Exact Long Tail Phrases in the Title are Not Ranking Well

[tedster]

There's an interesting side discussion in this thread Updating Question - no ranking movement despite on-page changes [webmasterworld.com]. I've extracted some quotes:

dvduval:
On page changes have zero effect even after waiting 2-3 months. This includes changing the title tag to the exact phrase we most represent

tedster:
I've notice recently that exact phrase match in the title does not seem to work well. It does work if you put quotes around the query phrase, but clearly only geeks do that.

MrFewkes:
Exact phrases in the title tag are (imho) a no no on longer phrases - say 4 or 5 words - if the competition is high. Now thats kind of a guess just browsing the serps. If you search a long string with quotes (a popular one) then get a feel for how many titles show in the serp - and compare that to a search without quotes for the same string you will generally not see very many of those full strings in there for high volume searches where normal distribution dictates a proportionally high volume of pages aswell (generally).

If however theres a high volume of searches but a low number of pages - quite rare - you may see the longer exact match strings in the serp without quotes.

I dont know the science behind this.

Indeed - I don't know the science behind this either. It does seem that Google's "fuzzy logic" has taken its toll here, and certainly exact text matching is long gone as an information retrieval approach.

In fact, even keyword proximity seems to be a minimal factor today, if it hasn't already gone extinct with its cousin, keyword density.

This seems to be directly in the area where seasoned search users have the most frustration with Google recently. The casual Google user, on the other hand, seems quite content for the most part - and there are a LOT more casual searches than seasoned searchers. That's who Google most wants to please.

So what factors do you think Google IS using (instead of keyword proximity) to generate those SERPs for non-quoted phrases?

[burcot]

Titles such as.. "Keyword|Keywords|Purchase Keywords Here"
seem to have lost their power.

I have noticed more and more structured sentence type titles doing well. eg - Compare the Best Low Rate keywords @ example.com

Not so long ago placing your keyword six words in to the title would be unheard of.

This obviously only applies to sites controlled by humans. Wiki still ranks with titles such as "Keyword - Wikipedia"

[aristotle]

Two thoughts:

1. I wonder if this could be one of the consequencies of the Mayday update and its downgrading of non-authority sites.

2. Another possibility is that the Google algorithm gives less weight to page titles now than it did in the past.

[steveb]

Exact matches of the first part of titles is the only way to compete for a phrase, but exact matches in later words in the title perform better if that exact match is actually in the body of the page, which normally is rare.

As always, these things are not black and white. Exact match of later words does obviously help, but it is a weaker scoring point for Google than domain authority, and exact match of domain name for most of the words in the search.

[doughayman]

I have found that this most recent update was my last straw with Google. Forget about SEO and my sites, I find that Google is no longer an adequate Search Engine for my search uses. I consider myself a moderate search user (I tend to use some quoted strings in my search), and the results returned, no longer have relevance, in 75 % of my searches. Moreover, the results returned to most of my queries all seem to be "recently" added pages (within the last month or 2). I think Google favors newer pages these days, and is tending to "ignore" older, more well-established authority pages, in my topics of interest. They have really blown this, IMO, and I now use Yahoo, Bing, and Ask, as my primary search engine sources. Remember guys, the Roman Empire didn't last forever, and neither will Google. A very clear case, of over-engineering.

[londrum]

i wonder if this is a possible reason why they're starting to favour new pages.

if you search for 'widgets' and get the same 5 totally amazing, but old, pages at the top every time, then people are bound to get to know them well. so next time they want the page the will go straight there. they wont need google, because they remember it.

but if google returns 5 new results every time then people will not get to know the pages so well, and will still need google to find them.

maybe that is a bit cynical, but if google wants to keep users coming back to them every day, then it makes more sense to weight new pages over old, even if they're not quite as good.

[indyank]

@londrum but how are they so sure that users will keep coming back to them if the search result pages are not so good :)

[yaix2]

G shuffles around results all the time. Those the users like will eventually float towards the top of the SERPs, others will go down. New pages are thrown into that process all the time, so its always moving. With user behaviour gaining more weight in the ranking process, other factors lose weight over time, such as the title.

[g1smd]

I made a throw away comment about this, several months ago. A new page on an established domain still does not show up for an exact match search for the page title - indeed, pages on more than a dozen other sites, pages that merely link to this page, show up instead. It's been another few months and still it does not show up; and that's just odd.

[sailorjwd]

I've been removing words from titles for two years now. A lot of titles have a single word whereas before they would have 3-4 words. I focus now on what is the single word that separates this page from similar pages.. i.e. rather than Ford Taurus it is now just Taurus. Rankings have gone up (quickly).

