General marketing
URL’s and branding through search advertising
A name is a big part of any organization’s overall brand. And having a synchronized domain name is a key component of an organization’s presence online. Companies often spend a lot of time and resources to ensure that their address on the Web reinforces their company’s brand, product, or product attribute.
Traditionally at Mozilla, we’ve used “getfirefox.com” as the promotional url for Firefox through a variety of channels. Everything from t-shirts to newspaper advertisements and everything in between. For a long time, we also used “getfirefox.com” in our search advertising — specifically, the display url.
We recently completed a number of tests using search engine marketing to try and gauge the effectiveness of our url’s. In other words, how does the way we display a Firefox related url affect a user’s likelihood to consider Firefox? How does that change if the user is actively searching out the product (e.g. through a search engine) or if they happen to notice the ad through contextual placement on a website (e.g. sponsored text ads next to website content)?
As Ken has pointed out in a recent post, one of the reasons that we engage in search advertising is to ensure that we have an excellent user experience. Many people believe that a user’s experience with a product begins with the product itself. Ken and I would suggest that the Firefox user experience starts when someone first engages with a brand. Through search advertising, one of these first engagement points is viewing a text ad.
We ran four tests, each with a slightly different flavor. The descriptions are below and an upcoming chart will provide a visual explanation.
- non-branded search advertising (e.g. “browser”, “download”, etc.)
- branded keyword search advertising (e.g. “firefox”, “mozilla firefox”, etc.)
- non-branded contextually relevant placement (words like “browser” and “download” that appear based on the content of a web page)
- branded contextual placements (words like “firefox”, “mozilla firefox”, that appear based on the content of a web page)
The chart below illustrates the relative performance strength (in terms of click-thru) of each of the display url’s in each of the four tests, with each quadrant representing a different scenario. In all but one case, “Mozilla/Firefox” outperformed both of the other url’s (“Firefox.com” and “GetFirefox.com”). In fact, the order in which each url performed (“Mozilla.com/Firefox” outperformed “Firefox.com” which outperformed “GetFirefox.com”) was identical in every case except one.
What are some of the possible reasons behind these results and what are the implications for search?
Users with high awareness of Firefox (those that are either using branded keywords like “Firefox” in their searches or those that are visiting web pages that surface a Firefox text ad based on branded terms that appear somewhere in the text of the page) are giving much more credibility to a url with “Mozilla” in the name. In addition, the use of a subdirectory, which indicates that Firefox is a product of Mozilla, likely lends credibility to the ad with the user thinking he/she is clicking through to an official Mozilla Firefox page.
Observation #1: Use “Mozilla.com/Firefox” when awareness of Mozilla and Firefox is high
The results also showed that users who are browsing web pages where our ad was triggered on general key words like “web browser” are more likely to click on an ad with “Mozilla.com/Firefox” than either of the other url’s. Like the above observation, the user may view the presence of a subdirectory in the ad in a more credible way.
The one exception in our experiment to “Mozilla.com/Firefox” (see the upper left quadrant of the chart) is when users are searching for a new browser solution, but have a low awareness of Mozilla and Firefox; in other words, they use general terms like “download browser”. In that case, text ads should use “GetFirefox.com.” “GetFirefox.com” has been used as the standard marketing url for many years — it’s possible that the familiarity of this url causes the performance improvement over the alternative url’s. It’s also possible this particlular segment is more receptive to a “Get” call-to-action relative to the other three segments, given the interaction between their low awareness and higher propensity to download.
Observation #2: Use “GetFirefox.com” for search display url’s on non-branded terms (when awareness of Mozilla and Firefox is low)”
In a future post, I’ll review the results of the effect of using a prefix like “www” on a display url, and add some additional thoughts and recommendations for Mozilla’s approach to search engine copy and branding.
Discussion
9 comments for “URL’s and branding through search advertising”
Post a comment
Categories
Great post, thanks David. So what’s your recommendation for the URL on the back of our t-shirts?
I registered GetFirefox.com four years ago. firefox.com wasn’t available at the time but I felt like it was a better name anyway because of the “get”.
I have always felt it works better both in print and when spoken as simply “GetFirefox.com” without any www. prefix.
Huh, getfirefox.com is legit.
I saw a Google ad for that domain once and assumed it was some sleazy bastards redistributing Firefox with malware attached. After all, how would an open source project afford paid advertising, and why wouldn’t it be hosted at mozilla.org?
Downloading windows software linked to by google ads is a risky proposition. Take Spybot Search & Destroy for example. There was a nasty incident where malware authors were buying adsense keywords for it and tricking people into downloading a fake which actually was spyware, rather than a scanner.
Like you were saying, using an apparently 3rd party domain raises questions if you’ve got expectations about where Firefox is hosted.
URL’s and branding through search advertising | Giant Spatula > a Rolnitzky blog…
A name is a big part of any organization’s overall brand. And having a synchronized domain name is a key component of an organization’s presence online. Companies often spend a lot of time and resources to ensure that their address on…
Did you try “mozilla.org/firefox”? I think the fact that we are not-for-profit, which is indicated by a .org URL, might have some effect on people’s assessment of our credibility.
Gerv
I remember years ago–when firebird had (somewhat) recently changed its name to firefox–I was looking to install it on a new computer or a friends computer or something similar and the first url that came up was “getfirefox.com,” and I actively searched for another address to go to because I thought that getfirefox.com sounded like a spoofing site.
I have since learned the error of my ways, but I’m saying that people with exactly the right combination of paranoia and ignorance/knowledge might be turned off by that address. While it’s now the address I tell people when I want them to “get firefox” because it is so easy to remember, it still feels kind of spammy to me.
I’m glad you’re doing this research, is the point.
@ Ben: You may be right — using “www” might not be necessary in an offline environment (like print). People have grown accustomed to a url structure in the offline world where just having a dot-suffix indicates a web address. However, we’ve been running tests through search using a “www” prefix that show an increase in click-thru rate vs. non-”www” prefixes. I’ll post more on that soon.
@Gerv: Excellent suggestion. We’ll try and incorporate a “.org” into our next round of copy testing.
@John: It’s difficult to make extremely confident conclusions on web url’s for offline use when tested in an online environment. With that said, it seems like “Firefox” would be preferable to “GetFirefox” based on this testing. The length of Mozilla.com/Firefox (the overall best performer) makes it less practical for offline use.
[...] (Note: This is the conclusion of a two-part post. You can find part one here.) [...]
[...] hours later when they’re at their computer * this URL actually performed better when we did some SEM tests earlier this year (search results and t-shirts aren’t exactly apples to apples, but it’s interesting [...]