Meta Descriptions Now Count for SEO says Google?

Alex Bennert (SEOSylph) tweeted a link this evening to an article by Sean Michael Kerner at Small Business Computing from today covering the SES Toronto keynote by Maile Ohye, senior developer programs engineer at Google.  According to the Article, Ohye said that Meta Descriptions are now once again factored into the relevance of a page.  That’s a huge change if it’s true…

According to Sean:

On the topic of metadata, Ohye said that Google is now once again pulling its search index descriptions, or snippets, from website metadata description information. Ohye noted that for several years, Google considered metatags as a spam technique, but are now again considering meta description information when determining overall search relevancy.

I’ve read and re-read that paragraph four times now, trying to determine if I’ve misinterpreted it.  I can’t see a flaw in what I thought I read the first time.  “now again considering meta description information when determining overall search relevancy” seems pretty straight forward to me.

I’ve tweeted a confirmation request to Matt Cutts, and we’ll see if he responds at all one way or another.

Update: Vanessa Fox Weighs In

In response to Alex’s tweet, Vanessa Fox, herself often misquoted, chimed in with

my guess is that was almost-right reporting. And that she actually said keywords tag not used, but description used for display.

She then went on to tell me that

There’s no quote in the story though. just a summation. this happens all the time when I speak

Well, if that’s the case, that’s a significant mis-interpretation on Sean’s part.  Which might explain why SEO types who attended that keynote haven’t lit up the blogosphere with the “news”.  :-)

As I’ve been digging around the interwebs for more info, Christian Maund-Anderson said he got the impression it was business as usual – that they may, or may not, use it, as overheard while at the conference.  Since he said he got that impression from “overhearing” it at the conference, I’m not sure how that figures into the equation.  When I asked for clarification, and if it was a google employee who said that, Christian then said

Yeah, they gave pacman aka googlebot a sugar pill, they are definitely ramping up on information retrieval.

LOL uh, okay then!

Testing Testing, 123, Is This Thing On?

What I want to do is some testing.  It’s just that this could be tricky, both because proper testing of this could take some serious time, and I’m also intuitively thinking it’s a minor factor, so any testing done needs to be done quite carefully and across multiple scenarios. So good luck finding the time for that any time in the very near future!

Call For Opinions, Tests, Data

What I can say is this – if you’ve been down this road recently and have conducted any actual tests, I’d love to see the data.  Or hear your thoughts. And as I learn more I’ll share it here…

UPDATE June 12th – Testing Begun

I have actually begun a Meta Description Usage Test for this scenario, and since Vanessa and Jill have weighed in (see comments below) I’ve redoubled my efforts.  But I don’t want to be isolated here.  So if you would like to help here’s what I’m doing.

I created a new page on my old business’s site just for this test.  I have a meta Description with text that’s unique to that page and not in the page’s content.  I’ve then linked to that page from both that site’s home page (the only page remaining on that site these days) and from my blog.  I then added it to my sitemap.xml file, and resubmitted that to Google to try and expedite it’s inclusion in the SERPs.  From there, I’ll see if that exact match wording will cause that page to show up in the SERPs.

That’s the first step in the testing process.  Up for it?  Let me know!

Also, if you know of any real world test conducted in the past few months, that info will help…

Update #2

Jill Whalen commented that new pages may not be treated the same as old ones, so I will now add an existing page into the test mix to see what happens with that and if there’s a difference between a brand new page and an existing page.

What To Do About Meta Descriptions in the Mean Time

Regardless of whether it’s true or not, for quite a while I’ve been advocating to clients that it’s my belief the best practice for Meta Description usage is to write content that engages the reader in helping them understand what the page is in a way that attracts them to visit your link as compared to competitors.  And in that process, you should integrate the top phrase of the page.

