If Ron Popiel were to invent a sponsored listing search engine marketing (SEM) tool, he might show you a way to stuff all of your keywords and phrases, titles and descriptions into a little machine available for five easy payments of $39.95 and tell you to "set it, and forget it." Sounds a little silly doesn't it? Advertisers wouldn't apply the highly successful Ronco Showtime Rotisserie and BBQ Oven mantra to SEM management, would they?
There are plenty of would-be search engine tools currently marketed and sold with a simplistic pricing structure and easy fix sales tactics, e.g. the "$99.95 gets you listed everywhere" search cancer pill pitch. Most, if not all, of those can be dismissed with a one-two punch of search sanity. However, one of the most interesting topics of discussion at the iMedia Search Marketing Tour was the all-too-common tendency for marketers to set up the paid search plan, check the search box as "completed," and move on to more important things.
Today's technologically sophisticated paid search management tools offered by search sites, paid search aggregators and third-party intermediaries can help manage bids and positioning, select keywords, and manage advertising return expectations -- which makes the "set and forget" tendency a lot more appealing. Overall, they can help bring the cost of managing a paid search initiative down, but leaving the tool in charge can have some pretty disastrous consequences. Let's take a look at the ups and downs of search engine marketing automation.
All too often, search marketers settle on keyword selection/management tools because they don't understand the unique requirements of search and a need for a simple solution. Checking the search box, as it were, should leave marketers wondering if automated decisions are the right ones.
A sound search marketing execution strategy effectively blends the efficacy and cost-savings offered by technology with the insight and experience of seasoned search marketing professionals. While bid management and keyword suggestion tools enable us to maintain optimal positions, minimize bid gaps, manage budget parameters and measure return, our program managers analyze campaign data and trends to optimize campaigns.
In the old days (last year) finding a search engine marketing tool meant signing up with a third-party bid management program that aggregated multiple paid search programs into one, easier-to-use interface. Overture maintains a list of approved providers of these services. Today, third-party bid management tools are just the beginning.
Search engine marketing is now more competitive than ever, and many specialist firms are developing their own tool-sets designed to reduce the cost of managing search programs. The paid search management space is in a constant state of evolution and keyword selection is now being married to systems that automatically generate descriptions and simultaneously remove keywords that do not meet expectations.
It slices, it dices, it will julienne search results
New search management technologies have many worthwhile advantages, critical in fact, for some advertisers.
"Automated keyword selection is important and efficient for advertisers with many products/models like resellers, distributors and catalogers," says Erik Matlick, chief executive officer of Industry Brains. "These tools are extremely effective in deriving detailed combinations of keywords using combinations of product/model names and marrying this with data from search engines to determine what is being searched for and what the advertiser has in inventory. I would go so far as to argue that not only is this a benefit to these advertisers, it actually puts their competitors who are trying to manage this by hand at a significant disadvantage."
So, manual stimulation of search listing selection is rapidly becoming a thing of the past. According to a recent Jupiter Research study, advertisers who used more than 1,000 keywords increased to 22 percent (March 2003 to March 2004) from only 4 percent the year before. Indeed, one of the key reasons behind this growth might be the use of keyword proliferation tools, but there are one or two problems with leaving keyword selection to technology.
Act now, quantities are limited
For the most part, automated systems cannot accommodate human search components and are based on algorithmic or computer-generated variational biases. That is to say, a keyword selection tool looking for matches to the term "Federal Express" will find multiple variations of the term, but not "FedEx." The same is true for commonly used abbreviations. Suggestion tools looking for matches and similar terms for "The Wall Street Journal" will not recommend "WSJ."
Brand marketers with compound words in their names may experience difficulties as well. "Dodgeball" brand searches are likely to occur as "dodgeball" and "dodge <space> ball," while keyword selection tools might not recommend the latter form. In May, 2004 Overture recorded more than 28,000 searches for the single word, "dodgeball," while the two-word format received just over 6,100 searches.
As of the writing of this article, the more frequently searched "dodgeball" is only $.03 per click more than the "dodge <space> ball" format. Often, an advertiser would find much lower click costs in the low volume keyphrase and most likely, very similar post click behavior, so it makes sense to buy them both.
New and improved space age formula
An automated system for selectively removing keywords for poor performance may not allow for trend analysis within keyword selection and messaging. Some systems will also not allow for multiple messaging elements to be considered as testing or trending parameters. A keyword is simply tied to one messaging element and removed if deemed ineffective. For the most part, automatic optimizers like these are designed to monitor key phrase performance and remove keywords that do not initiate a click or ones that generate click traffic from visitors who do not meet desired action requirements.
