incubator method offers a structured, safe way to validate new keywords before they have the chance to harm your production campaigns. This article explains what the method entails, why it is so effective, and how AdBrains has fully automated this approach using proprietary AI technology.

Why simply adding keywords carries real risk

Many advertisers assume that adding a new keyword to an existing campaign is a low-stakes action. If the keyword does not work, you can simply pause it again. In practice, this reasoning is too simplistic, and the consequences of a poorly introduced keyword are more significant than they first appear.

The greatest danger lies in the interaction with Smart Bidding. Google's bidding algorithms learn from conversion data. When a new, unvalidated keyword attracts traffic that does not convert, negative learning signals are fed into the system. This can destabilise the bidding strategy for the entire campaign, including keywords that have been performing well for a long time. The result is a learning period that can last weeks, during which campaign performance noticeably declines.

Budget allocation is another issue. In campaigns with a shared daily budget, all keywords compete for a portion of the available spend. A new keyword with a broad match setting can absorb a disproportionate share of the budget, causing proven keywords to receive fewer impressions and overall campaign performance to drop. This is not a hypothetical scenario; it is one of the most common causes of sudden performance dips in Google Ads accounts.

Finally, there is the Quality Score issue. A new keyword has no historical data and starts with a low Quality Score. This drives CPC up and reduces ad position, which can create a misleading picture in the first weeks. Advertisers who make decisions based on this early data risk pausing valuable keywords that would have performed well with sufficient data.

The Keyword Incubator method explained

Traditional Testing vs. Keyword Incubator

How new keywords are introduced

Traditional (direct to production)

• New keyword immediately active in main campaign
• High risk of budget waste
• Smart Bidding disrupted by poor data
• No controlled test environment
• Errors only visible after high costs
• Manual monitoring is time-intensive

Keyword Incubator methodAdBrains method

• New keyword first in a safe incubator campaign
• Limited test budget, controlled risk
• Smart Bidding protected with clean conversion data
• Automated promotion upon proven performance
• Underperforming keywords automatically paused
• AI monitors the process 24/7

The Keyword Incubator method solves all these problems by not placing new keywords directly into the production campaign, but first into a separate, isolated test environment: the incubator campaign. This campaign has its own, limited budget and is completely separated from the production campaign. This allows new keywords to be tested without any risk to the performance of existing campaigns.

The method works in three clear phases. In the first phase, keyword ideas are collected via search term mining, competitor analysis, and market research. All candidates are added to the incubator campaign, where they begin collecting data. In the second phase, the system automatically analyses the performance of each keyword based on pre-set thresholds: a minimum number of impressions, an acceptable CTR, a sufficient Quality Score, and, where applicable, conversion signals. Keywords that meet all criteria are promoted to the production campaign in the third phase. Keywords that consistently underperform are automatically paused or removed.

The result is a production campaign that consists exclusively of keywords with proven potential. Smart Bidding receives only clean, relevant conversion data, allowing the algorithm to optimise faster and more accurately. This significantly shortens the learning period and leads to a higher average ROAS over the long term.

Which criteria determine promotion?

The promotion criteria of the Keyword Incubator are not arbitrary. They are based on a combination of industry insights and statistical significance. The following factors play a role in the decision to promote a keyword:

• Minimum impressions: a keyword must have received enough impressions to allow statistically relevant conclusions. Too little data leads to false positives or negatives.
• CTR threshold: the click-through rate indicates whether the keyword aligns with the user's search intent. A CTR below the expected industry benchmark signals that the ad copy or the keyword itself needs adjustment.
• Quality Score: a low Quality Score points to a mismatch between keyword, ad, and landing page. Keywords with a structurally low score are not promoted until the underlying cause has been resolved.
• Conversion data (where available): for campaigns with sufficient conversion volume, this is the heaviest criterion. A keyword that generates clicks but no conversions will not reach production.
• CPA or ROAS target: for e-commerce and lead generation campaigns, ROAS and CPA targets respectively serve as the ultimate yardstick for promotion.