The thinking here and as I’ve seen others do and advocate, is that if someone types a phrase to search, the words in that phrase will be bolded in the SERPs.   And if those words are highlighted in the Title and the Description, it’s more likely that the person doing the search will say “this one really is on target for my search”…

And in my A/B testing over the years on several client sites, that’s been the case.  It’s critical though, that the description not be keyword stuffed, and really does make sense as a readable chunk of text…

About Alan Bleiweiss

Just another guy. Who happens to have a lot of experience living, breathing and sleeping organic SEO. So that's my primary focus - high end SEO audits and consulting for sites ranging from thousands to tens of millions of pages. In my spare time I blog, rant, write eBooks, and speak at industry conferences.

Read more from

61 Comments

  1. Stuart says:

    Ok it looks like I’m going to be the mug who goes first here.

    My experience over the years is that Google has never stopped reading the meta description tag. I have not seen a single listing for any site that we have built for a client … and there have been a lot … where Google did not take our meta description and display it in the search results.

    Is it any wonder that we rarely listen to anything anyone says about what Google is doing because … more often than not … we’re not seeing what the gurus are waffling about repeated in any of our sites.

    Stuart

  2. That’s about right Alan, on my pacman reference: Essentially Google is throwing more bandwidth on the collection and storage of more meta data information, whether or not it is an actual factor in their algo cannot be confirmed. I really did get the impression that this isn’t a huge impact play for overall site optimization, I wouldn’t give it any weight. To be accurate on the “OH”: http://twitter.com/cmaundanderson/status/15973360155

    Hope you don’t mind my “quirky” references to pop culture.

    Cheers,

    CMA
    Christian Maund-Anderson´s last blog ..Buzzkill: My Top 10 List of Online Marketing Overkill in 2010 My ComLuv Profile

  3. Thanks Guys – at this point, I’m adding the ? to the article title, and we’ll take it as it comes. Without testing, who knows whether it’s true. I do know that I hate when Google uses their own description snippet, which usually turns out to be crappy, garbled, and totally meaningless to the visitor doing the search. So I’ve personally always optimized Titles anyhow…

    • Thos003 says:

      Alan, beyond the fact that I believe meta description does play a role inadvertently in SEO and ranking, I more hardheartedly agree that your description is the hook on your fishing line. And that when included and relevant to the search query you will stand a far better chance of getting good traffic with a catchy, yet accurate, description.

      I believe meta description is a best practice.

      The only decent argument is that if you don’t include a description that the search term will show up in bold under your listing.

      ….But, out of pure laziness at times, I don’t take time to add descriptions to all of my blog posts.
      Thos003´s last blog ..Why use baits instead of spraysMy ComLuv Profile

  4. [...] This post was mentioned on Twitter by Arnout Hellemans and Sean I. Mitchell, Anthony Verre. Anthony Verre said: Meta Descriptions Now Count for SEO says Google? | Search Marketing Wisdom http://bit.ly/aPihl9 by @AlanBleiweiss [...]

  5. Vanessa Fox says:

    Stuart,

    Google has never stopped using the meta description tag. How things worked when I was at Google and as far as I know the way they continue to work now is:

    -Google doesn’t use the meta keywords tag at all.
    -Google doesn’t use the meta description tag for ranking/relevancy
    -Google *does* use the meta description tag for the snippet in the SERP

    Note that Google doesn’t *always* use the text from the meta description to display in search results. Google Webmaster Tools has a report of pages with meta descriptions they can’t use (too short, missing, etc.). In addition, if other text on the page is more relevant to the query (for instance, it’s a long tail query and the keywords aren’t in the meta description but they are elsewhere on the page), Google may use that other page text for the snippet instead.
    Vanessa Fox´s last blog ..May 26th, 2010 Office Hours My ComLuv Profile

    • Vanessa,

      My initial tests since this question arose yesterday show they’re still doing it the same way. If they’re using the Meta Description not only as the text to display in the SERPs, but as part of the page relevance evaluation, I haven’t been able to isolate a way to even test that yet.

      I don’t see WHY they’d start using it that way either, yet it is possible. And anyone who’s been creating quality descriptions shouldn’t, in my opinion, have any reason to change their ways, even if it’s true.

      But thanks to your chiming in last night, I’m much more skeptical that they did change their use of it. So until I hear otherwise, this has now gone down into the “people who mis-quote others” category.

