The importance of the ‘long tail’ in SEM campaigns

At a recent digital marketing conference I attended, I heard one speaker talk about the irrelevancy of the ‘long tail’ and how concentrating on ‘head’ terms was the key to building an effective campaign using search. This couldn’t be further from the truth, but does qualify for some explanation.

First, some facts from a Google presentation I got sent the other day:

Some analysts have even gone as far as to say more than 1/3 of all the queries Google processes haven’t been processed before, or in, the past 6 months, with Google’s VP of Engineering, Udi Manber pegging his guess at 25 percent.

Second, a definition (and for this, a picture of a dinosaur will suffice):

There are a few different ways to tell the difference between the head and the tail:

Length of Keyword: Head keywords are one or two words long, like: ‘porn’, ‘loan’, ‘flight’. Because head words are so short, they’re usually non-descriptive. ‘Porn’ for example, gives the advertiser no idea what kind of porn the user is searching for, is it ‘redheaded milf porn’ or ‘busty African femdom porn’? Tail words, then, have descriptors which enable advertisers to better understand the user’s intent, by typing in ‘flight from joburg to cape town’ the advertiser then knows that the person is looking to fly from Johannesburg to Cape Town, will show them the relevant advert and will take them to the corresponding page on their site.

A fixed number of Keywords: Some analysts have tried to make the case that the head is represented by the 250 highest traffic keywords for that particular industry, perhaps corresponding to the number that someone could manage by hand. This seems to be quite an arbitrary figure and an even more vague way of calculating it.

The number of keywords with more than X amount of traffic over a set period of time: Other analysts have tried to make the assertion that any keyword which generates X amount of click traffic in Y days is a head term, for example 100 clicks in 10 days.

Because I’m making the point about the tail being extremely important, I’m happy to accept any argument which serves to decrease the size of the tail. After all, if the tail still looks important when using the least generous definition, then it’s difficult to argue the point.

Onto our methodology:We decided to measure the impact of tail keywords across five different advertisers, chosen because of their varying industries; kinds of products; time of year etc and found that the tail performed differently for all of them.

We built out extensive keyword lists for all of the accounts using tools like our in-house Keyword Breakdown Tool and Google’s search query reports so that we could cover as much of the tail for each account as we could. The ads for these keywords were made as relevant to the keyword as possible, for example: If the keyword was ‘Cape Town to Joburg flight’ – we ensured that our client (in this case Kulula) advertised that the flight they were offering was from Cape Town to Johannesburg:

We found that of the client accounts we tested, the least important tail generated seven percent of the client’s sales during the period. In stark contrast, the most important ‘tail’ generated 92 percent of the sales — the ‘Head’ (top 250 keywords by search volume) netted only eight percent of its total sales! What backs up this finding is that the median tail for the group generated 42% of the competitive search sales.

What we found as a general rule for the long tail can be summarised in the following graph:

When applied correctly, the long tail increased Return on Adspend (Return on Investment) by providing highly relevant traffic at a very low average CPC compared to the head terms.

Overall, it’s important to recognize that users search differently for different products or services. Regardless of whether you’re trying to sell one or a thousand different products from your site it’s vital that you choose the kinds of keywords users are going to Google when they’re looking for your product.