At nexus-security, we possess a substantial amount of advertising data due to our complimentary tools and extensive clientele. Our analysis has covered over $60 million in Google Ads expenditures. We enjoy delving into this data and providing you with performance-related insights such as average click-through rates, conversion rates, cost-per-click, and cost-per-conversion across various industries.

However, we had never previously examined the ad text itself. Naturally, our content team is fascinated by words. This led us to ponder: What specific words are used in exceptional, high-CTR Google Ads? Additionally, we wondered about elements like numbers, punctuation marks, and the most effective calls to action. Are the best Google Ads positive or negative in tone? Do they employ keyword stuffing or a more creative approach?

In this article, we will explore nine often unexpected discoveries we made by conducting text analyses on some of the highest-performing Google Ads from the past year.

This analysis was a collaborative endeavor. We extend our gratitude to all those involved, including our data analyst Josh Brackett, web team leader Meg Lister, designer Kate Lindsay, and content team leader Elisa Gabbert.

About the Data Set: Defining “Best Ads”

Our methodology for identifying top-performing Google Ads involved utilizing our free Google Ads Performance Grader, which analyzes every account that generates a report to determine its best and worst-performing ads.

We extracted all data collected from Grader reports between August 1, 2016, and April 30, 2017. We then focused on the top 15% of ads based on impressions, clicks, and CTR, resulting in a sample size of 612 ads for analysis.

Our analysis considered advertisers bidding in USD, AUD, CAD, and ZAR while excluding non-English ads and duplicates. We also differentiated between branded and non-branded ads (335 brand ads, 277 non-branded).

Note: Branded ads, which appear when users search for a specific business name, tend to exhibit naturally high performance, including high CTRs and Quality Scores. This is because users searching for a brand name often intend to navigate directly to its website, and prior brand awareness fosters affinity (potentially increasing CTR by 2 or 3 times). Consequently, we excluded branded ads from most of our analysis to isolate the tactics driving high performance even without the advantage of strong brand affinity.

With that said, let’s explore the data!

#1: Most Popular Words in Top-Performing Ads

In 2016, Google introduced Expanded Text Ads, effectively doubling the maximum character count for ads. This expanded format offers both challenges and opportunities for advertisers.

Regardless of your perspective, having more space for ad copy emphasizes the importance of word choice. But which words are the “right” ones?

To answer this, we analyzed our top-performing text ads, breaking them down into individual words and removing conjunctions (and, or) and articles (a, the) until we identified the ten most frequent words:

Our data from 335 high-performing non-branded text ads revealed the following most frequently used words, in descending order:

1. Your
2. Free
3. Now
4. Get
5. Online
6. Our
7. Save
8. Best
9. Shipping
10. You Any noteworthy observations?

While “Your,” “Free,” and “Now” seem self-explanatory, the inclusion of “Shipping” (often associated with “Free”) is intriguing. It suggests a potential emphasis on e-commerce among top performers. Additionally, “Get” and “Save” are the only verbs on the list, highlighting their broad applicability across both B2B and B2C text ads.

The presence of both “Free” and “Save” implies a persistent focus on cost-consciousness among shoppers. Meanwhile, “Now” and “Get” indicate a desire for immediate gratification.

Crucially, the second-person pronouns “You” and “Your” underscore the importance of directly addressing the customer in ad copy.

In essence: When crafting compelling ad copy, remember: “Your Free Now Get!”

#2: Most Common Calls to Action

The call to action, or CTA, is arguably the most crucial element of ad copy. Ideally, it should be prominently featured in your headlines and reinforced in the description or URL paths whenever possible.

In this context, the CTA refers to the language instructing searchers on the desired action and how/when to take it. The most effective CTA varies depending on the industry, target audience, and specific offer. It wouldn’t make sense to urge someone downloading a mid-funnel whitepaper to “Buy Now.”

We excluded branded ads from our CTA analysis as they often prioritize brand presence over explicit calls to action. These ads primarily serve as text-based billboards on the SERP, aiming to establish brand awareness rather than directly driving clicks. While valuable, this approach doesn’t provide insights into the effectiveness of non-branded (often more commercially driven) text ads.

Although phrasing may differ, the verbs used in CTAs are relatively limited based on the desired action. Our analysis of top non-branded text ads revealed the following most popular CTAs based on frequency:

For product-based businesses, “Buy” or “Shop” are common CTAs, while “Get” offers a less transactional alternative. Conversely, lead generation campaigns often employ CTAs like “Learn,” “Discover,” or “Sign Up.” Notably, the most frequent CTA verb in our data set was “Get.”

None of the top-performing ads used the verb “Click” in their CTAs, such as “Click here.” In fact, doing so might even lead to ad disapproval.

#3: Sentiment Analysis of Top-Performing Ads

The sentiment expressed in your ad copy, whether positive, negative, or neutral, significantly influences how prospects perceive and engage with your ads. We’ve extensively discussed the use of emotional triggers to enhance ad copy and landing pages.

Using the Vader sentiment analysis tool and Python’s NLTK library, we analyzed the sentiment conveyed in our set of top-performing ads. Ads with a compound score above .5 were classified as positive, scores between -.5 and .5 were deemed neutral, and scores below -.5 were considered negative.

Our findings revealed a predominantly positive tone, with 45% of text ads conveying positive sentiment, 53% neutral, and only 2% negative.

These figures are surprising, especially considering they only represent non-branded text ads.

While a similar analysis of branded ad copy might yield comparable results (as brands generally avoid negativity), wouldn’t a degree of negativity be preferable to bland, forgettable copy?

