Pixel Acres

Ditch your meta keywords

One of my clients is currently optimizing their website for search engines. As part of the process I met with a rep from an SEO firm, who suggested ways the website design could be tweaked to be as friendly as possible to search engines. Most of the ground he covered was not new to me, but one tip was an eye opener: he advised me to entirely remove the meta keywords tags from the website.

I’ve known for a long time that meta keywords do nothing to help a site’s ranking in Google, as the keywords selected by a webmaster might be misleading or downright false. What I didn’t realize is that Google and other search engines may in fact penalize sites that they believe are cheating the system with their meta keywords. For instance, a website that specifies meta keywords that don’t actually appear on the page may in fact have their ranking decreased.

This is significant if you are in the habit of specifying the same meta keywords for every page on a site, perhaps as a global include. Or maybe you forget to update the keywords whenever a webpage changes, and your keywords no longer match the on-page content. While these scenarios may seem harmless enough, if search engines truly do penalize sites for sloppy meta keywords, then you are much better off to ditch the keywords altogether.

Simply put, meta keywords can never help your search engine ranking, only hurt it.

Note: It is important to make a distinction between meta keyword tags and meta description tags. Google uses the meta description tag to display within its search result listings, so it is still very useful to provide potential visitors with a succinct description of a page’s content, and a unique meta description should be included on every page of a site.

19 Responses to “Ditch your meta keywords”

  1. Dustin says:

    Do we have proof that this is true?

  2. Georges says:

    I always thought quite the same. I come across websites wich have more than 50 Keywords in their Meta-Description. Most of it just nonsense to rank up (what they tought).

    But since everyone knows that Google and Co not just scan the meta tags, it’s obvious that they scan the content to produce a page by page index.

    But nice to know that you can even be penalized by SEs for that :)

  3. Jonathan says:

    @Dustin: I don’t know if you can ever prove categorically the criteria Google use to assign page rank, but if you accept that keywords won’t boost your page rank then there is no harm in removing them from your pages. I guess it’s a case of “better safe than sorry”. This tip was related to me by someone who does SEO work for many of Australia’s largest companies, so I figure he probably knows his stuff. If you check out the wikipedia entry for meta tags, it says something to the same effect.

  4. Dustin says:

    Wikipedia says a lot of things some I assume might even be based on truth.

    The real question is do all search engines work this way.

  5. @Dustin – From what I have noticed from my stats collected the only other search engines that get used is msn.

    Anyone have any ‘offical’ stats on google’s popularity?? over 90% hey???

    @Jonathan – Great insight, I am at the moment helping redesigning for ICWA (this is the old site) and we are using keywords etc in our meta tags. The departments are doing the meta tags and its amazing just how many different keywords they have suggested. Most of the time its longer than the actual description.

    So its no wonder that google (wisely) do not count the meta tags in their page rank…

  6. Sean Fraser says:

    Friendly Bit said this was a good site to visit. I visited and I agree.

    The other thing to consider when using meta: descriptions – for Google – is to make them suit the individual page. There is growing speculation that they are penalizing sites in search rankings when page descriptions are identical for all pages. Sadly, some CMS setups only allow a global description.

  7. Jonathan says:

    @Jermayn: According to Nielsen Netratings, as reported in a February article on Ars Technica, US search queries can be broken down as follows:

    Google: 53.7%
    Yahoo: 22.7%
    MSN:10.6%

    So Google’s dominance is not as complete as popular perception would have it, but they obviously still have the lion’s share of the search market.

    @Sean: Couldn’t agree more, unique meta descriptions are very important. Apparently if a site has multiple pages with the same meta description, search engines will not index all pages on the site, deeming that the pages contain duplicated content.

  8. Waatti says:

    I think putting the same meta keywords to every page is downright spamming and therefore should be penalized. But it is also true that you can’t be sure that everyone is aware of this and might get penalized without knowing why. I guess these kinds of ‘forbidden optimization methods’ should be made clear by different search engines – and also a some sort of a common agreement on those punishable methods would be great.

