Sorry to Burst Your Bubble: The Long-Tail Keyword Myth

This article is a controversial one, with many PPC marketers believing that long-tail keywords are the holy grail of Google Ads and that they are the best way to get cheap high-converting traffic.

I am forever hearing from ‘PPC experts’ that they are able to double a client’s profit by taking all of their two-to-four-word keywords and turning them into long-tail keywords with four-plus-word phrases. 

This is based on the belief that long-tail keywords generate lots of cheap high-converting traffic, because they believe that 70% of searches are four-plus words long.

But in this article, I am going to explain (using real data from accounts that use our PPC management service and also data from an article by Sam Owen over at PPC Hero) why long-tail keywords are not the holy grail in Google Ads. 

It is simply not possible to double your profit by taking lots of short-tail keywords and turning them into long-tail keywords.

What are long-tail keywords?

Long-tail keywords are longer, more specific versions of keywords that are likely to appeal to people who are later on in the buying cycle. 

That is generally why they convert better and receive lower cost-per-conversions than head terms. Most marketers consider that long tail keywords are four or more words long.

For example, if you are a mortgage broker, your head term might be “mortgages” or “residential mortgage”; however, a long-tail keyword would be “buy to let mortgage brokers in Sheffield”.

Here, you will see that the long-tail phrase is much more specific and the user knows what type of mortgage they want and where they want it from. 

These users are late in the buying cycle and are highly likely to convert compared to people who just typed “mortgages”.

long-tail

Image source: SEOsitecheckup

It is generally thought that the exact-match head terms, such as “shoes”, account for around 10% of the search volume, with mid-length keywords between two and four words accounting for around 20-30% of the search volume, and the other 60–70% coming from long-tail phrases.

However, I will show you the real data from some of our accounts to show you that this is not, in fact, accurate. 

In reality, in most of the accounts that we have analysed, this is almost the opposite way round, with as much as 52% of the impressions coming from single-word phrases.

The long-tail keyword myth

There are three arguments that I am going to make, backed up by real-world data:

  1. From an optimisation-time perspective, your time is spent 33x more effectively by working on your top 20% of keywords as it is working on your long-tail keywords that contain four-plus words. Keywords that are over four words in length only account for 2.4% of the total number of conversions on the accounts that we analysed, so from an optimisation-time perspective, it makes more sense to spend your time working on the 10-20% of keywords that generate 80-90% of all of your conversions. For example, if you increase your Quality Score from 5 to 7 on your keywords that are four-plus words in length, this would reduce the CPC by 26% for these keywords (according to a study by WordStream). Overall, this would result in a 0.62% decrease in CPC in the account. But if you spent the same time increasing the Quality Score from 5 to 7 for your top performing keywords, you would get a 21% decrease in CPC for the account overall. 
  2. Over 90% of all impressions are generated by search terms that are four or fewer words long, so the long tail only represents 10% of all paid search traffic, not the 70%-plus that most search marketers believe. It would be virtually impossible to double your profit by making all of your keywords long tail, as you would lose 90% of your impressions and over 80% of your conversions.
  3. For keywords that contain five or more words, you would need 200 keywords on average to generate one click per month. I think you will agree that would prove very difficult to manage for only a small uplift in performance.

Argument 1: Your time is spent 33x more effectively working on the top 20% of keywords than your long-tail keywords.

Here is the data from the study that PPC Hero conducted that showed the performance of keywords based on their length.

screen-shot-2016-09-12-at-22-24-50

Image source: PPC Hero

One other thing to notice here is that keywords of five or more words in length only generated 15 out of the 608 conversions, which amounts to a measly 2.4% of the total conversions, which came from 138,638 keywords.

That is a lot of keywords to manage for just 2.4% of your conversions. It would make a lot more sense to work on optimising the top 20% of your keywords that will be generating 85% of your conversions, and making sure that they have a Quality Score between 8 and 9. 

This would have a much more drastic effect than adding 136k keywords that generate a few conversions per month.

For example, if you increase your Quality Score from 5 to 7 on your keywords that are four or more words in length, this would reduce the CPC by 26% for these keywords, according to a study by WordStream. Overall, this results in a 0.62% decrease in CPC in the account. 

However, if you spent the same time increasing the Quality Score from 5 to 7 for your top performing keywords, you would get a 20.8% decrease in CPC for the account overall. The choice is obvious.

screen-shot-2016-09-13-at-22-54-28

The graph below shows the percentage of the total conversions that were generated by different keywords containing various numbers of words.

screen-shot-2016-09-13-at-23-32-18

Here, you will see that only around 2.5% of all of the conversions have come from keywords that are five or more words in length, even though there is good long-tail keyword coverage with 139,889 keywords that are four or more words in length.

Argument 2: The next few charts back up the argument that if you change all of your short-tail keywords to long-tail keywords, you will lose 90% of your impressions and 80%-plus of your conversions. That means you can’t double your profit by relying on a long-tail strategy.

The pivot table below shows the performance of search terms by the number of words that each search term contains.

screen-shot-2016-09-13-at-23-05-51

Here is 30 days’ worth of search term data for one of our accounts.

From the search terms report, you will see that 434 conversions out of the total 532 were generated by keywords between one and four words long, accounting for 81.5% of the total number of conversions. 

You should also see that 233 conversions were generated by search terms that were four or more keywords long, accounting for 18.5% of the total conversions.

Furthermore, 94.9% of all impressions came from search terms that were four or fewer words in length.

