Putting SEO keywords in the page title (a) doesn’t actually help your page’s rank in search engine indexes, and (b) makes things harder for people trying to tweet a link, bookmark your page, or scan it from a list of currently open windows and tabs in their browser. Trust the Googlebot to figure it out.
I'm pretty sure point (a) is false [1], and I'm absolutely sure that the conclusion (trust the Googlebot!) is. You should probably optimize for humans first, but not humans only. That's just not the way the web works.
EDIT: I realize the SEOMoz data is based on SEO experts ranking what they believe is most influential, but what else are you going to go on? Maybe patio11 or someone else on here who is an SEO whiz can offer some of their experience.
I hate saying "trust me", but since I'm mentioned by name: he is as wrong as it is possible to be wrong about SEO. It's so wrong, its like me suggesting that all the cool iOS developers add their bezels with Photoshop, which is why you can't use Flash on an iPhone.
In my experience, Titles do matter. Titles matter not only as a main "ranking factor", but also because the Title is most often the link from the search engine result to the webpage.
Having keywords in there means that those keywords will be highlighted/bolded in the result, making them stand out -- I don't know if that actually makes people click on links more often, but I wouldn't bet against it.
Also, using Apple as an example from the article, Apple's Title is Apple because people don't get to the Apple site by searching for "mp3 player". People go to Apple's site because they want an Apple product. Apple doesn't need to try and rank for all that other stuff; it's not what they do and they don't need it. Branding is different than ranking.
It is odd that Atwood calls these people "scrapers" when I'm pretty sure they are just using the dump file that you can download and then doing better SEO than SO (in the past, seems like SO now ranks ahead of them in most cases).
Apple’s homepage is used as an example, since its title is simply “Apple”, but other pages at Apple’s site are named fairly inconsistently: some use curly apostrophes, others use straight ones. Some have sentences ending in periods; others don’t. Some use sentence case; others use title case. Some pages just state what the page is; others read like marketing babble.
Examples:
Apple
Welcome to the Apple Store – Apple Store (U.S.)
Apple – Mac
Apple – Play music and more on iPod.
Apple – iPhone 4 – Video calls, multitasking, HD video, and more
Apple – iPad – See the web, email, and photos like never before.
Apple – iTunes – Everything you need to be entertained
Apple – Support
iPad – iPad WiFi – iPad Wifi + 3G – Apple Store (U.S.)
Apple – MacBook Air – It thinks and acts like a full-size Mac.
Apple – iPod classic – Read the iPod classic technical specifications.
Apple – iPhone 4 – Design of the display, A4 processor, and more
Apple – iTunes – What's on – Discover music, movies, and more.
Apple – Why You’ll Love a Mac – A Mac is the ultimate upgrade.
Apple – In-Ear Headphones with Remote and Mic for iPod and iPhone
Apple – Science – Inside the Image – Quantum Views
(Still, overall, Apple’s is better than most websites: it’s pretty clear what each of the above pages is about just from its title.)
You should see Google products. Matt Cutts did a Google SEO 101 audit and released the results once. Suffice it to say that a bunch of intelligent engineers given mostly-free reign over their development roadmaps do not necessarily converge on a consistent, competent onpage SEO strategy. (Not mentioned in the audit, but equally true: if you work at Google, you can do everything wrong and still rank fine, because your projects will have more link equity than God. Plus, if Google desires to give you more prominent billing on the Internet, there are very easy ways of making that happen.)
Or, he just looked around at a few major website homepages, and used them as examples of decent titles. Gruber’s example did a fine job illustrating his point, I thought – I just found it interesting that even Apple, a company known for its fastidious care for such trivial details, has inconsistent titles across its sprawling website(s), and found the inconsistencies themselves interesting, because they illustrate slightly different approaches to writing titles by the different teams creating the pages. To be honest, I find your sarcasm entirely unhelpful to this discussion.
I apologize for the snark. It was directed at Gruber, not your post. I'm actually glad you pointed out the incompleteness of his example, and the frustration of my post stemmed from how most forget to look for this in every other line of Gruber's writings and suck down the subtle shilling without thinking about it.
Great, so if you have a bunch of shrunken tabs, you just see "Apple" over and over. Usability experts have said for years to put the subject first in the page title and then tack on branding afterward if you must.
Safari actually handles this case intelligently -- it dynamically removes overlapping prefixes from tab titles. Looks like neither Chrome or Firefox do the same.
