In adwords, everyone talks about targeting the long tail, but what do they mean & why should you target it? The long tail is all the very specific keywords that individually attract only a few impressions & clicks, but which collectively are highly significant. Called the LONG tail, because if done correctly there are a lot of them. If I can illustrate with an example: If you are a company that sells cars, you could target the key phrase ‘car for sale’. In doing this, you would access about 160,000 searches a month, and pay around £1.20 for each click. If on the other hand he targeted “ford focus for sale” you would only access 12,000 searches per month, but would only pay about £1 for a click If you now repeated this for every other ford model “Ford Fiesta for sale”, “Ford Mondeo for sale” etc, and then for every other make he sells, the number of available searches would soon be building nicely, and the average cost per click would remain lower. This all sounds sensible, but there is an even better reason for doing this than simply reducing the cost per click. If we think about the buying process; when someone decides to buy a car, they start their research and search for something like ‘Cars for sale’. On researching they decide they want to by a ford focus, so they review the ford focus, and ultimately decide they want to buy a ford focus 1.6 estate. This is the point they are actually ready to buy, and they are now at the point where their search moved from info gathering to looking for a supplier. Thus targeting ‘Ford Focus 1.6 estate’ you will only access about 900 searches a month, but a higher percentage of these are likely to be interested in buying, and you will only pay 4p per click. Thus targeting the long tail will keep your click costs down, and mean you are more likely to be targeting the real buyers. Obvious really! Only downside is that it means building big lists of keywords for your campaigns, rather than targeting just 1 or 2, but there are some great tools for building keyword lists now, so this is not so much of an issue. You may also find this paper on the subject useful
