One thing to consider is how Google now deal with uncommon searches, they often drop words and show what's missing from individual results if the combination of words gives few results.

Another thing is how does an individual website collect enough information to see what's unique to Google, I don't think you can - so the 15% figure is very interesting in that it pushes people to take care with all queries they can collect (not that it's as easy/cheap as it used to be). Outside services start to look attractive at this point, ones which do the legwork for you in terms of monitoring the popular searches.

I love the long tail, but there's another part of unique searches that can be overlooked; emerging vocabulary or descriptions of derived products/services. You can identify opportunities if you can spot the changes in how people describe their needs. Many people do take advantage of such information, are you?

Do you keep an eye on new Wikipedia pages? You need to be pretty keen to go as far as monitoring the page view stats on an hourly basis to spot popular new pages quickly but identifying pages that are in your business area can be a real opportunity. If your business caters for people who need to do basic research before looking for a supplier then you should also be looking at how WIkipedia articles describe what you do or sell. You can sometimes even spot people who have used Wikipedia before searching by the terms they use - that might tell you something about their stage in the buying process.

Do you take care to look at Google autocomplete suggestions? If you take relevant autocomplete phrases for your site and then think of single word modifiers that people could reasonably add you are likely to catch a fair bit of traffic. The number of searches that I have seen that are known autocomplete terms with something added to the end is huge - well worth the effort.

Long live the long tail.

But wasn't the original purpose of Panda to demote "content farms" like Ehow, that adopted a strategy of creating huge amounts of low-quality content which could act as a kind of giant basket to catch stray long-tail searches.

I've never tried to write anything solely intended to get long-tail traffic. But it comes naturally if your articles have a lot of detailed information that's hard to find on other sites.

Re Panda, creating huge amounts of pages like that is indeed a poor strategy. The long tail has much more to offer than many people think. The true value isn't a few clicks on adverts, it can give a great deal of insight without costly research.

drop words and show what's missing from individual results if the combination of words gives few results

... which kinda defeats the purpose of adding those exact words to your search in the first place, doesn't it?

But then, what's "unique"? If one student copies-and-pastes the entire text of their assignment* into the search engine, that's a unique search. Two from the same class, and it's no longer unique.


* This is literally true. Sometimes I'm tempted to paste the string right back into the search engine and see what school they're wasting everyone's money at.

They don't say that those 15% then recieve unique results. In my vertical you can get as unique and off the wall with your searches as your imagination can get and the results stay the same. I am not compaining because i show up in those results but i don't see unique searches getting different results.

What used to be keywords are more like trigger words now. If a trigger word is in your search, you get the same pre-determined results regardless of what else you include.

I changed to the PPC model of targeting for organic results and simply don't do an negative word list.

If my targeted phrase I want to rank for is search engine optimization

I target search
I target engine
I target optimization

And Google does what is does exceptionally well: Matches results to queries!

So if the query is:

SEARCH company that can rank in Google the ENGINE giant with the best OPTIMIZATION practices

My PPC positive words hit #1!

Nov 2008 [adwords.blogspot.ca...] (20% are ones we haven’t seen in at least 90 days, if at all)

Dec 2010 [googleblog.blogspot.ca...] (20% unique)

May 2015 [youtube.com...] (15% unique)

It depends mostly on the site's niches.

I still think that a certain number of people still post "question form" queries in Google, like "How do I get rid of acid reflux?" instead of "acid reflux remedies". This is only an example by the way.

These kinds of queries are unique in some way, not a lot of people use them, there's no competition for these kinds of queries - which in turn would render me to optimize keywords like "get rid of acid reflux". I may not get the lion's share of the market, but at least I'll get the audience I meant to reach in the first place.

I may not get the lion's share of the market, but at least I'll get the audience I meant to reach in the first place.

Google's Knowledge Graph answers many of these types of questions. Once the KG starts answering questions your site is currently answering, you will likely see your audience shrink considerably.

These kinds of queries are unique in some way, not a lot of people use them, there's no competition for these kinds of queries - which in turn would render me to optimize keywords like "get rid of acid reflux". I may not get the lion's share of the market, but at least I'll get the audience I meant to reach in the first place.

Can you really count on that in the post-Hummingbird "conversational search" era, when (to borrow a phrase from Matt Cutts) Google is focusing on "things, not strings"?

]Can you really count on that in the post-Hummingbird "conversational search" era, when (to borrow a phrase from Matt Cutts) Google is focusing on "things, not strings"?

Yes by taking hints from Google's Playbook e.g. Adwords. They suggest target broad match so "things" are one word answers. But targeting 5X one word answers is also a string... and not just a string but every string that contains those five things.

Matt's analogy is about Adwords not organic search. Google is focused on PPC.

On average 15% of all consumers are searching Google with their own custom search word(s).

Presumably that would include misspellings as well as gibberish. I see it in my logs daily, probably the results of "also including..." search returns. Depending on site theme, that variable could even be higher.

In May 2015, Google announced that 15% of all searches are unique. Think about that. On average 15% of all consumers are searching Google with their own custom search word(s).

REALLY! 15%?

I've always done that! Ever since way back when. I thought everyone did that?