This is highly inaccurate. Title tags have EVERY implication on ranking for a keyword. For the most competitive keywords in the world, it's very possible that certain sites could rank without having it in the title tag due to the sheer volume of people linking to them with relevant anchor text, but this is highly unlikely. And more importantly, it doesn't make sense not to do it, because it's pretty ridiculous to put UX in front of strong SEO in title tags, because the large majority of users completely ignore the title tags - UNLESS they show up in search results.
And, yes, you can't show up in search results (most often) unless you have SEO-optimized title tags. I would recommend serving up a specific SEO-title tag, and then another, closely related article title on the page itself - at least as it comes to news pieces.
And the other hand, I'd agree that it is in companies best interest to have the title tags short enough to get their brand in there - because brand relevancy helps CTR substantially. When CTR goes up, ranking may increase since Google is looking at these things in their algo. Then you get more clicks, views and links, and it starts a downhill, positive domino effect for your website.
SEOMofo has a great tool for making sure you don't breach Google's title tag/meta description character limit in the SERPs:
We've spent dozens if not hundreds of man-hours trying to come up with reliable heuristics to get rid of "title junk". The state of the web in this regard is shameful.
There are many sites that don't even include the article title in <title> at all (even hackers like Zed Shaw get this wrong).
It is yet another thing that proves the rule that "metadata that isn't prominently visible in the browser is guaranteed to be wrong" (see meta description, meta keywords, and comments for worse examples. These often aren't even escaped properly.)
"metadata that isn't prominently visible in the browser is guaranteed to be wrong"
Very true. Still, I wish title junk was my only problem. My bank's online banking home page has a title of "Home". Just that. "Home". No "Acme Bank Home", not even "Online :: Banking :: Home". In 2010. Go figure.
Not just SEO. Adsense also pays a lot more attention to the title tag than anything else on the page (at least it did about a year ago, when I last experimented with it). If you want nice targeted ads, unfortunately stuffing keywords into your title is an effective way to get them.
Unless you (Gruber) care prominently about Apple/Google and other places with enough inbound link equity that your marginal SEO increase is minuscule and marginal UX*appreciation gain is large.
I think it's a fair question--once you trade even the smallest marginal piece of UX for a marginal piece of SEO ranking, you start down a slope. It may or may not be a slippery slope, but it's certainly not a level plane.
Apparently the Chicago Sun-Times does on its website, as do many other sites. " :: " is a simple visual separator that makes it clear that what's intended is not just a colon. It also prevents any confusion that can arise when you use a single colon or an em-dash as the title/site separator and then use the same punctuation element within a title.
If anybody wants to know the punctuation symbol that "::" is, it's called Paamayim Nekudotayim, better known in programming as the scope resolution operator. PHP (because of its Israeli roots) calls it by the Hebrew name.
Apologies for the pedantry I display below, but your comment is so full of errors that I got riled up. Again, apologies, I'm sure you were just trying to add interesting information for other hackers.
'::' is not a punctuation symbol, at least not in English. It is two of the same symbol (colon) in succession.
Now in Hebrew, it may be considered a single symbol (though a cursory google makes me think Paamayim Nekudotayim might just be the Hebrew term for 'double-colon'). If Paamayim Nekudotayim were a single symbol in Hebrew, the term has certainly not migrated to English, at least outside the PHP community.
In programming languages, it is commonly used as a scope resolution operator, and dates back to C++ at least. In PHP, they often call it Paamayim Nekudotayim. But this does not mean that you would be correct to come across the syntax in a C++ source file and call it Paamayim Nekudotayim.
So to summarize:
- it is not a punctuation symbol
- if it were, it would not be called that in English
- it very likely isnt a single symbol in Hebrew
- it is the scope resolution operator is _some_ languages
- in those languages, it is not called Paamayim Nekudotayim, except in PHP.
(and after that rant, I fully expect to be proved wrong on some of those claims, which I will deserve).
No, thank you, I was indeed being imprecise, you are correct that double colon doesn't exist as punctuation in English or Hebrew grammar so it's not a "punctuation symbol". I guess I meant "programming symbol", but it is also used in other contexts such as ratio equivalence in math, or abbreviating analogies e.g. dog:house::fish:bowl. I guess I don't know what to call it, really.