[Pass the Dutchie]

I concur with everything said in this and Ted’s referenced thread. Of late, SEO as I have known it seems to be having zero or the reverse intended effect. So far I can only conclude that one of these major factors is that Trust Rank has gone ‘nano’. It seems that if you don’t have the highest level of Trust Rank, or are directly associated or connected to an authority domain, or brand and continue to incorporate traditional SEO techniques chances are your pages will start to tank. We have also noted that targeted pages are dropping out the SERPS and are getting replaced by other, slightly off topic pages within the same domain. Older, established pages which have had very little or no recent SEO seem to be doing fine for both Long and Short Tail. However, this does depend on type of site i.e Static, Forum, News, Blog, E-Comm, etc. Results will differ across the range.
Currently no positive, lasting movement for over 8-12 months. We see new pages after months of (traditionally) clean SEO raise, and months later fall. We are seeing new patterns all over. Established site’s who have done well in the past with little or no active SEO have recently floated to the top.

Hate to admit it but for the first time I am baffled.

[Whitey]

I have noticed more and more structured sentence type titles doing well. eg - Compare the Best Low Rate keywords @ example.com

Relevance and authority - maybe Google is placing stronger emphasis on the context supporting the title found in the relevance of the page which also has some correlation to the authority of that domain.

I also wonder about Google's " bridging dictionary " where phrase A can be found directly as Phrase B . This has been common with mispellings in Meta Titles and on page content for over 18 months now.

[willybfriendly]

One SERP I am very familiar with has 1 1994 .txt file at #9. In the past, this one shoed up only when G was trying to "get it right" (i.e. it would drift in and out during dances, or when knobs were given a significant turn). It has now been sitting their happily for 2+ months.

Being a .txt file it has no title, description, nor even any html. No outbound links. Nothing but exactly what you would expect in a .txt generated in notepad.

It is an "authority" site in as much as it is a subdomain of a .edu university site. It (the subdomain) also benefits from many, many well aged links - but not so many fresh ones from what I can tell, nor does it show any editorial activity to speak of since 2002.

There might be some clues in there for the astute observer...

[Sgt_Kickaxe]

You want longtail exact matches in the content, not the title anymore.

There also seems to be some kind of freeze in rankings changes as if our sites have been graded and positive movement is much harder and slower. Fast indexing but much slower movement to the top 3 is what I'm seeing as well. Google wants this to help out the less knowledgeable webmaster.

[1script]

You want longtail exact matches in the content, not the title anymore.

Maybe not so categorically - just quickly looked through a couple of my site's GA stats via "Top Content" -> "Entrance Keywords" and the best keywords (phrases, really) could be found exactly as entered by the searcher in neither title nor content. I mean, all of the individual words comprising the phrase are there, some in title and some in content, some in both, some multiple times. Also, I should admit the meaning was still pretty much dead on but the exact phrase was definitely NOWHERE on the page and sure enough - if you enclose the phrase in quotes, my page does not come up.

This was by far not a scientific investigation - just a quick look to see if it warrants digging deeper but it showed right away that it's not quite that simple.

By the way - a side note to this mini-study: I noticed that the quality of results in quotes (exact match) very quickly deteriorates, much quicker than the loose match. First couple results are fine but by the time you hit #11 some strange results crop up - error pages, spam, keyworded subdomains, long query strings as part of URL and such. It's also not so very "exact match" right after you leave the first page. On the other hand, if you don't use quotes, you get 6 pages of good results (I simply did not go farther, there could be more). I guess there is some merit to the idea that G* does not really care much about geeks searching and using any of the operators (site: vagueness also comes to mind) - general public would almost never use quotes.

[netmeg]

What happens if you toss a date onto the page? Anything?

[maximillianos]

One SERP I am very familiar with has 1 1994 .txt file at #9. In the past, this one shoed up only when G was trying to "get it right"

Is the content relevant to your search? That is more important than it being a txt file or from 1994.

[jkovar]

I have noticed more and more structured sentence type titles doing well. eg - Compare the Best Low Rate keywords @ example.com

I think you should collect thousands of sample titles, apply something like Flesch Kincaid readability scores to all of them, and see if there's any sort of pattern.

[freejung]

There also seems to be some kind of freeze in rankings changes as if our sites have been graded and positive movement is much harder and slower.

I definitely agree with this. A damping factor is being applied somehow. Even downward movement seems to be slow - the fact that you used to rank well for a phrase seems to be a positive ranking factor even if you stop using that phrase (and if you think about it, there are all sorts of interesting possible uses for that).