Personally, I’d favor ad copy that elicits a response, even a negative one, over generic, keyword-stuffed content. Provoking sadness, fear, or even a sardonic laugh is more effective than blending into the background noise of the SERP.

#4: Lexical Diversity in Top-Performing Ads

Lexical diversity refers to the ratio of unique words to the total number of words in a given text. In the context of text ads, it serves as an indicator of keyword stuffing.

Finding the right balance of keyword usage in ad copy is crucial. Too few keywords can negatively impact Quality Scores, while excessive keyword stuffing can alienate potential customers.

Our data shows that top-performing advertisers tend to repeat only two words per ad. For instance:

This is far more appealing than, say…

The takeaway? Maintaining lexical diversity is essential. While keyword stuffing might temporarily appease algorithms, it’s off-putting to human readers.

This is crucial because, ultimately, humans are the ones clicking on your ads and making purchases.

#5: Punctuation Usage in High-Performing Ads

Punctuation plays a critical role in conveying meaning and tone, whether differentiating between dining with or on a relative or marketing products to potential customers. A single character can drastically alter the interpretation of a text ad.

Analyzing punctuation usage in top-performing ad copy offers valuable insights into which types of punctuation to use and how frequently.

Punctuation in Branded Text Ads

Unsurprisingly, exclamation points dominate. This suggests a common tactic of ending ad descriptions with an exclamation point to create a sense of excitement and urgency.

The low usage of dollar signs in branded ad copy is understandable, as brand emphasis often takes precedence over pricing. Similarly, percentiles are less common in this context.

• Exclamation: 48%
• Commas: 38%
• Question Mark: 2%
• Percentile: 8%
• Dollar Sign: 11%

Interestingly, only 2% of the 335 branded ads surveyed used question marks. Experimenting with different punctuation marks, even if unconventional, can help your ads stand out.

Punctuation in Non-Branded Text Ads

Punctuation frequency in non-branded ad copy largely mirrored branded ads, with high usage of exclamation marks and commas.

However, dollar sign usage was 3% lower in non-branded ads. While this makes sense for brand-focused ads, it seems counterintuitive for non-branded ads where pricing is often a key selling point. A larger sample size might reveal different results.

• Exclamation: 42%
• Commas: 25%
• Question Mark: 11%
• Percentile: 9%
• Dollar Sign: 8%

These findings highlight the potential of underutilized punctuation marks like question marks and dollar signs to capture attention and differentiate your ads.

#6: Dynamic Keyword Insertion (DKI) Usage

In short, not really.

DKI allows you to tailor ad copy to match user search queries, creating more relevant and targeted ads. However, our data suggests that top-performing text ads, both branded and non-branded, tend to avoid DKI.

Only 10% of non-branded and 6% of branded ads we analyzed used DKI. While surprising given its widespread adoption, this supports the notion that DKI is a crutch rather than a solution. Over-reliance on DKI can stifle creativity, potentially explaining why ads using it are not among the highest performers.

#7: Use of Trademarks & Brand Signs

While analyzing branded ads alone wouldn’t reveal groundbreaking insights, we wanted to investigate how advertisers convey authority and authenticity in their copy.

Besides bidding on branded terms, using trademarks (™) or language that asserts official status is a common tactic. Past research has even shown that using symbols in ads can boost conversion rates. Surprisingly, only 28% of branded and 6% of non-branded ads in our analysis employed such tactics.

This might reflect the difference in brand equity between SMBs and large corporations. For example…

Searches for prominent brands often reveal attempts to assert authenticity, whether through trademarks or phrases like “Official.” If major brands utilize these tactics, you should consider them as well.

While not essential for high performance, experimenting with trademarks and brand signs could positively impact your CTR and conversion rates.

#8: Use of Numbers in Top-Performing Ads

Numbers are used in top-performing ads about a third of the time, which might seem low. Consider how often you encounter ads with numbers in them…

Most of them are doing it wrong.

Our data indicates that only 40% of high-performing branded ads and 37% of non-branded ads incorporate numbers.

This finding contradicts the assumption that numbers, especially in non-branded ads, are prevalent. The relatively low usage of dollar and percent signs further suggests that factors beyond price influence ad performance.

While numbers can be effective, this highlights that value proposition and messaging are equally important. Conveying value without relying solely on numbers is entirely possible.

While top-performing ads do utilize numbers strategically, over 60% relied solely on words, demonstrating the power of compelling language.

#9: Reading Level of Top-Performing Ads

We assessed the complexity of the best-performing Google ads using the Readability Calculator and the Fleish-Kincaid scale, a method for determining text readability.

Our findings revealed that the average reading level of top-performing ads targets a 14-year-old. In other words, writing at a 9th-grade level can lead to higher CTRs.

Interestingly, many best-selling authors employ a similar approach. Research by Shane Snow at Contently found that most best-selling authors write at or below a 9th-grade reading level:

Remarkably, Hemingway’s writing is even more accessible than Goodnight, Moon.

Constructing the Ultimate Google Ads Ad

By synthesizing our data, we can extrapolate the potential characteristics of the ultimate high-performing Google Ads ad:

This hypothetical ad incorporates:

1. High-performing words (your, free, get, our, save)
2. Effective CTA word (get)
3. Positive sentiment (emphasizing future savings)
4. Lexical diversity
5. Exclamation point
6. No DKI
7. No trademark or brand sign
8. 9th-grade reading level
9. A single number in the description line

Some of these findings were unexpected. What are your thoughts?