MSNBC is the fourth result at Google in a search for “news”, but its page title is so long that “MSNBC” doesn’t actually appear in the result entry title.
that is hilarious. i wonder how quickly you can a/b test a change on a website with regard to its pagerank.
The problem with pagerank is that it's exceedingly hard to a/b test.
There's multiple factors involved in this but here's a few of the obvious ones:s
For most sites your pagerank is typically determined by your root. (Unless google detects you're a hosting company like blogspot) So if you try to a/b test your leaf pages their pagerank is completely domainated by the root.
Pagerank is not that frequently updated. (Depending on how much traffic you have.)
There aren't individual and seperate googlebots to test against to seperate pageranks in tests so creating clean tests is very hard because Google very often updates their algorithms. (Try looking up ranking shuffles and you'll see what I mean.. every time there's a shuffle there's a furious scramble to figure out what changed and why and how to counteract it.)
Even if you could a/b test would you really want to? For now, a successful business only has to stay slightly ahead of the curve to trounce its competitors. Sure, a/b testing might get you valuable knowledge if your goal is to be an SEO consultant (well, one worth actually paying anyway, a lot of "SEO consultants" in my opinion read a few blog posts a couple of years ago and have been charging for that as gospel ever since) but if you're a company with a product it's probably going to cost you more money than it earns.
The problem with pagerank is that it's exceedingly hard to a/b test.
There's multiple factors involved in this but here's a few of the obvious ones:
For most sites your pagerank is typically determined by your root. (Unless google detects you're a hosting company like blogspot) So if you try to a/b test your leaf pages their pagerank is completely domainated by the root.
Pagerank is not that frequently updated. (Depending on how much traffic you have.)
There aren't individual and seperate googlebots to test against to seperate pageranks in tests so creating clean tests is very hard because Google very often updates their algorithms. (Try looking up ranking shuffles and you'll see what I mean.. every time there's a shuffle there's a furious scramble to figure out what changed and why and how to counteract it.)
Even if you could a/b test would you really want to? For now, a successful business only has to stay slightly ahead of the curve to trounce its competitors. Sure, a/b testing might get you valuable knowledge if your goal is to be an SEO consultant (well, one worth actually paying anyway, a lot of "SEO consultants" in my opinion read a few blog posts a couple of years ago and have been charging for that as gospel ever since) but if you're a company with a product it's probably going to cost you more money than it earns.
It saddens me when I work with a non-governmental organization (NGO) and need to copy some text from a website, only to find that the name and contact information is embedded in a graphic, with no text equivalent anywhere in the content. A simple search engine query usually confirms that the site's pagerank suffers.
The mere fact that SEO exists as an industry is an aberration and just a testament to the long way the search engines must go.
People that make business decisions based on SEO are putting the cart before the horse and will go away quickly. Fine-tune it, but don't obsess over it.
There are a lot of SEO snake oil salesmen out there, but that doesn't change the fact that there is a viable industry built around making your product more visible to your target audience. Until Google is psychic and can infer what you want by any search input, SEO will still have tangible and legitimate value.
Because we could usefully distinguish billions of sites by their tiny little icons?
ETA: To expand, I just opened the top 12 stories on HN right this moment in new tabs. I recognized the NYT logo in one favicon and the Apple logo in another. Of the other eight that did have favicons, those little icons told me absolutely nothing useful. A few of those sites did indicate their names in the titles, however.
An unfamiliar favicon lets me know I am not somewhere I normally would be, and to regard the site I am at more suspiciously. Without familiar favicons, I would treat every site that way before looking at the page more thoroughly. Obviously I do not rely solely on favicons, but they are a good initial way of orienting myself when looking at tabs.
In other words, the favicon is only useful for familiar sites or for knowing a page isn't from a familiar site. It's of absolutely no use in any other case, whereas a site name or a domain in the title tag will actually tell me what the site is in all cases.
I visit more than a handful. How useful do you think a dozen unfamiliar favicons on my tabs or bookmarks are?
And why would I have to visit billions of sites? The issue is that billions of sites exist, and a random little picture tells me nothing about what site it's from unless I already have it memorized.
Do you really want to distinguish a site by the title attribute? Think about it, a site already has two "branding elements": favicon and domain. Why not leave the title for just content, the thing that the site visitor is most likely interested in?
I don't know about you, but domains don't show up on my tabs or my bookmarks (unless I open one's properties). Only the title and the favicon do, and unlike the favicon, the title isn't a tiny picture that's only useful if I go to a site enough to recognize it.
