This test is pretty straightforward and it checks if we are wasting our time worrying about image file names and changing them to match our keywords.
Test Downloads
Transcriptions Below
CLINT: Test 490 is: Does the image name affect rank? This comes from pretty much every one of their brothers including this thought. When you do the images the first thing you want to do is make sure the file name contains your keyword.
So, you have keywords in your file name and then your alt tag is also going to have the keyword in it. Now, with the focus on accessibility because lawyers in some states are going after people because it's a good cash cow for them, you want to have the title in there as well which you can also put your keyword into.
There's three things there that are supposedly good ~ optimization pieces. So, we wanted to test if – and Hunter came up with this one is, does this really matter or we are just wasting your time worrying about it. We've done a lot of stock image tests and stuff like that. We can use those and we can use doctored images where you take some Photoshop or whatever and flip it or add text to it and all that stuff to make that image a little bit more unique for your purposes, but are we wasting our time changing the file names too?
So, he put it up as the non-keywords is basically at the filename and then this one has the keyword in the URL. That's pretty much how they landed for the duration of the test. He's got ten test pages but these are the ones that showed in the results and it's pretty conclusive that we're wasting our time really.
Google could care less one way or the other if we put the keyword in there or if we don't in that filename part. I think this even plays a bit more of a part when you're talking about people who are doing EXIF data and not paying attention to the sites that are uploading it and saying, "Oh, I'm doing all this EXIF data. It's great." And then putting it on to the site XY and Z. Then, you do a download and you pull that image out and you see the site XY and z are actually stripping all the EXIF data or all of the jpeg head data.
A nice platform as a good expression of that is Google Sites. If you put an image on Google Sites like how we build these as we put it on Google Docs, Google Sites is the same way. If you add the URL Google Sites strips out all of that data. All the heads is gone - the EXIF is gone, all the jpg heads is gone. If you do the upload it's actually flipped around.
If you do the URL it keeps the EXIF data but strips everything else, and it only keeps the description of the EXIF data. It doesn't keep the geo tagging or any of that it just keeps the description. If you do an upload it rips everything. All the EXIF data is gone. The famous counter-argument is, "Well, just because they deleted it that doesn't mean they read it."
Well, I think that if they go through the process of erasing and optimizing the image and pulling out all the data they could care less about the EXIF data. So, when you're using the EXIF data and you're worried about keywords and keywords in the EXIF data and keywords in your filenames from the looks of this you're really wasting your time unless you're paying attention to the sites you're uploading it on.
This test pretty much confirms my hypothesis. Then, SIA has done EXIF test before as well. I think the issue with the EXIF tests back then – and dori you can correct me if I'm wrong, was some of the data that was called EXIFs was actually IPTC data which are two different things and are used by the machine in different ways.
Regardless, if you're using EXIF data on your website make sure you're not stripping it off. Then, two of the tests from SIA says if it's on your website it really doesn't matter because it just didn't help one way or the other versus putting it on other websites that keep that EXIFs data and it results in backlinks and recognition to your website that may help. So, keep that in mind when you're doing all this stuff.
DORI: That's interesting and the only thing I would think possibly if you have it - the keyword in the file name, is if you were trying to rank that image because then the keyword is in the URL which we know always helps.
CLINT: Yes, that is a good sign if you're doing some image SEO stuff. I haven't looked at it very closely because I don't really mess with the image SEO stuff. I'm just, if it happens it happens. If not, well... Because I have all of my images for the most part set to be the 301 to the attachment page.
If you don't know what that means you just get into Yoast then you can turn that off and you'll see ~ that show up and basically there's the URL for the image itself. Then, Yoast and a lot of the other SEO plugins will just 301 that URL too with a page where that's attached too. So, ~ a URL and it would actually 301 to the document instead.
DORI: I didn't know that Yoast did that.
CLINT: Yes, that was one of the bugs that Yoast had and it actually stopped working for a little bit. Then, people started crying because they had a whole bunch of random URLs indexed. What it was is that the attachment page feature went away so now it doubled their URL count based off of their image count. At the end of the day, you can't say that you were doing image SEO and then complaining about that because it was not necessarily a bad thing. It was just a bug.
Alexander had a question. How would that affect the mobile ranking?
DORI: I think he was talking about the earlier test because that came in right before we were talking about placement and I answered it saying that I don't know. I haven't looked at it but it's a great question.
CLINT: I think it's something that we need to start looking at is checking our tests on mobile and reporting those as well. A little bit more complicated because there's a lot of proximity involved in that, to ~ or not.
Sweet's the one who does all of our test results. We get all of our test results – at least, for the tests that Hunter, Dori and I do, Sweet's the one that gets all those.
We get them all from one central location because I can see something that Sweet might not see just because proximity screw stuff up so it's going to be a little tricky in figuring out how to do that and the best way that we don't put Sweet into the position where – the poor guy is doing like fifty different things trying to figure out how to pull that one off. But I would like to see us tracking that mobile too ~ to say Hey, maybe it's not a factor in one versus another. It is going to be an interesting lesson learned.