I think this thing is just trained to like large hero images. Tried some of the most beautiful sites I know and got scores in the 10s and 20s, presumably from the presence of text.
Then I tried a number of sites with worthless huge hero images and got scores in the 80s.
I mean like all ai, take this with a grain of salt, but I think there is an obvious bias in this to the 'giant hero image that goes all the way to the fold' design that a lot of low effort start-up websites use.
Agreed. I tested two of my personal websites[1][2]. The one with the text-heavy poetry scored 38% with a "meh" comment. The one with the gratuitously oversized banner image got 72% with a "oh yeah!" comment.
It does have an AI running the analysis. I like my poetry website. Maybe the AI was trained on good poetry and reacted poorly to my stuff?
This sounds about right. I got dinged a lot on my website for "Visual Clarity" although imho it's a pretty standard blog page with sidenotes, and the "similar scores" all had a decent amount of text.
Looking at the "similar scores" carousel for higher scores was primarily huge hero images with little text.
My website's homepage which consists of a 128x128 picture of me, text and a table got 54. I guess the mere presence of a image made the AI like it a bit more.
Hi, Thanks for your feedback you are right. Currently algo evaluate how a user will feel...when they look at your site above the fold...it's like tinder you just look at the pic and decide to swap right or left.. this is based on : Users make lasting judgments about a website’s appeal within a split second. This first impression is influential enough to later affect their opinions of a site’s usability and trustworthiness...of course lot of areas we can improve...hope u enjoyed..
The results under "Here are few websites with similar visual scores as yours" were especially amusing for my site.
I've received a score of 26 and the similar websites were:
- Generic unstyled XML Access Denied document.
- Google account log in page.
- Generic unstyled "Your request has been blocked" text.
- "Please verify you're a human" HTML error message.
- Random JSON error message.
I entered my personal website, https://merely.xyz . That website is served with a single request; perhaps it's been classified as an error page because of that?
I tested https://gildas-lormeau.github.io/ (score 56) which is also a single request website. Almost all the websites with similar visual scores are okay for me.
I also got a 26 for a site I entered. It was a bit confusing because 26 happens to be a fixed point, the site gives you a score (26% of users will like your site) and also ranks you among all the sites people have entered (you scored higher than 26% of all sites entered).
Good ol' metadata strikes again. I wish AWS would make a virtual block device for accessing this content instead of network (at least as an option). Much easier to protect at the OS level.
The v2 service helps with this but it doesn't beat awareness of the issue.
Complete nonsense in the absence of any sense of why a site is designed the way it is. Is it a product brochure? Is it something functional (e.g. a search engine)? Does it target the general public? Or (say) highbrow art appreciators? Or tech geeks? In each case the aesthetic might be deliberately designed to attract/repel according to target audience. There is/can be no generic "attractive".
I'm not very convinced of this AI. I got 96% for http://peppermind.com and only 68% for https://talumriel.de - I'm almost certain that most users would prefer talumriel, and both scores them seem overrated to me. I mean, I really have no clue about design...
Thanks for this really. I looked it and trust me AI has got this completely wrong ..we will include this as exception when we train again.. thanks for trying..
My default Apache 404 page got a 38% of users will like it, 26 points for visually average, 20 for visual and 30 for clarity. After a few different URLs, my fail2ban kicked in, and the site complained about the URL being invalid. Fair enough.
Most of the page contains punctuation errors, even in the title.
Remove the space before the question mark.
In English, it is always an error. There should be no space between a sentence and its ending punctuation, whether that's a period, a question mark, or an exclamation mark. There should also be no space before a colon, semicolon, or comma.
I dont want to sound like a hater here, but why is it that so many people in this thread claim a 72% score?
This whole thing seems to me like it isn't any AI by the common understanding, but just a few hardcoded scores and criteria.
Now, thats probably wrong, and the AI is super neat, but I think it doesn't look very good when it does that.
I dont think the chances are high that so many different websites all score exactly 72%. Not 71, not 73, 72.
Its weird. Now the whole site looks nice and all, I saw the papers referenced, too.
But again, when I read that it's "97% accurate", I wonder how that is even measured? I understand that there are measures of how easy to use a site is, etc. So I wonder what cases those 3% were, where your AI was "wrong"? Did you hand-review them, or what?
