We previously tested if images helped a page rank. We are redoing this test because that test was inconclusive. However, since that test, we believe there has been a shift. Google has placed more importance on image SEO. SEOing your image can create more traffic flow so we wanted to see how many images does Google prefer.
Test Downloads
Transcription Below
This is a test 381 is how many images on a page boost or help the ranking? The idea behind this is none of the images are optimized. You'll just put on the page. The file name is there obviously but the rest of that stuff is all wiped out. The pages aren't optimized for keywords or they're all optimized for the same thing. 10 different pages, 990 words each and then they just have a different amount of images on each page.
The test setup, in case you want to try this yourself as simple as 1 through 10. You can probably go a lot farther if you wanted to but 1 through 10 seems like a good number to me. Then, this is the example. They just put it in between the paragraphs. Put the images in between the paragraphs and then you let it sit and see what happens.
At the beginning of the test you can see it shuffled around. It's 2, 6, 4, 10, 7 and then all of a sudden it dropped number 8. Kicked in, it dropped out and then it came right back again. For the purposes of this test the 8 is a good factor for images or is a good image count.
Here's a couple questions that this could bring in to you. First is you can implement this by just putting 8 images on there or I would probably recommend duplicating the exact test setup that they have here and then start implementing optimizations for the images. So, keyword specific file names, keyword specific exif data, etc. In there and duplicate it and see if you get the same result.
If all of them are all optimized for the keyword and then is that 8 still the numbers? So, now we might be onto something here and then that'd be worthwhile to put 8 images into your pieces of content or look at that as it as a thing.
Another build off of this test that we can look at too is is it images in the content or is it images total on the website? If the website is a blank WordPress install, for example, and the page is not the only image would probably be on a testing site would be what maybe some background or something like that. So, that could technically be one image and now we have 9 versus 8. So, look at that too.
If I put images or a slider or something like that do those count in this number two? There's definitely some build-off you can do on this test. I think if you're doing some link building and stuff there's probably safe bet to just go with 8 and you can optimize those and use 8 images off of your target site and do some image backlink building if you want to or set some exifs stuff and hope that the platforms that you're posting on don't strip the exif data.
I will tell you that most people – most sites strip the exif data unless you're specifically sharing photos like flickr or something like that. Those sites strip the exif data.
Again, this test is a cool test because it gives us a starting point to open up some other test ideas for sure. Right off the bat there I spotted three so I'm sure you guys can come up with some more and what this means and how we can implement it better especially since image SEO is a big thing.
Jesper pointed out is in Firefox she can inspect how many images are on a given page. You probably can use tools or you can use the inspects function or whatever. So, again, I bet this test was a great idea. It's a great setup and it opens an opportunity for you.
There's 4 or 5 more tests out of this one alone just as I sit here so check this out, look at it, duplicate it and then start tinkering with it. As Dori pointed out, the 990 words with 8 images makes it 8% density for images which is a good formula to use when you're building your tests. To me, because I'm a simple guy, I would just duplicate exactly what they did. I put 990 words on there and start messing around with these.
Here's another example. 990 words I got 8 images and I only optimize one for the keyword. Will that still win? So, 7 unoptimized images and 1 optimized image. Will that win or play with the numbers? Does one optimized image work better than 8 unoptimized images. Check that out and you can play around with those bounces and find that magic number. It'll take a little bit of time to pull it off.
Again, this one was run over months I think hunter said and is right up. So, after months of watching. This is definitely a long-term play but probably some worthwhile to check out.
Jesper pointed out something here is it's very cool. It seems strange at 8 images is number one but one image is number 4 and that's not a big difference. I wonder how strong the rankings ~ for the image number is. I think if you look at this it's 8, 6, 9. So, 9 maybe too much. 10, it started... It went... This one started number one, right?
Now, it's all the way down to number one. It's already down to the bottom then it moves around here. You see 10, 7, 1. It's really hard to look at these and say, "Hey, is there a too many? Is there not enough?", and if it is an image count issue why is 8 here, 9 here and 10s here. You think they would be clustered a little bit closer together. It would be almost cooler to wait and see if this washes out somewhere. It goes 8, 10, 9, 7 and then backwards. That would be cool to see but there's no discernible pattern as to volume.
Again, that's why we do testing though so it opens up all these more other questions for us and it expands us in our mind and our thought process and are we wasting our time. So, that would be cool to see.