    • radiodave says:

      Vanessa; I’m no expert but what u say makes a lot of sense. (Not that Google is expected to be so logical.) You get the distinction between the meta description appearing in SERP and same description weighting the search score. If I were Google, I’d add points if meta description matched text on page and subtract (or even ban) if no match.

  6. Jill Whalen says:

    Vanessa said:

    -Google doesn’t use the meta description tag for ranking/relevancy

    @Vanessa, then why do pages with a particular word just in the meta description show up in the Google search results?

  7. Jill Whalen says:

    Yeah have seen it for years (But it has been years since I’ve checked again). I’m not sure if I still have a test page up as I got rid of all my old blogs where I had them. But it’s always been that way. In fact, what’s even cooler is Google will index the words even if they’re like the 250th WORD (not character) in the meta description.

    Part of when Google says “these words only appear in pages linking to this page” is often actually just that the words are in the meta description, not in links to the page.

    I’ll see if I still have a test of this. Unfortunately, I rely on memory and forget to write down the test pages and words! I’ll run a new one too.

  8. Jill Whalen says:

    I should add that I think it’s possible that it’s only counted (the page shows up) if there are no other results elsewhere. So in the real world it’s not all that useful, which is why I don’t bother with Meta descriptions all that much if the page has great keyword rich copywriting anyway.

    • Yeah if they do use it that way at all, it’s got to be very insignificant overall in terms of weight for any truly optimized pages…

      Maybe I shouldn’t have even written this article since it would probably just get that many more half-clueless people trying to mess with the description field, thus further causing their clients grief rather than real results :-)

      On the other hand, I get to use testing methods I picked up from SMX this week, expanding my text methodology tool set :-)

    • Indeed, I’ve seen this situation on tests (from several years ago) with the alt=”” attribute and accents in French. In order for a page to show up in Google SERPS for an accented or non-accented version of a word (whichever is not visible on the page), placing it in the alt=”” attribute would work. However, that did not mean the page would be competitive for that version of the word. (CAVEAT: This is several years ago and the alt=”” attribute might be treated much differently now.) The bottom line is that just being indexed (relevancy) is not the same thing as being a ranking factor. And because the meta description is used for the SERP snippet, that is not the best place to put a misspelling or non-accented version of a word.
      David Leonhardt´s last blog ..Pros and cons of country-specific domains My ComLuv Profile

      • David L – This specific testing is being restricted to the Meta Description so anything to do with alternate attributes would be for a different thread. As for the indexing vs. relevancy concept, that could open up a whole new can of worms, even when discussing Meta Description.

        What I’m focusing on for this test cycle is limited to 1) – can a site be found for words unique to a meta description? 2) if so, under what circumstances?

        From there, IF it can, then we can explore the “indexing” vs. “relevancy” vs. “ranking importance” held up against “a low number of high value phrases” vs. “a high number of lower value phrases”…

  9. Jill Whalen says:

    Sorry for 3 comments in a row. I was just doing more real world testing with pages in existence and it seems that for exact phrase searches are NOT bringing up the page in question with Google right now.

    A non-exact phrase search can bring it up though, but it’s hard to correlate that to it being the meta description or just the words elsewhere on the page in other order.

    My previous tests were all one nonsense type word. Which as I previously said, is not a great real world test.

    But I’ve always equated Google’s indexing of the Meta description tag the same as Yahoo’s with the meta keyword tag. While they know the words in it (which I do believe) they don’t count unless there are no other pages that Google can find that have better relevance.

    • LOL – Jill! You’re apologizing for adding your input to the dialogue on my blog? You’re kidding right?

      Since I’m a recovering addict, my goal is to get every well respected member of our community to leave comments all day long every day! :-) More Comments. Need. More. Comments! Must. Have… LOL

  10. Ryan says:

    I have yet to see a case where a page ranks for a phrase that only exists in a meta description.

    Perhpas the best test would be to make up a new page, link it from your homepage so it gets juice, then add some unique phrases to the meta description. Without using those phrases on the page or in links to them, get some other people to do the same. Then, see if you rank for those phrases.

