Replacing H1 tags with Images
Ok while hopefully helping a new client out, one of the things I picked up on was the way they where replacing the H1 tags with images. I thought personally it was low level cloaking or hidden text, the inhouse team where pretty adamant that a certain Matt Cutts had said it was cool … this prompted me to say ..
“I don’t think it will get you a ban, but it could get you a black mark.. too many black marks will get you a ban, and for fucks sake we could have worked the H1 into the layout”, but in the back of my mind the fact that they had pulled the Matt Cutts line ( and yes I swore at the client, but I’m passionate about winning)
11 hours on a plane… time to think
ok I’m Cool with the -1000 indent in css IF ( <- big if ) the Logo example :
H1 = “DaveN logo” or even h1 = “DaveN a UK seo”
where it’s not cool and you will get shouted at if caught by your SEO is …
H1 = “Search Engine Optimisation, Good SEO, Search Engine Optimisation in Leeds, Search Engine Optimisation in yorkshire”
H1 = “Search Engine Optimisation, search engine optimisation, Search Engine Optimization, search engine optimization” <- thats way way out !!
so I suggest that you get a text reader, close your eyes and listen to your website.
DaveN
Sam I Am 1974 days ago
http://samiwas.stumbleupon.comThat’s exactly what Matt Cutts said. If you’re just replacing the text with an image that looks better, then no worries (although he did also talk about the extra notch being added to your site and you certainly don’t want too many of them!). There’s at least one thread in google groups where I asked this and it was confirmed back by MC via one of the other GWG Googlers. If you’re adding other stuff then you’ve got a potential issue on your hands. Some relatively known SEO was outed for doing this -actually just adding one or two words even so it was really nothing- a while ago actually and it was one of those storms of blog posts things (only in SEO land of course
). Can’t remember who it was though…. anyone?
Search Engine Optimization - Terry Reeves 1974 days ago
Is there ever a relevant reason to use a -1000 anything?
Hobo 1974 days ago
http://www.hobo-web.co.uk/So you think it is ok to put ‘seo’ in the logo? I’ve often wondered, and it certainly doesn’t seem to be doing you any harm (and I suppose it’s reflected in your main title tag on the page you are pointing to – so 100% “relevant” or “descriptive”) but what is the difference if your actual title was “Search Engine Optimisation, Good SEO, Search Engine Optimisation in Leeds, Search Engine Optimisation in yorkshire” surely that would be equally relevant, if a tad unfair on screen reader visitors, not to mention crap title crafting.
Jeremy Luebke 1974 days ago
I have this problem in my company. They did just this. Yahoo and Google let it slide across the board but we have a penalty in MSN because of this and I can’t get it lifted even though the issue has been fixed.
Paul-S 1974 days ago
http://www.northsouthmedia.co.ukIt is food for thought, what is the difference, unless your primarily concerned with peeps with screen readers.
Does having only a few words hidden matter less than a keyword stuffed spiel.
Do the bots actually surf your CSS file to see if there is any hidden text?
Or is it just down to the inform brigade and hand jobs?
Good question Dave, good response Shaun.
Chris 1973 days ago
I’ve been doings this across my blogs for a while now, nothing spammy mind just the keyword blog title in the H1′s. Never encountered any problems *touch wood*.
Joost de Valk 1973 days ago
My line has always been: if you OCR the page, there should be the same text as when you read the source.
peteT 1973 days ago
It is a pretty well established design technique amongst web developers/designers who have accessibility in mind. As noted the intention is to have a text version available to assistive technology like a screen reader (or any other device that does not understand CSS) but allowing designers more creative freedom to use an image in their design. And yes the aim is to have the text reflect what is shown in the image.
It would be a shame if this technique becomes a NO NO because the search engines take an exception to it because some people have been stuffing keywords into the text.
Joff 1973 days ago
Anyone stuffing keywords has to know by now that at some point it’s going to be frowned upon. Simple descriptive replacement is fine and at the point that Google blackmarks your site because of it, it’s time for Google (or any other search engine) to be relegated to the pile of “search engines we used to use”.
What happens with the case of sIFR then where the whole point of the technology is to replace the text?
Just to clarify, are paragraphs of white text on a white background still bad?….
b10g 1973 days ago
http://www.b10g.co.uk/blogThats exactly how I have always understod this matter, as long as it is relevant to text in an image it cannot be seen as malicious.
Matt Cutts 1973 days ago
http://www.mattcutts.com/blog/I knew if I waited, someone else would come in and say it for me. Just what Sam I Am and Joost said.
DaveN 1973 days ago
so the 10 million dollar question …..
is my H1 “DaveN a uk seo” for the image “can’t beat a good rank” lost my PR caused me pain or is it close enough to the title and theme of the strapline LOGO .. ?
