Results 1 to 1 of 1
October 27th, 2012, 12:01 AM #1
SMS: Keyword Individualization Is Coming Of Age
- Join Date
- September 6th, 2011
- West Chester, Ohio
Search Marketing Standard has just published the following:
Keyword Individualization Is Coming Of Age
For many paid-search marketers, when it comes to bid optimization, performance is king. Optimization methodologies and solutions are evaluated through the lens of their ability to impact financial performance while meeting business goals within well-defined constraints. For this reason, traditional rules-based optimization increasingly is being replaced by model-based optimization techniques, thanks to their ability to drive higher performance. However, even as model-based optimization methodologies emerge as the industry standard, not all flavors of model-based optimization are equal. Of the three most common model-based optimization techniques—global cluster level, local keyword level, and global keyword level—global keyword-level modeling packs the greatest punch in driving the performance increases search marketers seek.
Why keyword-level modeling improves performance
In a nutshell, global keyword-level modeling combines the benefits of individual keyword-level modeling, which analyzes each keyword individually, with the benefits of global optimization, which strives for the success of the campaign (or portfolio) as a whole. With global keyword-level modeling, each individual keyword is analyzed and bid individually, in service to the overarching global objective set by the marketer. Tying the ability to model at the individual keyword level, all the way through the extreme tail, to the marketer’s overall financial objectives and constraints offers the greatest potential for meaningful performance lifts.
Despite its clear advantages, global keyword-level modeling has been the most elusive to implement until recently. Many have tried, but have fallen short for a variety of reasons.
To start, keyword-level modeling traditionally has been limited to head terms where data is plentiful, making the manual building and tuning of models relatively easy. In tail terms, where data is sparse, clustering typically has been required to aggregate enough conversion data to make modeling possible—not entirely accurate or forward-looking.
Furthermore, the magnitude of analysis required for global keyword-level modeling presents an operational challenge for most organizations. Global keyword-level modeling requires substantially more manual number crunching than clustering or local optimization and significantly more mathematical permutations. A team of the most accomplished analysts would find the task daunting, if not impossible. In addition, it is difficult for even the most mathematically-inclined search marketers to even begin understanding their keyword data beyond the head, given the sheer amount of data. Without the ability to interpret this massive amount of data, marketers lack the ability to use it effectively to drive better performance.
Why is it coming of age?
The biggest driver of keyword-level modeling adoption today is its ability to act as a predictor of performance at the most granular of levels while working to achieve the goals of a campaign. All of this can translate into dramatic increases in paid search performance. On the other hand, cluster modeling, designed to create more data out of the “data-light” tail terms, by definition falls short in addressing the individual characteristics of the similar keywords it groups together—thereby negating the power of head-to-tail optimization in driving higher performance.
Fortunately, for search marketers for whom performance is paramount, advanced technology now makes the inherent complexity of global keyword-level modeling a reasonably simple and eminently more viable proposition. Advanced bid optimization technology now can handle the deep mathematical modeling without human intervention, relying solely on keyword performance data to inform optimization. The most advanced software can even do this global keyword-level modeling through the tail for terms that convert as few as ten times per year. This negates the need for the less-than-optimal practice of clustering, which not only requires some human management but also falls short in terms of the performance it delivers.
So, search marketers, breathe a sigh of relief that you don’t need to go back for an advanced mathematics degree in order to fully optimize your paid search programs from head to tail. The biggest factor in effectively implementing global keyword-level optimization is choosing the right bid optimization tool—one that can do all the heavy mathematical modeling lifting for you. From there, it is a matter of determining your campaign goals to align with your overall business’s goals and setting them up in your system. The right software will take it from there, and you can sit back and bask in the glow of better paid-search performance.
Did you know that we publish a print & digital magazine? Subscribe by October 31st to receive our Winter issue. More information and* subscribe using Coupon Code SMS2012P to save 25% off the regular price of $34.99/year ($5 extra for international) — digital-only subscriptions are $29.99.
Image: Math by Shutterstock
Save 50% on Search Marketing Standard Magazine; use Coupon Code "ABW50" - Offer expires 12/31/2012
Read More From Keyword Individualization Is Coming Of Age...
By brandonmbeard in forum SEO ArticlesReplies: 0Last Post: February 1st, 2013, 12:01 AM
By brandonmbeard in forum SEO ArticlesReplies: 0Last Post: September 24th, 2012, 12:02 AM
By PingoPrepaidCallingCards in forum Commission Junction - CJReplies: 9Last Post: March 21st, 2007, 11:45 AM