That works for Tabs - I just spent 10 minutes watching my browsing behavior, and realized the only thing I look at in Tabs, is the Favicon.
Personally, Google has replaced my bookmarking behavior. I don't bookmark pages any more - when I need to search for them, I just google for the appropriate keywords. In fact, this works so well, it's rare that I even need to add a "site:" to give a hint as to which site to look for it.
I like how Chrome remembers the sites I've visited the most and autocompletes them in the nav bar. Usually I'll pop open a new tab, type a letter (e.g. "s" for slickdeals") and hit enter. 2 letters takes care of any site I visit on a semi-regular basis.
I bookmark pages I want to visit later and take advantage of the location field to search my bookmarks, as I am paranoid a www search engine might forget about my obscure page.
I use Google to search my own sites. Typing 'tip my' into Google takes me to my 'Tip of my Tongue' page faster than looking for the bookmark in my browser.
In reality it should work fine because searches also include the URL so you can easily see the source (same idea as used by HN items). Titles are easier to fake so can't be entirely trusted anyway.
As someone who has about 100 tabs open in Firefox at any given moment, I firmly believe that the description of page content must appear before the title of the website or any other SEO junk in the title. This is what I see in my Firefox tab bar right now: http://imagebin.org/128865
What does "Hacker..." tell me? Or "Timoth..."? On the other hand, "SQL Co..." at least lets me know the page is about SQL. Same goes for "n-gram...", "Sylistic..." and "HTML5".
The New Yorker’s home page title is shamefully sloppy:
National and world news, Profiles, culture, reviews, fiction, poetry : The New Yorker
The New Yorker — arguably the most precisely punctuated and copy-edited publication in the English-speaking world — would never use a colon like that (i.e., with a preceding space) in print. And why in the world is “Profiles” capitalized?
--
Who are these title-junk keywords aimed at? Google? Do you they really think that putting “breaking news” in their home page title...
--
Why "Profiles" is capitalized?. What does he mean by "Do you they really..."?
People make mistakes.
I agree with his point. Wish he didn't pick on small things.
Given that <titles> do matter for SEO, I wonder if there are any drawbacks to stuffing the <title> with keywords but, on page load, changing it to something short & meaningful, for a more pleasant bookmarking & browsing experience.
It's especially annoying when websites prefix their title by "Hacker News". Now all your tabs read like "Hacker News | Foo Bar [... other 6 relevant words omitted]". That's what the favicon is for.
"Surely, the name of the site should be the first thing (and in many cases, the only thing) in the title of the home page."
Surely, this man has investigated known facts about SEO before making this lengthy blog post about it. Surely?
The domain name alone will give you more then enough to go on regarding searchability for the name of your site. Several studies have shown that putting the branding last achieves better results for relevant searches than putting it first.
Perhaps the author should have tried Googling for "Breaking news" and then opening the top 3 sites and looking at the titles of each, before making false statements. I can't find any value in this blog post.
Complaining about proper punctuation in web page titles seems a little pedantic. Using colons and other strange "punctuation" is a substitute for graphics and styling since these are not available in title bars. For those cases where it is meant to be sort-of punctuation, I'd say titles on the internet simply follow a different set of rules and conventions than prose, or anything else prior, because they serve a different purpose.
The New Yorker (e.g.) is a publication known for its typographical pedantry, for example writing diaereses in words such as coöperate or naïve; it’s entirely uncharacteristic of them to have sloppily punctuated titles.
In this case, it's not typography - it's the closet thing they can get to graphic design using ASCII in one line and still being dignified. If instead of a colon it was a pretty little fig leaf we'd all feel more pedantically satisfied. Unfortunately, that's not possible in titles - so they use a colon.
Regardless, I'm guessing the editor of the New Yorker has been to their website. Her approval defines what it is that the New Yorker does, and what it is known for. Sniping at such an organization previously known for pedantry is a bit presumptuous.
I'm pretty sure point (a) is false [1], and I'm absolutely sure that the conclusion (trust the Googlebot!) is. You should probably optimize for humans first, but not humans only. That's just not the way the web works.
http://www.seomoz.org/article/search-ranking-factors
EDIT: I realize the SEOMoz data is based on SEO experts ranking what they believe is most influential, but what else are you going to go on? Maybe patio11 or someone else on here who is an SEO whiz can offer some of their experience.