Seven Habits of Highly Effective SEM, Part III: Creating the Right Ad Text


I often tell prospective search engine marketers that to be good at SEM you need to be good at three college courses: statistics, psychology, and creative writing. Creating good ad text is probably the one SEM skill that combines all three of these skill sets.

Here are my top tips for creating and optimizing effective ad text.

1. Test, Test, Test . . . Wisely. As with any aspect of SEM, testing is fundamental to ad text success. The key to ad text optimization is to test efficiently – it’s difficult or impossible to successfully test ads across hundreds or thousands of ad groups. I recommend a two-pronged approach. First, find a “global” winner – i.e., an ad text that can work well across multiple ad groups. This should be an ad with a slightly more generic message than one that you would create for a targeted ad group, but one that you feel pretty confident will work well in many situations. Create two or three of these global ads and test them in multiple and diverse ad groups over a period of a couple of weeks. Once you’ve found the winner, delete your two losers and add two new challengers (I call this an “ad text war”). Wash and repeat until the increase in performance becomes minimal or your ad text winner holds the top position through many ad text wars.

Next, perform the same ad text war for your top five to 10 ad groups. This time, of course, your text should be targeted to each specific ad group. Continue to test until you’re confident you’ve found your winner. In many cases, you may find that you find a few winners – two or three ad text that all perform at high levels. You may wish to keep two to three ad text in each ad group as the rotation of the ads can help your ad look fresh to repeat visitors.

Combine these two techniques and you can continually improve the performance of your entire account and your best ad groups without spending your whole life testing ad text.

2. CTR is Just a Part of the Equation. Absent in my discussion of ad text testing was the definition of what qualifies as a “high performing ad text”. Many people have been duped into thinking that the sole determinant of ad text success is click through rate (CTR). If you are optimizing for Google’s stock price, then this theory is correct; if, however, you are trying to run a business, you should be optimizing for a combination of CTR and conversion rate. While it is true that your CTR will reduce your CPC on the search engines, sending thousands of unqualified visitors to your site will quickly deplete your budget and ruin your ROI.

My recommendation is to set up a simple two factor equation to determine effective ad text: CTR multiplied by conversion rate. For example, let’s say that Ad Text A has a 10% CTR and a 5% conversion rate, and Ad Text B has a 5% CTR and 20% conversion rate. Ad Text A gets a score of .005 (.1 X .05) and Ad Text B gets a score of .01 (.5 X .2), thus Ad Text B wins over Ad Text A, even though Ad Text A has a higher CTR.

To put this same example into more concrete terms, if both ads got 1000 impressions, Ad Text A would get five conversions and Ad Text B would get 10 conversions. Optimizing just for CTR would be a win-lose situation – a win for Google and a lose for you!

3. Research Competitive Ads – Copy . . . But Not Exactly. When first developing your ad text copy, you can save yourself a lot of time by first looking at what your competitors are doing. More specifically, look at what the advertisers in the top positions are using for their ad text. Most likely, these folks have refined their ad text over time and have found a winning combination.

This does not, mean, however, that you should just outright copy these ads and run them. There are two very good reasons not to do this. First, your company’s value proposition might be different than your competitors. For example, if you are a budget hotel chain buying the keyword “hotel” and the Ritz Carleton is in top position, you don’t want to copy an ad that says “Luxury Hotels Redefined.” Second, and more importantly, you want to create an ad text that stands out from the crowd. If the top advertisers are all basically saying the same thing, the odds that your ad (which will, at least initially, show up lower down in the rankings) will attract a user’s attention is slim.

What you want to do is to understand the gist of the top ads, but then create an ad with a ‘similar but different’ message. In this way, you can use your competitors’ research to quickly understand what ad text is effective, but also come up with positioning that helps you stand out and hopefully get a lot of qualified clicks.

4. Always Use a Call to Action – Start Now! Humans are herd animals – they like to be told what to do. This rule definitely applies to search engine advertising. If you don’t end your ad with a call to action like “Learn more now!” or “Download your free white paper today” or whatever you want someone to do, many users will likely not click on the ad, simply because you didn’t tell them to.

I know there may be some of you out there that are opposed to such call to actions, on the grounds that you don’t want to be “too market-y” or “offend potential customers.” While I think that there is some merit to this argument (see my last point on marketing too much), I still think a call to action is essential. The difference – for those of you who want to do a softer sell – is that you just need a more subdued call to action. Rather than ending with “Hurry and get your 4 absolutely free quotes this instant!”, you should try something like “Schedule a free consultation” or “Research available options.” In other words, you can still create a great call to action without feeling like you are yelling at your potential customers.

