.gif)
Are All Non-Converting Keywords Bad?
By Anna Johnson on June 26th, 2009Jeff Sexton has written an interesting article in GrokDotcom that may have you rethinking your pay-per-click advertising campaigns. Whilst conventional PPC ad thinking has it that keywords that don’t lead to conversions are ‘duds’ to be eliminated from your campaign, Jeff Sexton suggests that they might actually be responsible for up to 30 percent of your sales.
Essentially Jeff explains that some keywords do not lead to immediate sales… but may lead to sales after several visits to your website. In fact, one of his clients discovered that removing those ‘dud’ keywords resulted in losing 30 percent of their sales.
This implies that PPC advertisers are wise to be a bit smarter about their keyword-to-conversion tracking. Unfortunately, says Jeff Sexton, relying on conventional web analytics tools such as Google Analytics can lead you astray.
Because of the limited capacity of Google Analytics to track repeat traffic that may have originated from a keyword, you may end up disregarding useful keywords.
Consequently, you may need to become more sophisticated in your keyword tracking, as well as apply some common sense. For example, some keywords are, by there nature, more likely to be used by people who are still in ‘research’ mode rather than those in ‘buying’ mode.
Source: Jeff Sexton, “Are Your Analytics Causing You to Lose 30% of Your Sales?” GrokDotCom, June 16, 2009
June 26th, 2009 at 12:25 pm
This is absolutely correct! Many of the tools that show you how your keywords convert, only provide you with the last click conversion metric. This means that 100% of the conversion credit goes to the last clicked keyword.
Keywords that typically convert the best are an advertiser’s branded keywords. Words like “iphone”, “blackberry curve”, “sony bravia tv”. A majority of the time however, visitors start their search by using more general terms such as, “cell phone” and “plasma tv” and then after doing some research end up using the more branded terms later.
The problem is that most of the time the branded terms end up getting the credit even though the general terms assisted in driving the visitor to those branded terms. But because many analytics tools do not give credit to the general term, many advertisers assume that they are not performing well. Incorrect assumption!
Problem solved. For those advertisers running Yahoo! Sponsored Search PPC ads, they have access to analytics tools (Conversion Only or Full Analytics) within their account that will give credit to keywords that contributed to driving conversions to other keywords.
For example, a visitor goes to Yahoo! and searches for “cell phone”. They see the RIM Blackberry ad and click on it. They reach the site, do some research on the 8820 model, but decide not to make a purchase that day. 15 days later they go to Yahoo! again and this time search for the more branded “blackberry 8820″ term. They like what they see and make a purchase.
The result? The general term “cell phone” is attributed with an Assist while the term “blackberry 8820″ is attributed with a Conversion.
The blackberry advertiser, can now see the total number of times that their keyword “cell phone” assisted in driving conversions to their other keywords. Of course on top of that, they are also able to see how many times the term “cell phone” converted directly (visitor searched for “cell phone”, clicked on the Blackberry ad, and converted).
Conversion Only and Full Analytics is available right now to every single advertiser that used Yahoo! Search Marketing.
For more information, take a look at this write up on a presentation we (Yahoo!) did with Havas Digital at the AdTech conference this year:
http://www.adtechblog.com/blog/detail/multi-click-attribution-the-way-conversions-actually-happen-draft-only/
Matt Lillig
Yahoo! Search Analytics
June 26th, 2009 at 8:35 pm
Thanks so much for your input Matt. Great to see Yahoo has a solution for this.