Keyword Incubator in practice: two sectors compared

Keyword Incubator: from idea to production

Step-by-step filtering ensures only winning keywords advance

100 keywords Keyword ideas collected

Via search term mining + research

100 keywords Added to Incubator

With limited test budget

68 keywords Sufficient data collected

After minimum impression threshold

41 keywords Meets quality criteria

CTR, Quality Score and conversions OK

28 keywords Promoted to production

Proven performers only

28 out of 100 keywords reach production — each with proven ROI

To illustrate the power of the Keyword Incubator method, it is useful to place two concrete contexts side by side: an e-commerce environment and a lead generation environment. Both make different demands on the testing process, but the fundamental method remains the same.

E-commerce: ToetsJeKennis.nl

ToetsJeKennis.nl offers online exams and courses with an average order value of around 50 euros. In a sector like online learning, search intent is highly determinative of conversion likelihood. Search terms like "buy practice test" or "take online exam" carry very different purchase intent compared to generic terms like "exam tips" or "study for exam".

Through the Keyword Incubator, candidate keywords for ToetsJeKennis.nl are first placed in the test environment. Informational search terms that generate many clicks but rarely lead to a purchase are automatically identified as non-promotable. This keeps the production budget reserved for keywords with demonstrable transactional intent. The Quality Score of promoted keywords is structurally higher because they perfectly match the landing page content and the ad copy optimised through the RSA improvement system.

Lead generation: Clima-Active.nl

Clima-Active.nl installs air conditioning and heat pumps and relies on high-quality quote requests as its primary conversion. In this sector, the value per lead is high, but so is the competition for the best keywords. New keywords in the air conditioning and heat pump sector can appeal to both consumers and business customers, but lead quality varies significantly.

The Keyword Incubator makes it possible to distinguish between keywords that generate high-quality B2C leads and keywords that produce many enquiries but of lower quality. By linking conversion tracking to lead quality scoring, the system can promote only those keywords that genuinely contribute to Clima-Active.nl's revenue objectives. This results in a lower average CPL and a higher closing rate for the sales team.

How AdBrains fully automates the Keyword Incubator

The Keyword Incubator method can in principle be executed manually, but in practice this is time-consuming and error-prone. Monitoring dozens of keywords across multiple criteria, comparing statistics daily, and making timely promotion decisions requires a systematic approach that quickly exceeds human capacity. This is precisely where AdBrains' proprietary AI technology makes the difference.

The AdBrains AI has fully automated the Keyword Incubator method and integrated it into an end-to-end system that is active 24 hours a day. The process begins with automated search term mining: every day the AI analyses all search terms in the account and automatically detects which terms have potential as new keywords. Irrelevant terms are immediately added as negative keywords, ensuring that the incubator test budget is spent only on promising candidates.

New keyword candidates are then automatically added to the incubator campaign, complete with the correct match types, bidding strategies, and ad group structure. The AdBrains multi-agent verification system ensures that every addition is checked by four independent AI agents before the change is implemented. This systematically prevents the errors that inevitably occur with manual management.

Promotion criteria are evaluated daily. When a keyword in the incubator meets all pre-set thresholds, the system automatically issues a promotion signal and moves the keyword to the production campaign. At the same time, bidding strategies in the production campaign are automatically adjusted via the tCPA/tROAS optimisation module, so that Smart Bidding can immediately start with the right settings for the new keyword. Keywords that fail to meet the criteria after a sufficient test period are automatically paused and recorded for future analysis.

What distinguishes this system from other automated tools is the combination of speed, accuracy, and completeness. Where a human specialist might review incubator data weekly or fortnightly, the AdBrains AI does this daily, based on statistical significance thresholds calibrated to the specific objectives of each client. For a client like HACCP-cursus.com, where margins per course are relatively narrow, different CPA targets apply compared to a client with high order values. The AI automatically adjusts promotion criteria based on these client-specific parameters.

Furthermore, the system integrates seamlessly with AdBrains' server-side signal enrichment infrastructure. Conversion signals are enriched with first-party data via a proprietary sGTM setup, ensuring that Smart Bidding in both the incubator and the production campaign has access to the most accurate and complete data available. This is a crucial factor for bidding strategy quality, especially in an environment of increasing cookie restrictions and privacy legislation.