5. Use DKI . . . in Moderation. Whenever I train young search engine marketers, they always get particularly excited when I teach them about Dynamic Keyword Insertion (DKI) – the ability to automatically add the user’s query into the ad text. DKI is a very cool feature and I use it a lot, but it doesn’t always work and can sometimes backfire.

There are a few problems with DKI. First, it is often the case that everyone (and their mother) is using DKI. As a result, you end up with ten ads that all have the user query as the headline. This ‘tragedy of the commons’ scenario means that all advertisers nullify each others ads, and no one gets a particularly good CTR. For this reason, I always recommend that you create ad text variations in your ad groups with and without DKI. You may find that your non-DKI ad text will win.

Another problem with DKI is that it often makes for unintentionally funny ad texts. It’s always fun to type in random searches like “kidney” or “belly button lint” and see who has a created a generic ad using DKI that reads something like “Buy Belly Button Lint on eBay.” If you don’t think through how your keywords might look with a DKI ad text, you may end up running such ads. This is especially problematic when you use DKI within the text of your ad. For example, an ad that reads “We sell {KeyWord:widgets} online!” could end up reading “We sell Find Widgets online!” if you don’t think through all the possible DKI combinations in your ad group.

I’m all for DKI, but it can actually do more harm than good if you aren’t careful!

6. Get Bonus Bolding. As I’ve noted in a prior column, there are several ways to get extra bolding, extra lines of text, and extra relevancy in your ad text. Adding Google Checkout, for example, gets you a big Google Checkout logo under your ad, and geo-targeting gets you the name of the geographic region you are targeting. The more real estate your ad takes up, and the more bolding it receives, the more likely it is that users will click through.

7. Use the Four Human Emotions. Internet marketing is new, but direct marketing is not – we’re all accustomed to getting dozens of letters from credit card companies and circulars in our newspapers. All of those direct marketing messages you receive rely on four basic human emotions to get you to buy something: fear, greed, exclusivity, and vanity. While the medium has changed, humans have not – use these emotions in your ad text to drive CTR through the roof!

8. To Be Market-y . . . or Not To Be. My final point is to never assume that you ‘know’ what your users want to see in ad text. As part of your testing, you should always test different tones – some very direct response and aggressive, others soft and unobtrusive. A few years ago when I was marketing mortgage quotes, I tried an ad that said something like “Get four free mortgage quotes, serious inquiries only, please.” The conversion rate and CTR was incredibly high on this ad (and I subsequently saw my competitors start to use it as well). I think that the ad worked because it was very different from all the very agro ads out there, and it gave users a sense of credibility (we are a serious company, we only want to work with serious consumers). This doesn’t work all the time, but by putting down my marketing pen for a moment, I was able to surprise myself (and apparently consumers).

Stay tuned in a couple weeks for Part IV, choosing the right bid.

To see the other parts of this series, click these links for the overview article, Part I (choosing the right search engine), and Part II (keyword creation).

About the Author

David Rodnitzky is CEO of PPC Associates, a leading SEM agency based in Silicon Valley. PPC Associates provides search, social, and display advertising management to growing, savvy companies. To learn more, visit, or contact David at

Add Your Comments

  • (will not be published)


  1. I particularly like your point when it comes to testing ad text. Many people simply optimize ad texts for CTR...big mistake. Others test for conversion rates. Very few understand and follow the CTR X Conversion Rate calculation when testing ad texts. THIS is the way to find the real top-performing ads! Steve LaLonde

  2. I like this article very much. Many people who involved in PPC mangement are only considering the CTR as a main thing but they never think about the conversion rate or other stuff involved in PPC like cost per conversion, CPA etc. In my point of view, we should consider all the above factors to measure the success of top performing ads. GemKing

  3. This is a great blog article highlighting the fundamentals of SEM ad text. If you don't mind I will bookmark it and refer to it at a later date. Many thanks. .-= Quality Hotels´s last blog ..Prince Hotels =-.

  4. Your are absolutely right to focus on both CTR and Conversions when comparing ads. However, one could argue that CTR should be assigned a higher weight given how important it is to your keyword / ad's Quality Score.