Online Exclusive: Rate, Not Volume
Methodology. The Index looks at conversion data from across several fields: media/entertainment, finance, travel, online service, telecom, vertical search, retail and business to business. We defined conversions according to each marketer's unique search goals. For one company, a conversion could be an acquisition of an online lead; for another, it could be a purchase from a bricks-and-mortar store.
That left us with a whole lot of conversions types, from a slew of different kinds of companies, so we normalized the data to make sure we were comparing apples to apples. The data came from our clients' campaigns (a total of many, many millions of searches). To make sure our findings were indicative of the entire search world and not just of the Did-it world, we ran our numbers against data from online research firm comScore Media Metrix. Then, we added everything up.
The final result is two lists. One ranks the relative value of a search source for driving conversion volume. The other ranks how well search sources drive conversion rate. The index begins with June 2005 data; it's updated monthly. The list ranks the top 20 sites in each category (a rank of 1 is the highest; 20 is the lowest).
The findings. Some of the findings were a bit surprising; others, less so. We'll focus mostly on December to keep things simple.
Conversion volume pretty much follows traffic volume. Google was the top source for getting high conversion volume, followed by Yahoo; Web sites that aren't specifically search sites but happen to have search bars on them didn't fare as well. And that starts to flip when you look at conversion rates: lower-traffic sites with search bars did better and the high-traffic, generic sites started to slip. So while AT&T portal att.net, which features a Google search bar, didn't even make it to the conversion volume top 20, it scored as No. 1 in December for conversion rate. Google.com, by contrast, had the conversion rate No. 14.
Quality, not quantity. What seems to be happening here is the classic case of traffic quantity versus traffic quality. If you make yourself generically attractive, more people will be interested in finding more about you -- but it will be hard to filter the kinds of people who come to your door. So you'll get a lot of the people you do want (who are looking to convert) along with a lot of the people you don't want, too. As you get more targeted, a lot of people will lose a lot of interest in you -- but the people who do keep interest are more likely to be the people you're most interested in, too.
That's something crucial to keep in mind in paid search, where every click-through -- whether it converts or not -- costs you money. To increase efficiency, you've got to increase targeting -- and, very often, you'll get a higher ROI if you bring your overall traffic volume down (even if you bring your conversion volume down with it).
That targeting/volume tradeoff happens in every aspect of search: your keyword choice, your ad copy and even your landing page. And it also happens in engine choice. A click from the AT&T portal site is bound to attract people who are closely attached to that particular brand, and to the telephony world; meanwhile, the brand presence throughout the site will only serve to deepen those interests. Which is why, even though there are far fewer people searching on att.net than on Google.com, it's likely that a searcher for "cell phones" on att.net will be more excited about buying a cell phone than a searcher for "cell phones" on Google.com will be.
At the same time, there are a lot more people searching on Google.com than there are people searching on the AT&T/Google search bar. So while the conversion rates will be higher on att.net, the sheer numbers will push the conversion volume up.
Quality, not quantity -- sort of. Are we saying that you should pull your campaign out of Google and ship it to AT&T? Absolutely not. First of all, that's not an option for technical reasons. Google, which provides the search bar on AT&T's site, won't let you pick and choose amongst its syndicated partners. But even if it were an option to choose AT&T over Google, you still probably wouldn't want to take it. After all, if you want to achieve ROI and scale (and most advertisers do), then you need a good conversion rate and a strong conversion volume. Meanwhile, you can't build a good campaign without testing; and you need volume to create statistically relevant tests.
But what you do need to keep in mind is that there are strategic decisions to be made and optimizations to be made. The higher your volume, the harder it is to target and the harder it becomes to hit a good conversion rate. The more precisely you target, the more likely you are to build efficiency, but you also stand to lose the high revenues that come with scale. So whether you're talking about engine choice, keyword choice, ad copy or even landing pages, you need to find that perfect medium. This means both getting highly sophisticated -- and highly picky -- about which are the very best keywords, engines, ads and landing pages to hit that perfect volume/targeting combination for your campaign.
Which, by the way, is entirely doable. Just look at MSN Search. The third most popular spot for Internet search traffic, it scored No. 4 on December conversion volume, and only dropped down one space, to No. 5, in the conversion rate rankings. That ability to achieve volume and efficiency is probably mostly because MSN's AdCenter, which serves up ads based on demographic knowledge of its searchers, is highly targeted.
So a high volume and efficiency-driving targeting can be done. Which is fortunate, because it's what you have to achieve in search if you want to win.