Not a great title, I know, but accurate.
Recently, I began a test of the ALT and Title elements in the img html tag. Here is the breakdown of the testing:
- Conducting the research by creating a page on the topic of “dogs” that has the same image of a Dalmatian that link to four separate pages. The parent page is only linked to from one page, the other four are linked to from the parent and the same location as the dog page. These pages will be further influenced by placing them into a deeper directory /dogs/.
- These four links will have the following elements:
- Page One – Link – No Alt – No Title
- Page Two Link – Alt – No Title
- Page Three Link – No Alt – Title
- Page Four Link – Alt – Title
- These pages will have unrelated but different names.
- The parent of these pages will be located on a rarely visited page (my html sitemap) by users and crawled regularly by search engine spiders.
- These pages will contain an H1, H2, H3 and page content. Lorem Ipsen will be used to incorporate the content with keywords added to the content in a consistent manner (same density, character count and placement of keywords).
- These page will target the keyword “Dalmatian.”
- When cached by all major engines*, we will conduct a site search for the priority of pages in the SERPs across all engines for “Dalmatian.” This will measure the impact of alt and title elements by seeing which of the pages ranks higher with no external impacts.
* Two notes to this research:
one > the pages within this study were originally published December 11, 2008 and adjustments were made to these pages the following day (due to my lack of spelling prowess) and the original pages were quickly cached by all three engines; and
two > despite returning cached dates of 1/2/2009 and 1/3/2009 MSN/Live Search’s index has failed to pick up the spelling changes that have been made, from “dalmation” to the correct spelling of the canine breed “dalmatian,” except for one of these pages that was reportedly cached on 1/25/2009.
These two notes being taken into account, the results for MSN/Live Search have not been included due to incomplete data. Ask.com was considered for this study, however, it has not, at the time of publication of this study cached any of the tested pages.
This being said, here are the results:
| Position | Yahoo! | |
| 1 | 4 – Alt + Title | 3 – |
| 2 | 2 – Alt + |
2 – Alt + |
| 3 | 3 – |
1 – |
| 4 | 1 – |
4 – Alt + Title |
| 5 | dogs | none |
click on the engine name to see the SERPs for the tested query
My initial assumption was that the title element has little, if not no value, however, based upon the results of this study, it appears that this is not the case.
Looking at the results from Google, you can see that the alt element is clearly more valuable than the title element (noting that both pages with alt elements outrank the one with title only). It is also curious that the page with both alt and title elements is the highest ranking page of the four. I would have assumed that this page would have been considered over optimized, but the results differ. The fifth result is interesting, seeing that it only has the target keyword visible once and in the image elements.
Yahoo!
Looking at the result it appears that Yahoo! does not see that utilizing both of these elements are ideal (my assumption is that it appears to be a case of over optimization based upon the fact that the page without either alt or title elements outranks the page with both), however, they put more weight on the title over the alt element. Not having a fifth result was interesting as well, seeing that the targeted keyword was contained within the visible text.
Seeing the results of this study, it is clear to see that Google and Yahoo! assign different values for elements within an image tag, however, it is interesting to see the differences, an apparent over optimization ding from Yahoo! and an increase in ranking from Google. So, with the results in, it looks like if you are optimizing your images for improved performance, the only thing that you can do to positively impact your ranking across both engines would be to utilize only the alt element, however, if you are looking to improve your ranking in Google, you might utilize both alt and title elements to give your page the best chance for ranking better.
I have tried to be as patient as possible with MSN/Live Search to see the impacts, however, I was more excited to release the results of this test than patient. I would absolutely love to hear if you have conducted similar testing and have differing results. I also know that these results will undoubtedly be different after the publication of this post, seeing that additional links will be pointing to these pages and this may/may not skew the ongoing results.
Let me know what you think of my study.
