Yahoo!’s demo of its upcoming Project Panama ad platform at last week’s Search Engine Strategies conference was novel for a number of reasons. Prime among them was that the audience broke into applause at certain points in the presentation—something this reporter hasn’t encountered at a trade show before.
One of the bullet points that elicited a smattering of spontaneous love was the news that the new platform would have the ability to credit keyword “assists” to conversions. In other words, while marketers might be able to tell that a searcher converted after clicking on the term “Sony Cybershot”, the Yahoo! tracking software will also let them know that before that, the same user searched on the generic phrase “digital camera”.
Knowing that, the thinking goes, will give advertisers a truer picture of what path their users are taking to get to the point of conversion. That will help them better allocate their keyword bids and beef up the return on their search marketing investment.
Investigating that helper connection among keywords has been the aim of a series of studies conducted this year by search marketing firm 360i and search tech provider SearchIgnite. The latest study in that project, delivered just before the SES convention, found that more than one-third of search marketing conversions take place with the assistance of at least two clicks on a search result or paid search ad. And when multiple clicks lead up to a conversion, the chances are that the early clicks are not getting their proper due.
The study, in fact entitled “Giving Clicks Credit Where They Due, Part Two”, also found that 12.6% of the conversions credited to clicks on marketers’ organic search results were in fact preceded by at least one click on a paid search ad. So if conversion credit is being misallocated, it’s probably at the expense of paid search.
“Marketers need to look at all the clicks together,” says David Berkowitz, director of strategic planning at 360i and one of the report’s authors. “A lot of marketers use different firms for their natural and paid search marketing, or they do some work in-house and outsource other jobs, split up campaigns into branded and non-branded terms. But if you’re not enabling some kind of communication among all those channels, you’re not going to see the full process.”
SearchIgnite and 360i looked at the clickstream data for more than 250,000 users who transacted with marketer clients of the two companies, and for whom they manage large amounts of both paid and natural SEM. It looked only at clickpaths that ending in a purchase conversion; lead generation and sign-ups were not counted, and clicks that did not end in a conversion were discarded from the study.
The report found that while more than 60% of conversions came after a single click on a search results page (either a natural listing or a paid search ad) multi-click clickpaths accounted for more than two thirds of the total clicks examined during the investigation.
Obviously, Berkowitz says, search is a much more complex activity than marketers might conclude from holding their online conversions up to their log files.
“Search is a process, and marketers need to respect that,” he says. “Consumers aren’t deciding to do a paid search today and a natural one tomorrow. The most important thing any marketer can do is to make sure that they’re viewing this conversion process end to end. At some point, they need to make sure they’re able to see everything as a whole.”
Besides looking at paid and organic search clicks, the report also categorized search terms as branded (for example, “Joe’s Electronics Store”) or unbranded (“DVD player”). This links the study to another produced in May by the same pair, which looked only at paid search ads.
That May report found that the accepted wisdom that branded terms convert better is true; when a user’s first click and last click were on paid search ads tied to branded keywords, the conversion rate reached 9.3%. But the study found that conversions can be almost as good (8.7%) when a user starts by clicking on a paid ad from an unbranded term and then switches to a paid ad on a branded term. In other words, the unbranded keyword played a role in the conversion. The report concluded that marketers should focus their efforts on capturing searchers with unbranded terms and then locking them with branded ones.
“It’s tremendously valuable to marketers to know that unbranded terms can be just as good an assist to conversion as branded ones,” Berkowitz says, speaking of the first report. As with the interplay between natural and paid search, advertisers know this innately, he says. But because the ROI on branded terms is so good, they may mistakenly try to protect that return by not allocating as much budget to unbranded terms as they should.
This new report added organic search to the mix. It found that a slight majority of all the searches studied started with a click on a natural result (55.4%). A larger majority of first clicks occurred on branded terms (63.5%).
The study also found that most users chose their search flavor and stuck with it. Only 8.5% of the subjects, including 22.9% of the ones who took more than one click to convert, flipped between paid and natural search.
Among those who did switch, the most likely path was to move from a first click on a paid listing to a final pre-conversion click on a natural result. That happened in 12.6% of the instances studied, more than twice as often as the counter-move from a first natural result to a last click on a paid listing.
Multi-click searchers also showed some strong loyalty to branded terms. Users who started with a natural result on a branded term ended up with the same kind of click 81.5% of the time. When users began by clicking on a paid listing for a branded term, their last click before converting was on the same type of term in 74.3% of cases.
Non-branded terms didn’t exert as strong a hold. Only 57.6% of those who began with a natural result on an unbranded term finished up the same way. An even smaller proportion of those who set out by clicking on a paid listing for an unbranded term finished with the same: 46.4%.
Zeroing in on users who clicked several times before converting and who switched between branded and unbranded terms, the most common path went from a click on a natural result and a non-branded search term to a natural result and a branded search. That happened in 32.5% of switch cases.
But the next most common transition—and the assist most often unaccredited-- was from a paid listing in a non-branded search to a natural result of a branded search, which occurred in 22.4% of the cases studied.
Of all the instances where a searcher switched modes during the search, it’s this transition from an initial paid-ad click to a final organic click that is most likely to be misread by marketers, Berkowitz says.
“Those paid/ non-branded first clicks are really the least likely to receive proper credit for the assist, even though they helped bring about the conversion,” he says. “When these searchers started out, they were not sure whom they were going to do business with. But they saw a paid listing, clicked on it, and the brand resonated so strongly that they ultimately wind up converting through that term.”
Berkowitz concedes that checking keywords for their “assist” contributions adds a level of complexity to SEM strategizing that some marketers may not want to think about. After all, the large search engines are already adding complications to their inner workings with quality scores and demographic targeting; search engines must also decide whether to expand into second-tier or specialized search engines; and those keyword lists keep growing.
But that’s exactly why they should make the time to understand more about the path their customers take to the online conversion, Berkowitz says. “The more complex their SEM ad campaigns get and the more money they pour into them, the more valuable this knowledge will be, to make sure they’re budgeting and running those campaigns as efficiently as they can,” he says.