Why Keywords Suck

Paid search has proven to be one of the highest ROI channels in digital advertising, but keywords are both the life’s blood and bane of an SEM’s existence.  With an infinite number of keywords that could be associated with any retailer’s product catalog, it’s damn near impossible to predict and manage all the right keywords that will maximize your revenue and return on ad spend.

Keywords promise SEMs pin-point accuracy as to who they're marketing to, how much they're willing to pay to get a click from that person, and what ad they want to show them.  However, as any SEM will tell you: keywords actually suck. Here are five reasons why:

1.    Keywords don’t scale. Growing an SEM campaign typically requires adding more keywords, and adding keywords requires hiring more SEMs, or demanding SEMs work longer hours. Every keyword demands the same amount of work to get live.  Without doing the additional work for each additional keyword, relevance of ad groups, ads, and landing pages will start to suffer, artificially depressing ROAS.

2.    Keywords don't map cleanly to products. One keyword can market multiple products in your catalog (e.g., "blue jeans").  Conversely, many keywords can be used to market one product (“Calvin” “Klein” “blue” “jeans”).  Keeping track of which keywords map to which products is a complex, ongoing problem for online retailers.

3.    Keywords are organized in a completely different way from how a business is organized.  Google and Microsoft force SEMs to organize their keywords into three tiers: 1) Account, 2) Campaign, and 3) Ad Group, even though few retailers organize their product catalogs or stores in a three-tier hierarchy. This makes it difficult for SEMs to see trends that might span different ad groups. For example, a certain designer or style might be popular and converting better, but this could be difficult to see. Furthermore, it might be difficult to implement certain promotions. An ad for "2-for-1 women's Levi's Special" would need to be manually added to (and eventually removed from) every ad group with women's Levi's keywords.

4.    Keywords leave marketers in a Catch-22.  Marketers find it difficult to make money on high-volume head keywords because they are more competitive and, as such, more expensive. Conversely, it’s hard to manage low-volume tail keywords because their workload grows proportionately with the number of keywords in their campaign.  The marginal benefit of adding more tail keywords to a campaign doesn't outweigh the marginal cost of managing those tail keywords, despite the fact that the long tail keywords tend to provide better results.

5.    The more products you have in your catalog, the more potential keywords you can bid on.  As a result, it's difficult for retailers to market their entire product catalog -- especially with long-tail keywords -- leaving a lot of search advertising opportunity untapped. This is particularly difficult for retailers with catalogs that change frequently.

Unfortunately, instead of tackling these issues head-on, many SEMs have developed a laundry list of coping mechanisms -- such as depending primarily on high volume head terms, overuse of broad match, or letting ad groups become too large -- to deal with an unwieldy number of keywords. These shortcuts are band aids and inevitably prevent search campaigns from scaling for real growth.

There are existing assets that retailers can and should use to discover new keywords and create more relevant ads. Product catalogs, on-site search query logs, and broad match query reports for Google AdWords and Microsoft adCenter, for example, can help increase exact match keywords to serve more relevant ads to consumers.

However, the only way to really make keywords suck less is to take a step back and focus not on the keywords themselves, but their underlying intent. Here’s how it works. Let's say I sell Nike men's basketball shoes.  Instead of manually creating a list of thousands of keywords, grouping them, and writing relevant ads for each of these ad groups, I can manage just one intent -- "Nike men's basketball shoes." With available technology, it’s actually possible to generate nearly every traffic-producing keyword related to Nike men's basketball shoes, group them by similarity, and easily publish highly relevant ads to each group.

Keywords have helped the paid search industry grow to where it is today.  However, keywords are also the primary impediment for the paid search industry to achieve its full potential.  The sooner the paid search industry can move away from keywords alone and provide advertisers with a less complex and more scalable unit of purchase -– like intent -– the faster the entire industry will grow.  And the less keywords will suck.

Tags: keywords, search
Recommend (12) Print RSS
3 comments about "Why Keywords Suck".
  1. Kevin Bullard from ILFUSION Creative , May 18, 2012 at 11:51 a.m.
    Keywords don't suck at all. It just takes work to b successful. Is that news to anyone?
  2. Timothy Daly from Vincodo , May 18, 2012 at 11:17 p.m.
    Unfortunately, intent cannot be determined by simple quantitative analysis. If it could, then we'd live in a world of 100% CTR and 100% order/lead conversion. Until that day, we'll continue to love keywords over black boxes that overpromise and underdeliver.
  3. Rob Schmults from Intent Media , May 22, 2012 at 12:24 p.m.
    Good thought provoking piece Thi. Two thoughts it provoked in me: First, most marketers have an implicit or explicit sense of "return on effort" when it comes to SEM. They spend time not just in pursuit of ROI (which as you say has the potential to be never-ending), but the right mix of volume AND ROI. Striking the correct balance there is central to an good SEM effort. Second, in regards to #3, it will be interesting to see if the continued evolution of the search engines' front-end experience will trickle back in terms of similarly continuing to align around the consumer and their intent. This is a better answer to the mismatch between search and business lack of taxonomical alignment. If search engines were to organize their hierarchies the way retailers traditionally organized their product catalog taxonomies, it would be a short term win for some marketers, but be counter to the increasing trend of retailers to rework their taxonomies in a more consumer (vs. merchant) centric model.