I should have some good evidence on this during the next few weeks, as I've recently improved exact matching across a fair number of phrases. So far this doesn't seem to have hurt, but it doesn't seem to be helping much either.

Generally speaking, many recent observations seem to support the conclusion that exact match in title matters much less than it used to. I'm not sure I'd go so far as to say it actually hurts but it doesn't seem to help much.

Here's another interesting observation on semantics that stands out very clearly in my niche due to overlexicalization of the subject matter:

Suppose widgets, woozles, wingdings and doodads are essentially synonyms, with possible slight differences in connotation.

There is a point where if you rank well for "red widgets" (exact title match at the front of the title) it becomes very easy to rank for "red woozles" etc without exact match in title. Indeed, sometimes the ranking for "red woozles" will actually be better than that for "red widgets."

Changing the title to contain an exact match for "red woozles" does not seem to affect ranking.

However, here's the point: in some cases, it appears you have to have the word "woozles" somewhere on the page. There have been several specific cases where I ranked for "red widgets," "red wingdings" and "red doodads" but nowhere at all for "red woozles." On examination, I was using the words "wingdings" and "doodads" in the page (not an exact phrase match necessarily, just using the single word somewhere). Merely adding "woozles" somewhere on the page popped me up to the first page -- but not to the top 3, where I rank for the others. Thus my thoughts on the dampening effect previously mentioned. I bet in six months I'll be in the top 3 for "red woozles."

My takeaway from this is, if you rank for a phrase, you may rank for synonyms, but it's best to use those synonyms somewhere on the page.

[willybfriendly]

Is the content relevant to your search? That is more important than it being a txt file or from 1994.

Relevant? Perhaps. Dated? Certainly! Accurate? Questionable.

What is relevant is aged links to a subdomain on a .edu authority site.

What is also relevant to this particular thread is that the markup we might think influences ranking is entirely lacking on this particular document. No title tag. No description. No Hx tags. No li tags. Nada.

Just bare text served up as naked as the day it was created.

It is certainly not the best resource on the subject, but there it sits at #9, solid as a rock.

Will add - on another related SERP the #1 result does not contain the search phrase, but comes close with a noun in place of the verb in the search, i.e. 'dairy farm' rather than the search phrase 'dairy farming'

[freejung]

I think some experimentation is in order.

[gouri]

Generally speaking, many recent observations seem to support the conclusion that exact match in title matters much less than it used to. I'm not sure I'd go so far as to say it actually hurts but it doesn't seem to help much.

If you are trying to rank for a 4 word phrase, for instance, have you guys seen recently that adding a couple of words to the title tag helped to rank for the 4 word phrase?

Exact matches of the first part of titles is the only way to compete for a phrase, but exact matches in later words in the title perform better if that exact match is actually in the body of the page, which normally is rare.

Recently, I put a four word phrase that I am trying to rank for in the title tag (with no other words) and the same phrase is in the body of the page as well but I have not been able to rank for the phrase.

It is a little hard to make sense of that situation.

[mhansen]

Great topic.... Titles.

What I have found, are that titles seem to be broken into chunks, then ranked on how each chunk would rank, before delivering a result to the serp.

An example of this would be something like:

[brandname] [model-name] [product-type] [action words]

Before Mayday update, it would be ranked based on the phrase:

"BigName MOD8765 Air Conditioner Bargains"
After the update, it seems that each section of the title is being assessed THEN ranked.

HowDoesSiteRankFor - [brandname] = score
HowDoesSiteRankFor - [model-name] = score
HowDoesSiteRankFor - [product type] = score
HowDoesSiteRankFor - [Action words] = score

After these initial rankings are completed, then a final position is sent to serps.

I have seen looking through sites I manage (i know, me, me, me, sorry)
that simply pulling off the brand name, a serp can go from 40ish to top 10. The MINUTE you search with the brand, poof... gone.

Likewise... pulling out the [product-type] can take a bottom 50-ish rank, and move it to top 10 for [brand-model-action] phrases.

So I really think its a cumulative combination of how your page ranks for the "chunks" or the title, that determine where the page ranks for the whole title.

To us, the Mayday update was:

Complex.... fun (got me thinking SEO again)... VERY costly in both time and money!

[1script]

@willybfriendly:

Just bare text served up as naked as the day it was created

You give a great example with this .txt file. The pure distilled content of a .txt file should make it very desirable to bots because most modern HTML marked-up pages serve at best 40% of content and the rest has to be thrown away.

