 |
|
While we begin with the adtext, the searcher begins with a keword, the concept, or subject about which he is seeking further information. We need to anticipate what words a potential customer is likely to use in order to ensure that the adtext, which we have now refined, appears in response to his inquiry. Some of these words will be relatively obvious, and for these we use a strategy called head bidding. Others will be a little less likely, or somewhat more obscure, and for these the strategy we use is called long tail bidding:
a. Head bidding
Head bidding involves such words and phrases as will be of obvious relevance in the search for certain services and products, ‘London hotel’ or ‘running shoes’, for instance.
Though the words here may be apparent, careful consideration is necessary when bidding on them. Because of their obvious relevance such words are extremely popular and the competition to bid on them drives up their price.
Because of their expense significant losses can occur when bidding on such words. Such losses can rapidly accumulate if this aspect of the campaign is not being stringently monitored.
Not only close monitoring but also fraud overlay systems are necessary here. Since these keywords are so costly unscrupulous rivals can, without such systems, easily drive up costs by creating ways of clicking repetitively on the relevant words making them seem far more of an expense than they actually are.
Such complications aside, the number of potential keywords of this kind is extremely limited, making head bidding a relatively simple affair, relative, that is, to long tail bidding. While head bidding may involve even less than a hundred keywords, the potential keywords relevant for long tail bidding can run into the millions. To properly manipulate this gargantuan number of words, both processing power and specialized models are necessary. These in turn require operators with a certain level of expertise as well as a fair amount of experience.
b. Long Tail Bidding
The majority of words and phrases worth bidding on are not, as it turns out, immediately apparent. We locate some of these words in our clients’ data base. (For companies without either data feeds or product listings we have developed spider technology which can extract the necessary data.) We then introduce our proprietary algorithms which cross reference between the data listings and a variety of source references: dictionaries, thesauruses, and brand listings, for example. This way we source not only (all) relevant key words but also (all) relevant key word combinations, ensuring complete coverage.
While head bidding (as we saw) targets such word/phrases as ‘London hotel’ and ‘running shoes’, in our long-tail bidding we focus on more specific phrase/words like ‘five star, bed and breakfast, Picadilly’, or, ‘New Balance, cross country, for men’
One incidental advantage here is that these are keywords used by searchers who have already narrowed down their search (they know, for example, exactly what make of running shoes they are after) and are usually much closer to actually making a financial commitment.
The key advantage, however, is the relatively low cost of these key words. Because they are so much less expensive it is viable, in some cases, to bid on millions of them. Long-tail bidding is, strategically, like casting an enormous net: though only a few people are likely to click on any of these individual words, together they are usually more successful than the popular keywords which are in such high demand.
Head and long-tail bidding are two strategies we employ when deciding which words we will bid on and at what cost. Another choice, however, must be made. ‘Keyword matching’ determines how similar a searcher’s entry would have to be, to the word/phrase being bid on, in order to elicit the relevant adtext.
Next is Keyword Matching
|
|
|