    My guess: you won’t.
    Ryan´s last blog ..Who My ComLuv Profile

  11. safcblogger says:

    Alan, i will also give this a go on a weak domain with some gobbled nonsense word and get back to you. I suspect it won`t but lets try.
    safcblogger´s last blog ..Google Parameters My ComLuv Profile

  12. Jill Whalen says:

    Why make up a new page?

    • Ryan Jones says:

      So that you can isolate as many factors as possible. With an existing page, you have no idea what type of terms are used in anchor text pointing to them. Also, you don’t want to screw with the rankings of an already existing page – at least I don’t.

      In My case, I didn’t pick a nonsense word, but I picked an exact match phrase that had 0 results in google – just like the Testing panel advised at smx.
      Ryan Jones´s last blog ..Who My ComLuv Profile

  13. safcblogger says:

    Well the page has been indexed but nothing so far
    safcblogger´s last blog ..Google Parameters My ComLuv Profile

  14. Jill Whalen says:

    But new pages may not be treated the same as old ones. Especially if they’re just created for a test.

    It’s good to test those of course, but looking at pages in the real world (to me) are better indicators.

  15. safcblogger says:

    I regret to inform that after 2 seperate tests. This is bullshit.

    A brand new page on a weak domain yielded no index of the term that i had in mind. In this case a BS nonsensical word on a weak domain.

    I know this goes against the grain of the recommended, but sureley an indicator would have been present. No Joy.

    Then after Jill suggested looking at the same scenario on an existing page, i tried a similar re-inclusion i tried via XML on a more authorative domain. Alas the same.

    I suspect as we all may that a “first inclusion” will come via anything other tha the meta desc. and in this instance i hold my hands up and say i have found nothing that will support that the meta desc will form an indicator. As we have all already achieved that staus quo.

    If i am proven otherwise i would love to see that test method.
    safcblogger´s last blog ..Google Parameters My ComLuv Profile

    • Thanks for contributing to the test Dean. So far I’m seeing similar results but have more to test because I want to take it further in terms of if/then and A/B, multi-directional testing. If you would like to write up a blog article on it I’ll gladly link to it when I publish my findings.

      Alan

  16. David says:

    @Alan I could be reading the experiment wrong but wouldn’t it be easier to find a page or a random site that already has meta keywords stuffed with key terms and then optimise it to be so called useful and monitor the page’s organic traffic and exposure in Google Webmaster Tools?

    Much easier and quicker to test what happens if you update a page that maybe benefiting from that meta description and take away or reduce that benefit.
    David´s last blog ..GM fails with Chevy My ComLuv Profile

    • David,

      The core experiment is to determine whether having keywords only in the Meta Description field will result in Google showing that page for that keyword phrase. Beyond that, there are other issues that could arise. We aren’t however, looking to see whether a Meta Descriptions presence has any impact on click-through or conversions.

      • David says:

        @Alan,

        Yes but i understand that i was thinking it would be more transparent to use an existing site/page that ranks very well for a term that is not showing up for a similar keyword which is not featured on page or in description but once you change the meta it now ranks #… for that term.

        the test would show you the results faster….
        David´s last blog ..GM fails with Chevy My ComLuv Profile

        • I’m doing a test along those lines as well as a “new” page test. It turned out that the using a brand new page got fast results – indexed in less than 2 hours. It’s the “change an existing page” then “wait for the refresh of the Google index” process that takes longer. At least for the pages I’m using. If you think you have a way to do it faster, by all means – please test and report back. The more input we have, the closer to a consensus we might get, and the more confidence we may achieve.

    • Jill Whalen says:

      What does the meta keyword tag being stuffed have to do with tests for the meta description?

      • David says:

        @Jill i was just thinking that some clients that i just monitor that haven’t done any onsite for several years appear to be showing decent gains.

        They all have meta keyword stuffing in common… i was just thinking if this test proves to be correct it might lead back to the days of spamming via meta tags…
        David´s last blog ..GM fails with Chevy My ComLuv Profile

        • Jill Whalen says:

          The Meta Keyword tag is a completely different animal, and imo, shouldn’t even be introduced into this discussion as people already confuse the phrase “meta tags” as meaning the meta keyword tag.