Dave Carlson, president of Atlas OnePoint (formerly GoToast) calls this process "modeling." "Modeling takes existing keyword sets and uses historical success criteria to map out keyword goals for campaign optimization," he says. "At minimum, advertisers must have 30 days' exact return on investment data. Ideally, there would be more performance history data because without it, November will be treated the same as March in a search program."
Often, otherwise effective keywords could be removed prematurely without allowing for positive trend analysis. A good example of this would be low-traffic keywords showing negative purchase behavior in the short term, which may prove valuable over a six-month time period. For example, the low-volume key phrase "liberate apes before imprisoning apes" may not move any View Askew merchandise for six months. In that time, 20 searchers may click on a listing result at a cost of $2. In month seven, when "Jay and Silent Bob Strike Back" re-enters the movie channel schedule, 75 searchers might visit the Secret Stash Web site, generating more than $700 in revenue.
Although a number of keywords may be deemed ineffective in the short term, that doesn't mean said keywords should be removed and replaced. Smart search marketers allow for trends, be they seasonal, tied to events or considered acts of HBO. Lesson learned: Keeping these keywords in your search program can prove effective over the long term.
Don't be fooled by similar looking products
Another shortcoming of many automated systems is an inherent lack of understanding in search term iterations from one search site to another. Also, many search sites do not compensate for plural or spacing disparities while others do. In this instance, having power over a keyword list designed specifically for each search syndicate is critical.
"Automated generation is a good thing, assuming the list can be controlled," says Kevin Lee, chief executive officer of Did-it.com. "Automated selection brings with it a loss in control. We must retain the ability to separate the wheat from the chaff."
Closely related to the differing term set problem is another small dilemma -- these tools also fall short in accommodating messaging performance analysis variations from site to site. For example, Google's creative requirements vary greatly from Overture's, but if you don't separate what works in a 70-character Google description from a 140-character Overture message, effectiveness measurement suffers.
A unique strategy has to be implemented across all sites and in these instances, the human variable starts to look highly favorable. Once again, keyword performance might vary across phrases with certain seasonal ties and other market conditions as well.
But wait, there's more
Search automation tools are in a constant state of refinement and enhanced evolution. Every day, they are getting better at addressing some of the issues we have touched on here. The question is, will there ever be a time when one can just "set and forget" a search initiative? Most search marketing professionals and search site functionality designers agree there will always be some level of need for human intervention in a search marketing program.
Although I have heard the perpetual need for a human role in search explained at least a dozen ways, I'll leave you this week with my absolute favorite, from Dave Carlson, using a commercial aviation analogy.
"Imagine settling into a pre-flight beverage aboard a large passenger jet bound for your Parisian vacation," he says. "A voice is heard over the intercom congratulating you for being randomly selected for the first fully automated, pilotless commercial flight. Although today's modern jets have very sophisticated auto pilot systems they can't handle unpredictable weather, hijacker intervention or the random errors of other pilots. Likewise, search program management technology has advanced, but it still can't fly on its own."
I don't know about you, but I'd rather swim to Paris.
Ryan proves geographic ignoramus
How catastrophically ironic would it be to get a geographic reference wrong in a world search article? Well, in last week's column, I mistakenly referred to Singapore as a Malaysian country. Actually, Singapore separated from the Malaysian Federation in 1965 according to the CIA world fact book. As Jay Shapiro, chief executive officer of Blue Interactive wrote, "In fact, you will find that Singapore is a full fledged member of the APEC nations (including Canada and USA) and is currently the No. 7 ranked country globally as 'e-ready' by The Economist (USA was No. 6) for the attractiveness of online business, adoption and infrastructure."
There you have it. I was wrong, and I am sorry for the error.
iMedia search columnist Kevin Ryan's current and former client roster reads like a "who's who" in big brands: Rolex Watch, USA, State Farm Insurance, Farmers Insurance, Minolta Corporation, Samsung Electronics America, Toyota Motor Sales, USA, Panasonic Services, and the Hilton Hotels brands, to name a few. Ryan believes in sound guidance, creative thought, accountable actions and collaborative execution as applied to search, or any form of marketing. His principled approach and staunch commitment to the industry have made him one of the most sought after personalities in online marketing. Ryan volunteers his time with the Interactive Advertising Bureau, Search Engine Marketing Professional Organization, and several regional non-profit organizations. Meet Ryan at Ad:Tech http://www.ad-tech.com/ July 12-13, 2004.