6 AdWords CTR Boons & Bombs
August 21st, 2013 by
You might not need a NASA mathematician, but a keen eye can help you identify your own CTR boons & bombs.[/caption]
Many internet marketers tend to forget that they sit on a veritable gold mine of big data. Each visit to a website is one data point. For paid advertising such as AdWords and Facebook ads, marketers end up with literally thousands of data points to test even the most minute characteristics of ads. While pulling the data, aggregating it, and analyzing it can be time-consuming, the value to your marketing plan is immeasurable. Recently, we pulled a sample of over 16,000 ads from the past 7 years of our AdWords accounts to see what aspects of ad copy increased click-through-rates, the overwhelmingly largest factor in keeping your costs per click low.
These ads predate the full rollout of enhanced campaigns, and are not segmented for anything such as ad position, device, or industry. Furthermore, of the 16,000 ads in the study, only ones with over 1000 impressions were sampled, an arbitrary but significant amount. When compared to large-scale CTR studies, the size of the sample falls short; however, we reached a statistical significance of at least 95% using single-tailed t-scores* in all 6 of these variables. Finally, it’s important to note that these six actionable insights were gleaned from 48 interaction, linear, and binomial variables that were tested. Of course, it’s hoped that these flaws do not invalidate the study; at worst, it’s an invitation for someone to check under the hood of their own accounts.
Three main factors were the most certain to be the cause of increases in click-through rate: “Sentence Format,” which changes the display of the ad; Calls to Action, in either line of the body copy; and DKI in the headline, which inserts the triggered keyword. In general, factors that increased relevance to keywords and uniqueness among competing ads increased CTR, which isn’t terribly far from the general status quo of ad copy literature.
Use Periods in Both Description Lines
Two variables checked for sentence-ending punctuation, one for each line of the body copy. Both lines lead to higher click-through rates, which is a bit of a surprise. Because we at Search Influence tend to keep our ads in the top positions, using punctuation will bring up the first body line into the headline of the ad, making the ad stand out against competitors on the page. However, the second line seems equally as influential, lifting the average CTR by .6% at a near-perfect confidence interval.
Ending the second line with an ending punctuation makes the user feel like the thought is done, a natural fit for an English-speaking user. A further aspect to look into is an ongoing debate in the office: Do exclamation marks entice or drive away users?
Call Your Users to Action in the Description
Putting calls to action in the description has a palpable effect on conversion rate, but highlighting the value to the user entices .52% more clicks at a 99.99% confidence interval. Placing the CTA in the first line teams up with using ending punctuation, raising CTR by .6% at 99.9% confidence, but the second line performs admirably as well, raising CTR by .38% at a 99.99% confidence.
Calls to action are treated very broadly, matching a variety of words often used in our campaigns. These words like “call,” “download,” or “contact” make your audience know that you’re expecting a certain action and can highlight unique selling propositions along the way. Interestingly, having a call to action in the headline was loosely correlated (90% confidence) to a significant drop in CTR. Users seem to subconsciously like to be guided, not bashed over the head with the sales pitch.
Use DKI in Your Headline
Finally, DKI remains an effective way to match user queries to ad copy. By inserting the triggering keyword into the headline, you raise your CTR by .44% at a 99.99% confidence interval, though DKI anywhere else in the ad has very little measurable effect on click-throughs. Relevance is by far the easiest way to get a users’ click, but it seems users are used to a matching keyword only in the headline, leaving the rest of the ad for setting yourself apart from the other people on the page.
In the same vein, I checked for the existence of the ad group name in the copy, expecting the same idea to hold true in static ads. Either due to inconsistent naming from the various alternative campaign structures that exist such as personas or highlighting the wrong keyword in the ad group, such as [breast augmentation] rather than [breast implants], there was either no correlation or a negative effect between including the ad group name in the ad copy and CTR.
On the other hand, these three aspects of ad copy were pretty certain to have caused drops in click-through rate. Unlike the CTR Boons, the Bombs generally fly in the face of conventional advice. In fact, blogs often suggestusing copy that drops these bombs.
Use Obscure Parts of Speech
Previously, we had suggested that a natural language answer to a search query is best. Yet “little” words like “the,” “an,” and “on” anywhere in the ad copy lowers click-through rates by over 1% at a 99.99% interval. This was the most consistent bomb across positions and the biggest drop to CTR, showing that it’s absolutely hated by users.
Users seem to need clear and direct writing, something along the lines of a grammatically correct Tonto. Simple, clear sentences win over verbosity every time. Furthermore, bullet-point style copy can also quickly outline main unique selling points and make the value of clicking on the ad clear to users.
Ask Your Reader Questions
“Try asking a question in your ad copy,” say so many copywriting guides, but our dataset shows a 1.18% drop (99.99% confidence) in CTR from asking questions in the headline and a .55% drop (98% confidence) due to questions in the body. This may tie into the previous bomb because questions often require little “helping” words to make sense. However, it could also be a question of relevance — asking a question like “Are You a Candidate?” introduces doubt into the user whether or not the page is really relevant to them, discouraging the click.
Finally, including “Now” or “Today” in the headline seems to turn off users from clicking, to the tune of a .97% drop at a 98% confidence. At lower confidences (90%), including the word “Now” in the body copy drops CTR by .33%. It seems that too much immediacy is a bad thing for users, making them feel pushed. Most verticals have a buying cycle, and trying to artificially shorten that may gain short-term clients but won’t net the same results.
Interestingly, at a 90% confidence interval, the word “Today” doesn’t have the same effect in the first line of body copy, boosting CTR by .4%. It seems that, again, playing to the buying cycle of the industry will ensure that your ads, and not your less pushy competitors’, will get the click.
After all this, it should be made clear that CTR isn’t the only holy grail, as being able to vet users before the click is extremely important. As with any data, more research will always be needed. The next step for this dataset is to look into collinearities that may shed further light onto the data. Finding data-based “power combos” for ad copy can jump-start any campaign and gives your copywriters real formulas for effective, large-scale advertising solutions.
Unfortunately, we can’t share all our data, but the next-best thing is to test your own data and test your assumptions — it’s the only way to make sure you’re doing the best job for your campaigns. Do you have any data you’d like to share?
*Disclaimer: Please excuse any oversimplifications of terms, though feel free to correct anything outright wrong with how I’m describing processes or data. (Get back to reading)