DaveN
Lea de Groot 1972 days ago
Dave, in theory that could get you a hand penalty (I’m NOT going to say the obvious pun!)
The alt/ replacement text for an image should match the words in the image.
As someone said – if the OCR doesn’t match the screen reader result, you may have a problem.
Lea
~ not a Google employee, nor playing one on TV…
inkode 1971 days ago
http://www.inkode.co.nz/inkodemissionstatement.htm“so I suggest that you get a text reader, close your eyes and listen to your website.”
Seriously that is a really great tip, actually hearing how your copy reads.
Brent D. Payne 1971 days ago
http://www.brentdavidpayne.comWhen I was working for OneCall.com as their in-house SEO, we ran across a problem similiar to this. However, our problem was more about page load times. The ‘Essentials for the . . .’ section of the product page was taking too long to load (db logic processing lag). The solution the web team put into place was an Ajax implementation. However, that wiped out one of our H2 tags. We decided to H2 the keyphrase off the page and I asked Matt Cutts about this at SES San Jose 2007 (shortly after I denied him access to my wall outlet, before realizing who he was). I explained that the text is a replica of the Ajax’d text. He said it was fine.
Here is a sample page of this in action:
http://www.onecall.com/ProductDetails.aspx?id=87213
Bronson 1970 days ago
http://www.twenty20.co.zaThanks for the information Dave, I hope that flight doesn’t take too much out of you.
For me the general rule of thumb for keeping it clean would be to say that whatever you would get from OCR on the logo file should be what your H1. Then you are clearly coming at it from a usability and accessibility angle rather than keyword stuffing an H1 tag.
Most of the time a small background image file (for non-text elements) with a sprinkling of css-styled text elements can render almost identical results – but keep the file sizes down.
I agree wholeheartedly that a small black mark is acceptable, but the problem comes in when you start setting off other flags (stuffing H1′s, buying links, and various other questionable actions) you are running the risk of review – and when it’s review time, all the little black marks can add up to a big foot in your backside.
Byron 1970 days ago
http://www.byronrode.co.zaHi Dave,
I had always had suspicions and it was why I always made sure it was just dynamic text REPLACEMENT and not keyword stuffing. I did confirm this with you beforehand, but glad to see that the matter has been brought to light. I have made certain that all implementations of Image Replacement are in fact replacements and not a keyword stuffing exercise.
Just to clarify though, would it not be better to use display: none; than text-indent: -1000px;? It would be interesting to hear if there is a difference in the scoring of penalties in that regard.
DavidT 1970 days ago
Hi Dave,
The company I work for uses this technique using css and sIFR so when you disable css you see the text and not the image. Recently we have seen 3 or 4 sites drop of the msn rankings that use but we have not keyword stuffed the image just used replacement text relevant to the image or sIFR text. I posted a question seomoz but they where of the opinon that Msn is running an update at the minute which is true, but the image replacement seems more of a logical explanation for the sites dropping off. If you want I can post up a url of an example site????
Cheers David
Anthony 1969 days ago
Hmmm, what about other elements hidden for good reason – such as a h1 of ‘Global navigation’ preceding the nav? That’s helpful for indicating underlying structure to a screen reader but should not be displayed visually…
DavidT 1968 days ago
I sent an email to MSN and it seems alot of people are asking for re-inclusion, they are saying they have a quite a back log and that should be reincluded sometime soon (some hope). So it looks like Msn have tried an algorithm change and anyone using an image replacement code has been penalised much to there oh so diligent hard work looks like they haven’t taken into account that alot people use this technique, hence the backlog in re-inclusion requests.
I get anymore info I will let you know.
David
Betty 1967 days ago
http://www.articlesgarage.comGoogle doesn’t how to discover hidden words or phrases if CSS is used. I have seen so many sites where, if you choose to disallow CSS from the webdeveloper toolbar, you will discover so many things that weren’t there before.
Take care of the big white spaces on the footer-is a good sign someone is using text indenting to hide words. And after reporting so many sites to Google, I just gave up. No use whatsoever..
Cloaking on the BBC Home Page - SEO Skeptic 1535 days ago
[...] relates back to a post on Dave Naylor’s blog I saw a year or so ago where there was a discussion of using CSS replacement techniques for image-presented text. Matt Cutts responded there, giving his thumbs-up to a comment added by “My line has always been: [...]
CSS Image Replacement, what's up Matt? - Yoast 725 days ago
[...] a simple rule: if you OCR the page, it should be exactly the same as when you read the HTML.I've said so on Dave's blog back in January '08, and then, Matt Cutts agreed with me. Well, Dave pinged me [...]