A .txt file is also easier to parse because there is nothing to parse, it's ALL content which is I think where it becomes very relevant to the theme of this particular discussion: title tags are and have always been in the most parse-able part of the page's code. I've always put them right after the <head> tag and I'm sure most of those SEO types hanging around here did the same. The lower you go on the page the more weird stuff like embedded CSS, scripts and so on gets on the way (and that's assuming the markup is error-free to begin with which it often isn't) and hinders the bot's ability to distill the content they are after.

This is a pure speculation on my part but what if importance of titles in the past was only brought about because they were always most available to G*bot, unlike the rest of the content of the page? Now, with the much celebrated Caffeine muscle G* may now be able to parse pages better (apply CSS rules, execute JS and so on) and therefore importance of title tags is diminished because there is just more other content from the page to index?

[gouri]

Now, with the much celebrated Caffeine muscle G* may now be able to parse pages better (apply CSS rules, execute JS and so on) and therefore importance of title tags is diminished because there is just more other content from the page to index?

1script,

That is definitely an interesing thought in my opinion.
------------------

I am working on something that is similar to what mhansen said and I would appreciate if someone could please help me?

Say, for example, that I am trying to rank for the phrase Signs of Getting Sick

Would it be better to change the title tag from Signs of Getting Sick
to

Signs of Getting Sick | Cold & Fever

or

Signs of Getting Sick - Cold & Fever

I guess these examples that I am providing would be considered Signs of Getting Sick - [factors]

Would being more descriptive (adding a couple of more words to the phrase that I am trying to rank for) help me to rank for that phrase (Signs of Getting Sick)?

Also, does it make a difference which of the two options I use (| or dash sign)?

[tedster]

Whether you use a dash or a pipe character is not likely to make any difference, gouri - not in my experience. And whether adding the phrase to your title will do the trick for you, well a couple months ago I'd have said yes. And certainly it's still very useful, but this thread is discussing the way things have changed recently. So I'd say you'll need to test it and not hope for a guaranteed answer ahead of time.

[Whitey]

I did some more digging around and noticed that sites with exact phrases in the meta titles, were slightly better than those with the phrases in the content. But only just.

On top of this i noticed the phrase on the better sites were contained in the URL , had matching internal link phrases and H1's with the phrase.

The better sites had higher quality backlinks into the site , but not into the site's pages.


The others simply had the phrases in the body of their content.

Given the very minor benefit all of this optimisation had on the better sites , I'm willing to bet that the only factors that mattered were the authority the sites had.

Anything with low authority , i feel will be ignored, or receive a low score using these tactics.

... I went away and did some more digging. Strange things :

A UK University page with a strong authority ranking ( sorry TBPR6 , linked from home page and with external links ) ranks No 1 in the UK for it's exact unique phrase contained in the meta title in Google.co.uk.

In the US it ranks 10. Similar story elsewhere. Yet there is no reason for it not to rank No1 anywhere in the World , especially as the phrase contains it's name.

So there appear to be some miscellaneous geographical factors playing into this as well.

Links? Place names ? Some attempt by Google to try geographic relevance whilst overiding the content quality maybe and indeed the meta title . Things seem a bit awry.

[edited by: Whitey at 8:50 am (utc) on Aug 31, 2010]

[seoN00B]

Adding |, : or - doesnt mean anything.

That kind of practice is for the visitor to understand. It has nothing to do with seo.

[tedster]

I think you're right. The Mayday update was aimed at sending long-tail traffic to the more reputable SITES, not pages. And there were too many bottom-feeders taking long-tail traffic and offering low value in return, no doubt about that. However, now it becomes harder to get started with a new site of good quality because the long tail traffic that used to start things up is not as strong.

[MrFewkes]

I was doing some medical research recently on goofle regarding the liver. In this search term of 4 or 5 words (I cant remember exactly) - the pages returned were ones with words in them - but no exact match - and hardly enough semi exact match.

The results were atrocious - insulting infact - in that where I was searching for medical information on the human liver - I was actually returned page on page of cooking a cows liver.

Dont forget - most importantly I issued the slackers with 4 or 5 word strings which were highly specific - in order that I may locate pages which were highly specific - hence the long search string.

I went to bing - and my search was nailed. Period - Job Dome.

Goofle should keep their shoddy work to themselves - and let professional programmers and language analysts get on with the job at microsoft.

Shoddy goofle - the shame of it is - many webmasters then start distorting their own pages to match the algo requirements to rank - and the net ends up a whole load of garbage dross and drivel.

Cooking a cows liver - I ask you.