The graph below shows the total percentage of conversions that have come from search terms by number of words per search query.

screen-shot-2016-09-13-at-23-13-50

The first takeaway from this data: by changing your short-tail keywords out for longer-tail keywords that are four-plus words in length, you lose 80%-plus of your conversions. That would make it virtually impossible to double your ROI.

Furthermore, you will see that 87,018 impressions came from the search terms that are between one and four words long out of the total of 91,716. This equates to 94.8% of impressions.

Here is the data from the study that Sam Owen did over at PPC Hero:

screen-shot-2016-09-12-at-21-49-30

Image credit: PPC Hero

Here, you will see from PPC Hero’s data that 1,028,543 of its impressions came from search terms that were one to four words in length, equating to 93% of impressions. 

By changing all of the short-tail keywords here to long-tail keywords that are four or more words in length, you would lose out on 93% of your impressions.

This shows that the ‘experts’ who keep telling you that only 30% of impressions come from search terms that are one to three words long are, in fact, wrong. 

From PPC Hero’s study, it found that, in reality, 51% of its conversions had come from single-word search terms and that 74.5% of impressions came from search terms that were between one and three words in length.

Furthermore, 1,071 conversions out of the total 1,253 conversions were generated by search terms that were generated by keywords that were between one and four words long. This equates to a total of 85.4% of the conversions.

So, from our own data and from the data of PPC Hero, it is fairly conclusive that by changing out your short-tail keywords for four-plus-word keywords, it is going to be impossible to recoup 80% or more of conversions that are generated by search terms that are less than four words in length.

Argument 3: You would need 200 keywords containing five words or more to generate one click per month, which would be inefficient to manage.

One of the biggest issues with catching all of these search terms is that you require a huge number of keywords to do this. If you look at the data below that Sam Owen over at PPC Hero put together, you’ll see that with keywords that contain four or more keywords, the number of impressions per keyword on average drops significantly.

This results in you needing to add a significant number of keywords to your account to be able to capture them, which is going to be nearly impossible to manage efficiently. 

For example, to generate one impression, you would need on average 10 keywords that are five words long; with an average CTR of 5%, you would need 200 keywords to generate one click per month.

screen-shot-2016-09-14-at-00-35-58

Image Source: PPC Hero

Even if you do succeed in adding all of these search terms as keywords, the truth is, as I’ve mentioned before, you are unlikely to actually increase your performance, because improved performance with more specific keywords comes from pairing them with more specific ads – but it is virtually impossible to make adverts any more specific than they already are with four-word keywords, because of headline character length restrictions.

When your keywords get to around four words in length, you are generally limited by the 30-character headline limit and are unable to make your adverts any more specific to the search query. 

This then results in diminishing returns: even though the cost per acquisition would be lower for long-tail keywords, because of the buyers’ intention, you could not reduce it by creating more specific adverts.

There is also the argument that you can bid more accurately if you use long-tail keywords. 

However, this is also likely to be inaccurate. To make bidding decisions, you need a large amount of data, and if your long-tail keywords are generating 0.1 impressions per keyword per month, you would have to wait about 100 years to get enough data to make a statistically accurate judgement, even if the keyword had a 10% CTR.

So, you are going to have to aggregate bids to make a statistically significant decision, just in the same way that using a shorter-tail keyword would use the same bid for a large number of longer-tail variants.

All in all, long-tail traffic from search terms that are five to seven words long is still likely to generate around 10-20% of your conversions and should not be ignored. However, adding huge numbers of long-tail keywords to try to capture this traffic is going to be a massive drain on your time.

Instead, spend your time on your top 20% of keywords that are going to generate 80% of your conversions, then build out your account with hundreds of mid- to long-tail keywords (three to four words) in BMM or phrase match to capture these longer-tailed enquiries as accurately as possible.

So, how long should your keywords be?

Generally speaking, between two and four words. In some industries, keywords that are one word in length do perform well.

When keywords become over four words in length, they tend to generate very few impressions and conversions, and you reach the point where the time spent optimising them would provide a much better ROI if invested in improving the Quality Score of your top 20% of keywords instead.

So, we suggest building out your keyword list so that the largest proportion of your keywords are in the three-to-four-word length zone, because that provides the best ROI when taking into account optimisation time.

Conclusion

Overall, when analysing data from our own accounts and from the accounts that Sam Owen managed over at PPC Hero, it was very quickly apparent that the long-tail curve that most marketers believe to be correct does not hold true for PPC marketing.

We found that around 94% of impressions came from keywords that contained four words or fewer, compared to the 30% that most marketers have been taught. This makes it virtually impossible to double your profits by making your account long tail, as most marketers believe to be possible.

When it came to my second point of argument, we found that 2.4% of conversions were generated by keywords that contained four or more words, in comparison to the 80%-plus of conversions that were generated by the top 20% of keywords. So, you would be 33x more effective by spending your time optimising these keywords as opposed to working on adding new long-tail terms.

Finally, when analysing the data, we saw that, on average, you would need to add 200 keywords of five-plus words in length to generate one click per month on average. That would prove very difficult to manage, and would consume a large amount of time for the very small increase in performance that you would see.

Of course, these stats will vary from account to account; however, we have not yet found an account where the main arguments that I’ve made don’t hold true. 

If your account is lacking when it comes to mid- to long-tail keywords, the search term report is a really good way to pad out your keyword list with relevant new keywords that you know already perform well with your audience.

With this being such a controversial article, I’m interested in hearing other people’s views on the study and welcome some debate, though I ask members to be respectful in doing so.


About Wes

Wes is the Managing Director of DemandMore. He is columnist for several leading marketing publications including Campaign Magazine, Econsultancy and Search Engine Land where he shares his expertise in search marketing.