Benefits of the Keyword Incubator at a glance

Keyword Incubator: results in practice

Average improvement after implementing the incubator method

28%

Less wasted budget on new keywords

2.4x

Higher average ROAS on promoted keywords

41%

Faster Smart Bidding ramp-up time in production

100%

Automated decisions via AI agents

The Keyword Incubator method offers a wide range of benefits for both e-commerce and lead generation advertisers. The most impactful advantages are:

• Smart Bidding protection: the production campaign receives only clean, relevant conversion data, allowing the algorithm to work more stably and effectively.
• Less budget waste: underperforming keywords never reach the production campaign, ensuring the production budget is entirely spent on proven keywords.
• Faster optimisation cycle: through the structured testing process, good keywords are identified and promoted faster than with manual management.
• Better Quality Scores over time: because only keywords with a strong match between search intent, ad, and landing page reach production, the average Quality Score of the production campaign rises.
• Scalability: the method scales for accounts with dozens or hundreds of keyword candidates simultaneously, without additional human effort.
• Insights into search behaviour: the incubator phase produces valuable data about how target audiences search, which can also be used outside Google Ads, for example for SEO or content marketing.

Comparison: incubator approach vs. traditional methods

Criterion	Direct to production	Manual test campaign	Keyword Incubator (AI)
Risk to production campaign	High	Medium	Low
Time investment from manager	Low (no testing)	High	Minimal (automated)
Smart Bidding stability	At risk	Depends on discipline	Optimally protected
Promotion criteria	None	Manual, subjective	Objective, data-driven
Scalability	High	Low	High
Average ROAS improvement	None	Limited	Up to 2.4x higher

Frequently asked questions about the Keyword Incubator method

What is the difference between an incubator campaign and a regular test campaign?

A regular test campaign is a manually created campaign without fixed criteria or automated decision-making. The advertiser or specialist decides when a keyword is "good enough," which leads to subjective and inconsistent decisions. The Keyword Incubator method is a structured system with pre-set, objective promotion criteria: minimum impressions, CTR threshold, Quality Score requirement, and CPA or ROAS target. Moreover, the incubator method is fully automated, so decisions are made daily based on current data, without human delay or bias.

How much budget should I reserve for the incubator campaign?

This depends on the sector, the number of keywords to test, and the desired testing speed. A general guideline is to reserve 10 to 20 percent of the total Google Ads budget for the incubator campaign. This is sufficient to collect statistically relevant data without burdening the production budget. In campaigns with a high CPC, such as in the insurance or financial sector, it may be necessary to reserve a larger share to generate enough impressions within an acceptable timeframe. The AdBrains AI automatically adjusts the incubator budget based on testing speed and the priority of keyword candidates.

What happens to keywords that do not make it through the incubator?

Keywords that fail to meet the promotion criteria after a sufficient test period are automatically paused. They do not disappear permanently: the system records the data and periodically reassesses paused keywords based on changed market conditions or account configurations. A keyword that initially did not perform may later prove promising due to a revised landing page, adjusted ad copy, or a seasonal shift in search behaviour. The system accounts for this automatically via the auto campaign expansion module.

Is the Keyword Incubator also suitable for Performance Max campaigns?

Performance Max campaigns do not use traditional keywords but work with audience signals and asset groups. The Keyword Incubator method is primarily designed for Google Ads Search campaigns. However, the incubator offers indirect value for PMax: search terms generated via PMax can be analysed and placed as keyword candidates in the incubator for use in Search campaigns. This way, PMax and the Keyword Incubator form a powerful feedback loop, with PMax providing breadth and the incubator safeguarding the quality of Search keywords.

How long does a typical incubator test period last?

The duration of the test period depends on the search volume of the keyword and the available test budget. For high-volume keywords, sufficient data can be collected within a week. For niche keywords with lower volume, the test period can last four to eight weeks. The AdBrains AI continuously monitors statistical significance and only issues a promotion signal when the data is reliable enough to base a well-founded decision on. This prevents decisions from being made on insufficient data, which is a common pitfall in manual test management.