Search Engine Strategies, Chicago 2005 - Day 4
by Chris Boggs
Thursday morning came fairly quickly, much like Wednesday had. Fortunately, I had attended most of the sessions available at 9:00, so I spent the morning packing and getting ready to head home. The second session period provided a choice of good topics, but I chose to attend “Search Head or Search Tail? Getting the Mix Right.” This turned out to be a very informative session. For those that do not understand the title, a “head” and a “tail” in terms of search means the following: “head” search terms or phrases are usually no longer than two words, and most often just one word. A good example of a “head” query would be typing “rug” into the search box. A “tail” search term is longer and more descriptive, such as “oriental area rug,” or “Nourison round oriental area rug.”
This session was geared more towards using tail or head terms within paid search advertising, but, generally, the longer the tail is, the easier it gets to gain organic rankings. Also, these types of terms usually indicate a more sophisticated searcher, who most likely knows exactly what they are looking for. Longer tail terms tend to be used later in the buying cycle, after the searcher has essentially decided what they want. However, as the first speaker, Kevin Lee of dit-it.com pointed out, “Positive actions vary throughout the buying cycle.” Kevin had been introduced by the Moderator Misty Locke, of Range Online Media. He gave the attendees enough information in 15 minutes to greatly improve their chances of success in paid search efforts, and I thank him for that.
Kevin started by describing the value of tail terms, even though they may only be searched a few times per month. He explained that analyzing searcher behavior revealed a power curve, or one that that never reaches an axis. He valuates keywords based on their positions on this curve. Pretty advanced stuff, but I am sure Kevin or an equally qualified expert (not many of them) could explain in more detail if contacted (and maybe hired). He went on to teach the importance of bidding higher for longer tail keywords, since people are more likely to convert when using longer queries. He also wisely added that you must remember that you are competing with other bidders that may have bid for a head keyword as a broad match, meaning that their listings will also appear for tail terms, since the tail includes the head query.
Kevin finished with two more gems: when you are bidding on a head term, it is better to employ more advanced bidding techniques to include geographic targeting and day-parting, in order to cut costs. However, the higher ROI from tail terms prompts him to not use the same tactic, opting rather to maintain full exposure. His last point was to reinforce three main tenets derived from looking at head and tail terms separately: allows for reduction of waste, targeting of best customers, and increased profit. The next speaker was Harrison Magun, of Avenue A - Razorfish Search, and he took the whole topic to another mathematical level.
I have always been a proponent of the mantra that states that the “key to effective marketing is research.” Harrison provided ample proof that this was also the case when helping to decide which keywords to target in a paid search campaign. He started by showing an example describing how to rate the confidence of a sample. Hypothetically, he calculated that if 5% of the people in the room of 120 were twins (not necessarily both present), how accurate could we say that percentage is comparison to the overall population? The answer was that with that size sample, we could only be 20% sure that the results were indicative of the general population. It would take a sample of just over 5000 people in order to be 90% sure the results were a good cross-section.
This led him to his suggestion that for a paid search test campaign analysis to be accurate, it would need a large sample of data. To accurately define conversion rates of a particular campaign or ad, you need 25,000 clicks with a 1% conversion, or 2500 clicks with a 10% conversion rate to ensure 90% accuracy. So if an advertiser only has 400 clicks with 2% conversion, there is a 60% chance the analysis and decision based on it could be wrong. As Harrison says, if you are not willing to pay for the number of clicks required to gain such data, you might as well not even embark on the campaign. This matches nicely with what Darren Kuhn of Resolution Media said during day 2.
In my opinion, that is the crux of search engine marketing: the data derived from testing is more accurate than with almost any other marketing (pay-per-call is pretty close); so advertisers are most likely to gain the best ROI from large-budget search campaigns, compared to traditional medias.
The last session of the conference that I attended was the new “Measuring Success – Case Studies and Tactics,” moderated by Mike Sack of Inceptor, Inc. The first speaker, Geoff Karcher of the Karcher Group in Ohio, spoke fast and informatively. His main point was that it is just as important to measure failure as to measure success. He described the value of building trust with a client from the onset of the business relationship. The client should know what could go wrong, and feel confident that the outsourced company is looking out for their best interests. He then outlined a case study which I will not go into detail about. The bottom line was that through the careful examination of a shopping cart abandonment rate issue, and subsequent improvements to site usability, the conversion rate increased by 56% and return visits also increased measurably.
The next presenter was Kent Lewis of Anvil Media, who shared a case study involving a client in the hotel space. The most impressive tactic he detailed was adjusting PPC bids based on hotel availability. They discovered through analysis that the majority of one of the hotel’s clients were businessmen that booked on average 2 days before arrival. This allowed the hotel to increase bids two days prior to nights with availability, and essentially shut down the specific campaign two days before fully booked dates. This was a great example of using pertinent data to effect change within existing campaigns, and is in my opinion only the tip of the iceberg as measurement tools increase in their abilities to provide such valuable information.
The last speaker, Alan Rimm-Kaufman of the Rimm-Kaufman Group, delved into the topic of tracking various methods of online marketing, as well as protecting your brand. He explained that if people are bidding on your brand within paid listings, it is important to battle them, even if this causes what could be viewed as channel conflict with distributors. Furthermore, when creating reports, brand efforts and conversions should be counted separately from non-branded campaigns. Alan then described the importance of using different tracking for different online efforts such as email, affiliates, oragnic search and more. He advises to make sure orders are not being counted more than once and also to place test orders from “clean” browsers (both IE and Mozilla) that have been cleaned of cookies. Lastly, he suggests he would like to see dynamic barcodes created - similar to online coupons using static barcodes - that provide information such as search term, portal, etc… You can visit rimmkaufmann.com/ses-2005-12 for his presentation.
Well that’s it! The conference is over and I get to go home and see wife and baby tonight…oops not so fast! Blizzard hits Chicago, causing my flight to be cancelled (fortunately prior to my leaving the hotel- I heard many complaints about sleeping on carousels etc the next day when I left) as well as a plane to run off the runway at Midway, ironically from Baltimore. So I caught up on some sleep and left Friday, feeling full of fresh new information, as I always do when leaving an SES conference. Once again, I highly recommend this conference series to anyone that wishes to learn more about search marketing. There are events worldwide, and you can see an entire list of them here at the Search Engine Strategies website. As usual, feel free to email me with any questions at cboggs (at) instantposition.com, or comment in our forum linked below.