  9. Aaron says:

    What about search engine text on the page? I assume that Google doesn’t look at WHERE a page’s content is, but simply that it’s there. But I have a problem with companies that include a paragraph or two of text that is designed only for the search engines. It’s relavent, but not really part of the site’s content in any way. Do search engines look at the placement of content? (As in, is it a heading or a paragraph or a div?) Where I work the sites all contain such a paragraph in the footer but to me this is cheating. What are your thoughts?

  10. Jonathan says:

    @Aaron: I would like to think Google is smart enough to notice when keywords are bunched together in high concentration, such as in a paragraph at the bottom of a page. I’m not an SEO expert, but my understanding is that Google pays attention to the placement of text of the page, and that weight is assigned to keywords that appear near the top of the page, and also in the last paragraph (ie: the introduction and summary of a document). Also, text that appears in title and heading tags are assigned greater importance. Keyword density is also significant, and if the trophy keywords/phrases appear too many times on a page Google will assume you are trying to cheat the system.

  11. Zach Katkin says:

    Personally Google drives all of my traffic. I rarely see Yahoo or MSN in my stats. Personally I doubt Google or any other search engines will ever de-value, or rank a site lower because of its meta keyword info. Many web designers are guilty of doing a global include for the meta keywords tags, because it is very time consuming to include exact, page appropriate keyterms for each page.

  12. Jonathan says:

    @Zach Katkin: It would be nice to have proof one way or the other whether there is truth to the speculation that Google penalizes sites with keywords that don’t appear to match the on-page content. But even without proof, I don’t see any value in doing a global meta keyword include for every page on the site. It certainly won’t help your search engine ranking to have identical keywords for every page on the site, so why not just leave them out altogether and save a little energy?

  13. I have just started using a meta data tool that allows me to use separate keywords and meta data for WordPress (read more hear)

    @Jonathan, thanks for those stats but on my stats that I receive I rarely get yahoo, whether thats my bad coding or they are just not searching for me :(

  14. Jonathan says:

    thanks for those stats but on my stats that I receive I rarely get yahoo, whether thats my bad coding or they are just not searching for me

    Yeah I dunno – maybe a regional variation? We’re both based in Australia which might make a difference. I checked one of my client’s Australian based sites, and the referring search engine stats were skewed massively in Google’s favour. Google combine with Google images accounted for 90% of search engine referrals, MSN 4.5% and Yahoo a paltry 1.8%.

    On the other hand, most of the research I’ve done indicates that Google’s market share is much lower in the US, ranging from 45% – 70% depending who you ask.

    If anyone has SE referral stats for a site whose audience if predominantly from the US, I’d be curious to hear if there is a significant variation from the referral traffic Jermayn and myself are seeing?

  15. Interesting point about Yahoo being more popular in USA than Australia.

    SOT – In several movies (National Treasure for one), when they do searches I remembered being surprised that they were all using yahoo instead, so maybe your right about yahoo being more popular in other countries like USA.

  16. Luke says:

    I’m not sure I agree entirely.
    I’m no SEO expert but I’ve found in my endeavor to get our site ranking highly on Google for the phrase “web design Melbourne”, keywords have played a significant roll. In a competitive field like web design, I think Google considers every aspect. In less competitive markets, I would agree that keywords aren’t worth squat. We also use a global include for keywords too.

  17. I think meta words are important but not as important as they used to be (when people put bogas words in). I agree that its important to use them as it can increase your SEO

  18. Johann says:

    About time I found your blog because I just threw away meta keywords on my site, too.

  19. Dan says:

    Nice article, found it by chance after a friend of a customer complained that I left out meta keywords from his website.

    As has been mentioned some of the search engines don’t give meta keyword the importance they once did because of spamming by webmasters.

    I have not, for a while used meta keywords in any of the websites I have developed over the last year or so and have not noticed any downturn in results. I like to add unique meta description tags as well as page title and heading tags with keyword rich text. This seems to get me the best results.

    I do how ever use a meta tag to for WordPress that adds a meta description based on the first paragraph of your post and adds your blog’s categories as keywords for the “keywords” meta tag. This seems to be working also.