          Google has never indexed the Meta keyword tag any time I’ve ever checked and I doubt they ever will.

          The Meta description tag is a different story as I know for 100% that they have in the past indexed words in it previously and ranked pages based on them if there were no other pages with those words on them.

  17. So I was in that session at SES Toronto, and that’s not exactly how I remember it going :)

    Maile said that the meta keyword tag was not used because it was spammed so much, and then she said something along the lines of the meta description tag is not spammed so they do take it as a signal. That is not an exact quote, of course. I think I was only half-listening at that point. She certainly didn’t say it in any way that made me perk up and go “oh? that’s new”.

    I didn’t get the impression from what she was saying that it is a new factor in the ranking algo, more like “business as usual”. She was talking about meta data in general (incl. microformats) and that they were using those more heavily now.

    I know, that totally didn’t clear anything up, did it :)
    Dawn Wentzell´s last blog ..Creating Custom Blog Pages in WordPress My ComLuv Profile

    • Dawn,

      Thanks for providing your take on it. So it appears that this was all a case of mis-understanding, yet here YOU are, saying

      and then she said something along the lines of the meta description tag is not spammed so they do take it as a signal.

      And that further backs up Jill and Edwards perspective, especially given the “business as usual” aspect.

      This is going to once again prove to be both fun and frustrating as we start publishing more tests – people participating in the test have slowly been reporting back to me their findings and I’ll be publishing those once enough broad spectrum data is in :-)

      More important though – can I pay for you to attend aggressiveness training? I REALLY would love to see you take on TOPSEOs and Visibility magazine at one of the conferences you attend! :-)

  18. I’m excited that you are doing this test Alan. (For the record, I’d be taking part in the test if I wasn’t leaving the country for 2 weeks!) I have always used the meta description – not just for the ranking element, but for the marketing aspect. If I can control the snippet to include a strong call to action, a phone number, an offer to encourage click through, I am going to include it. I try to make it as relevant to the page as possible. I have never seen it impact the rank, but I’d like to think it affects the overall marketing of the site.
    Monica Wright´s last blog ..Validation Rocks My ComLuv Profile

    • Agreed Monica – you’re flaking on the community that brought you this far, and we understand – it’s business as…

      oh.

      wait.

      uh…

      Ah – yes – you’ve got a life and you’re living it and it doesn’t ALWAYS revolve around Search Marketing… Of course! (I’m just jealous that you get to go back home for two weeks! )

      Seriously though – I completely agree with the concept of using the Meta Description as a marketing touch point. I hadn’t considered putting the phone number in it though – my preference is to get them to click through. At least that’s the goal until we have proper call tracking in on a client’s marketing initiative…

  19. I’m in for the test! It’s a good opportunity, no matter what was said, for us to have some data. I’ve been working under the same premise as Monica and based off Vanessa’s breakdown of how Google interprets the meta description tag.

    Dawn may have cleared some things up. This may be a situation of “telephone” and misquoting/interpretation changes the message.

    I’ll begin testing later today, Alan. I’ll do the following:

    * Use a unique phrase that has 0 search results for exact match, “unique phrase in quotes”
    * Put phrase in meta description on an new page, no use in body copy
    * Send links to new page.
    * Put phrase in meta description on an old page, that doesn’t have phrase in incoming links but is a page that has some juice already. (My previous business site, PR 4, is live still. Page ranking isn’t a big deal for me on this site, so I can play.)
    Dana Lookadoo´s last blog ..Formula for Reaching the Top – Pain is Growth My ComLuv Profile

  20. Alan et all:
    Thanks for this interesting discussion. I’ll keep using meta description tags since there seems to be no negatives. Looking forward to the results of all the testing…

  21. [...] week after SMX, Alan Bleiweiss started running a Meta Description Test to see whether or not the Meta description has any affect on search engine rankings in [...]

  22. [...] seem to use a slightly different method of determining things to some of my seo buddies with tests here and [...]