No wonder I always get immediate search answers 95+% of the time?

People actually search for "Widgets", without any qualification?

Blow me down, proving you can still learn something new everyday. Oh in the past, when I bothered to check search terms to my site? They were all long tail. That's why I thought everyone did it. And I bet every reader here does exactly the same.

My advice - don't waste your time trying to plan for longtail keywords, and I have an example to share that explains why.

I have a 9 year old site about "widgets" that used to get xxxx visitors per day until 5 years ago. 5 years ago I wrote an article that offers a solution to a widgets problem that was, and still is, hugely linked to and considered authority by search and by visitors. The site now gets xxxxx visitors per day, 90% of which is to that one page. My competitors have created countless variations of my page since then to no avail.

The incoming link anchor text, according to GWT, is so varied that most of the terms do not appear on the page at all. 2 years ago a NEW model of widgets was introduced to the world, complete with a new name that I obviously did not mention in the older article, but the solution still applies. Google now credits hundreds of variations that include the new name as pointing to that page when NONE of the terms are on the page.

So it would seem that you do not need to actually have exact match text on a page to rank for a longtail keyword. That being the case don't go trying to target longtail phrases, it will end up sounding like spam if you match even 0.1% of the likely variations that will eventually be associated with your pages.

exception: a brand new page with no history will only rank for the terms used somewhere on the page, and slight variations thereof. Since that's the case your BEST bet is to create quality content so that people can share it and expand it's reach in terms of longtail keywords.

Obviously if you are well establish longtail makes lillte sense so the alternative is what when you are not well established? Just become well established.

I love Matt Cutts video on SEO & Wordpress [m.youtube.com...] get well established by starting small, a niche, within a niche, wthin a niche if you must.

I've read the thread with minor interest... and there's some good value in much of what has been discussed, but I don't see this as a long tail thing at all.

G says 15% of all searches are unique.

Okay. Unique for who? Us or them?

Back when we got reliable referer (sic) info, dang near 99% of those were right on the money. I suspect that number has done down some over the years because the web is gizillion times larger now than it was just 7 years ago. Still, what hits me is spot on. And is greater than 85%

Not sure I want to chase a 15% I know nothing about (might be aliens from another world, for example) and my sites/products would have no interest (or conversions). but 15% of trillions of searches means there's plenty to go around.... and most of those were not destined for "your" (that's we/our) websites anyway.

Some long tails just ain't worth chasing.

So you're saying "unique to Google" which suggest unique to "Adwords Publishers" (I presume are included? Not sure why Google would release privileged information to the Adwords public if not to gain some addition return.) ...and with the right organic approach ... unique to all advertisers using Google as an advertising platform.

My question here... is that 15% per Adwords ONLY?

Thus if Google's Search Traffic is a 50/50 split (for example only) that is 15% unique in Adwords & assuming organic gets as much attention 15% for organic and thus your 85% lock is only 70%.

Which unless you are running Adwords & optimizing at the same time your lock is only 35%.

The source video came from the Adwords team, so it may be pointless information without Adwords running.

<added>The reason I point this out Trends show the reverse effect suggesting the levels of unique queries are increasing, year by year.

If you take any single word query and drop it into trends the year over year search data is diminishing suggesting searchers are getting more sophisticated by getting more descriptive.</added>

I found that chasing longtail was a waste of time and money at least in Adwords.

Logic would assume that you could eventually get enough negative keywords to weed out the chaff, but mathematically I think the possibilities are infinite.
In other words, you will just keep getting people searching for things that are not relevant. For example, "widgets in Boise, ID" no matter how many negatives you add.

Another consideration is click fraud. It's much more likely a competitor or robot could go down the line with a broad keyword and deplete your budget. While with an exact match it would be much harder to do or take more effort making it less likely.

Logic would assume that you could eventually get enough negative keywords to weed out the chaff, but mathematically I think the possibilities are infinite.

While there may be an infinite number of negative words there will also be an infinite number of positive words which also means an infinite number of conversions.

Interestingly, I went into Google Analytics for my website, then Acquisition > Channels > Organic Search, changed 'Show rows' to 5000 so I could see all keywords, added up the keywords with sessions count of one, and then divided that into (session count - (not provided)). Answer for my site is 17% for past month.

Looking at searches where I received 1 session for a search term, I can't believe they were unique for all searches to all sites.

Lies, damned lies, and statistics
[en.wikipedia.org ]

I'm Reading Web Analytics 2.0 by Avinash Kaushik again by the way. Written in 2009 and still current. What a classic.

While there may be an infinite number of negative words there will also be an infinite number of positive words which also means an infinite number of conversions.

There is a finite number of sales for any given product or interest in any given topic. When the buyer keywords are heavily saturated or diluted by Google's own products or services, conversions are still possible but the rate at which those conversions occur is much lower.

Conversions are still possible but the rate at which those conversions occur is much lower.

True but many aspects are equally lower including the ease of capturing due to lower competition where there is less need for natural links as on page merits are enough or internal links passing PR around from elsewhere. While your conversion rate can be extremely high (per phrase) as it is 100% or 0% when there is only one click.