Related Posts >>
- Top 10 Posts of 2009
- Measuring the Value of Page Elements for Search Engine Rankings
- Further Study on the Value of Image Optimization
- Move Completed – Welcome to SEOpittfall 2.0
- Blogger Offers New Widgets
{ 2 trackbacks }
{ 21 comments… read them below or add one }
Good information!! this is exactly what I were looking for
Wow, not was I was expecting to see at all. I would have thought that the alt tag had more weight than the title tag did. I hope you will update the post when you get the msn/live results.
Thanks for the comments. This is certainly not the results that I initially assumed I would get either. I am not sure the MSN/Live Search results will be reported on because the study has been opened, but if the Google and Yahoo results remain consistent, then I might be able to report the MSN findings later.
Very interesting and revealing research, Stephen. It’s not surprising, though, that Google and Yahoo! put a heavy weighting on the use of title attributes on images since the use of image title tags is a W3c standard – http://www.w3.org/TR/html401/struct/global.html#h-7.4.3 – while alt tags are meant for accessibility purposes, not as tool tips (which is why Mozilla browsers like Firefoox don’t display alt tags as tool tips).
I think Google’s parameters make sense — rank an image higher if it has both title and alt tags as this is good universal design.
Steve,
Great work here. My suspicions have been confirmed! I have used ALT and TITLE attributes with keywords thinking they provided a slight value to search engines. The Yahoo! information is interesting. It is known in the xHTML world that ALT descriptions help validate the page and Google generally values sites with W3C xHTML validation. The TITLE attribute is known to have value in describing the anchor link resource and can have some benefit when internal linking content.
Yeah , it is great article, I use these in my own projects too… Thanks lol
I’m seeing Live/MSN results now.
(1) 3 – Alt + Title
(7) 4 – Alt + Title
(8) 1 – Alt + Title
(9) 2 – Alt + Title
Thanks everyone for your comments.
@Craig
The results in MSN still do not include the updated title that has corrected spelling, rather they still contain in “Dalmation” in the SERPs and the cached page. Hopefully, this will be updated… when is the better question.
Very interesting results. To think that google and yahoo are at odds over the use of both an alt and a title attributes. Looks like adding both attributes are the way to go. Google just brings in way more traffic than the other guys.
That’s a really nice results. Until know I always use alt for my images but I never did research on what the impacts is for search engine. Thanks.
I am always amazed at how many people do not add alt tags to their images. It takes two seconds to do and obviously there is a pay off. Thanks for the blog post.
Great article Steve! Looking forward to more updates!
We have always used alt tags as they are important to users using text readers – if you have images and want the site accessible to users with text readers they are important but handy for seo as well.
I wonder how and when Google consider indexing images on my pages. I’ve put both title and alt tags but haven’t seen my images on image search yet. My site is 4 months old.
Google has developed a tiny mini little bug right now. I was thrilled to see that has worked. For having your video related to google for a keyword research all you have to do is to make a good youtube description and to add a screen tag on your video. Talking about photo, i just changed the name of the jpeg file with a name that has the keyword. The jpeg must be placed on a page that google is considered important for one of the words from the keywords. And don’t forget to alt it.
I have two of the picture on the majority of the searches in google roumania related to baptist photo and wedding photo. And i can see visitators on analytic every day from those google image searches.
Great experiment! I had often wondered how much weight each of these elements carried and this clears up a lot of the mystery. Yahoo’s algo always seems wacked to me. Google seems a lot more logical and for the life of me i can’t figure out what Yahoo is thinking with some of their results. This just proves it again. Why would no alt plus no title rank higher than alt plus title in Yahoo?
I like to see the effect or Google caffeine in this test.
Another thing, Bing already improved its algorithm and it is now giving high rank to important keywords the same with Google. Do you have any update for the Bing?
Interesting experiment…
I like to see the effect or Google caffeine in this test.
Another thing, Bing already improved its algorithm and it is now giving high rank to important keywords the same with Google.
It looks like the results were for a standard web search. Do you think that an image search would follow the same formula? Thanks.
Kurt,
Thanks for the comment. I don’t have very much experience in optimizing for image search because I haven’t needed to optimize for images except for regular SERPs. I would love to see findings for this if you are up to testing it.
Cheers!