  23. a. Meta Kw tag – no longer works since Bing signed deal with yahoo..
    (no discussion needed.. testing has been done both ways, is NOT used now. – FACT)
    b. Meta description tag – yes.. shows in serps as page snippet
    1. downstream indexing is where SERPs relevance comes from.
    2. test needs to run AT LEAST 6 weeks (unless you are splitting spiders)
    3. duplicate descriptions penalize and invalidate.
    4. the rest will be on twitter stream
    #whitecoatseo´s last blog ..whitecoatseo: RT @theGypsy: @whitecoatseo lol… nice…. smart ass My ComLuv Profile

  24. [...] mettere una pietra sopra gli ultimi dubbi ci ha pensato Vanessa Fox, che nei commenti di un post su un blog (che titolava proprio “Meta Descriptions Now Count [...]

  25. [...] week, I posted an article in response to a claim I read where it was reported that Google is now using Meta Descriptions in page relevance.  We got a flood of input, both on Twitter and in the comments of that article.  Vanessa Fox [...]

  26. Jay says:

    Alan,

    Interesting test regarding the use of meta descriptions… It sounds like a meta description alone for a specific keyword phrase is not sufficient to find a page through search unless the page itself contains those relevant keywords?

    If I’m understanding that correctly, is it possible that the reason for this is to prevent spammers from abusing meta descriptions in the same way that meta keyword tags were formerly abused?
    Jay´s last undefined ..Response cached until Fri 25 @ 16:55 GMT (Refreshes in 23.93 Hours)My ComLuv Profile

    • Jay,

      I have to assume that’s the reason – Google doesn’t need to encourage any more spam, yet the Meta Description “can” be an ideal place to provide the descriptive text for search results pages. So unlike the Meta Keywords field, they don’t want to abandon it altogether. Especially since their ability to select content on the page itself for use in the SERP snippet is quite often very messy :-)

  27. [...] knows it.” How many times have you said something like that? Read through the comments on Alan’s META description test and count the passionate opinions there that are soley based on ego without any data to support [...]

  28. Scott says:

    Hi Guys and Girls,

    So what happened with all the testing? Where are the results? Is there more about it in another blog post somewhere as i couldn’t find any on the link provided near the top of this page?

    I’m very interested to here the results although i assume you found what Vanessa said and what Google states in Webmaster tools is the truth??

    I have a site with really good rankings that i want to change the Meta description for but have been too scared too…..I still want to see the results first ;) but i believe Google…i think ;)

    Yea i Do :), still be interested to see what turned up.

    Thanks,

    Scott

    • Scott,

      all the testing that was completed (by several people) got the same results – none of us was able to get a site ranking based solely on meta Description content. Feel free to work on changing yours as you see fit, with the understanding that the whole point / value of that text is to reinforce to people doing a search that “this is a site I want to click through to because it really is relevant to my search”…

      • Scott says:

        Hi Alan,

        Thank you for the very quick reply!

        That is an awesome bit of info to know as many people, myself and a lot seo experts included believe it does and write tags more so for rankings than necessarily getting clicks, or at the very least, certainly don’t write the exact tags we’d like to.

        Thank you very much
        Scott´s last blog ..Aboriginal Art Gallery PreviewMy ComLuv Profile

  29. SEO Matt says:

    I am just reading this now (in 2012), and am curious about how much of this is SEO wisdom is still relevant given the constant changes to Google’s algorithm?

  30. David says:

    How does google, searches engines feel when you add a phone number in the meta description.

    “Call 817-797-9047. Ask about a home you like or to sell your home. We do Video of your home! Let us show your home the best way!

    If I had 3,000 pages and they all ad this. What effect would it have?

    • This would be acceptable but only if the description matches Google’s evaluation of the page content. If not, Google will replace your description with a snippet of content from the page in question. So unless you can figure a way to get 3000 pages to all have truly unique focus and content, and ensure that description makes sense for every one, you’ll likely see the overwhelming majority of your descriptions replaced.

Leave a Reply

CommentLuv Enabled