What's it actually doing under the hood? Just taking a screen grab of the first few hundred vertical pixels x the full width? What's the actual resolution that it's grabbing, and how does the algorithm work and come up with scores? Is the system backed by some sort of data from a representative group of users?
> Visual Mind is an AI engine specifically designed for understanding and scoring visual appearance of a website. Visual Mind has analyzed over a million websites to achieve an accuracy rate of over 97%.
How does accuracy work in a project like this, where the result is subjective?
“You can improve visual score by using better images and improving site layout to make it little more denser.”
These suggestions need to be improved if they are going to be the output of this tool.
Better images? Define better. It doesn’t ask what the images or site are trying to achieve. Maybe for their intended purpose these images are the best.
And making it more dense? Again, why?
If you’re going to give design feedback it might be valuable to consider what valuable design feedback looks like. There are books on this.
Amusingly, from the comments here, it seems like everyone wants some kind of "HotOrNot" service for websites. Crowdsource the ratings! Then you can drop the "A" from "AI" :)
Yeas your's is militaristic...may be the algo though you have too much text..and no images so low visual appeal..may be a case we need to consider in the next training set..don't take the recommendation from algo on face value ..but just as a overall guidance..thanks for trying and sending over feedback.
We use Bootstrap at CareDash with larger hero images on landers with clear calls to action. This structure tends to perform well on A/B Tests too as it leads the user to the action meant to be taken.
I gave my website without the login credentials and got 38% for the 401 page. Then I added the credentials and got the same exact percentage. I know the creds worked, because the screenshot changed. I'm not going to say this assessment site is completely useless, because the colorized screenshots at the bottom are pretty cool and give me more theme ideas, but other than that...
And I can't even save the images, because they are just all the same screenshot, base64-embedded in the HTML and then altered with JavaScript-based filters! What is that!?
[+] [-] kgran|5 years ago|reply
Input: https://www.wikipedia.org/ Output: 38% of users are expected to like your site
[+] [-] BbzzbB|5 years ago|reply
[+] [-] z3t4|5 years ago|reply
[+] [-] gonzo41|5 years ago|reply
[+] [-] donatj|5 years ago|reply
Then I tried a number of sites with worthless huge hero images and got scores in the 80s.
I mean like all ai, take this with a grain of salt, but I think there is an obvious bias in this to the 'giant hero image that goes all the way to the fold' design that a lot of low effort start-up websites use.
[+] [-] rikroots|5 years ago|reply
It does have an AI running the analysis. I like my poetry website. Maybe the AI was trained on good poetry and reacted poorly to my stuff?
[1] - https://rikverse2020.rikweb.org.uk/
[2] - https://scrawl-v8.rikweb.org.uk/
[+] [-] jszymborski|5 years ago|reply
Looking at the "similar scores" carousel for higher scores was primarily huge hero images with little text.
https://myraah.io/index.php/visualmind/report/aHR0cDovL2pzen...
[+] [-] rasso|5 years ago|reply
[1] https://rassohilber.com
[+] [-] kroltan|5 years ago|reply
[+] [-] myraahio|5 years ago|reply
[+] [-] merelysounds|5 years ago|reply
I've received a score of 26 and the similar websites were:
- Generic unstyled XML Access Denied document.
- Google account log in page.
- Generic unstyled "Your request has been blocked" text.
- "Please verify you're a human" HTML error message.
- Random JSON error message.
I entered my personal website, https://merely.xyz . That website is served with a single request; perhaps it's been classified as an error page because of that?
[+] [-] gildas|5 years ago|reply
[+] [-] furyofantares|5 years ago|reply
[+] [-] kome|5 years ago|reply
[+] [-] stevekemp|5 years ago|reply
Visual Mind summary report for http://169.254.169.254/latest/meta-data
38% of users are expected to like your site
Yup. That's a security hole:
http://13.232.106.1/thumbs/visualmind/aHR0cDovLzE2OS4yNTQuMT...
[+] [-] eat_veggies|5 years ago|reply
[+] [-] myraahio|5 years ago|reply
[+] [-] jcims|5 years ago|reply
The v2 service helps with this but it doesn't beat awareness of the issue.
[+] [-] mikro2nd|5 years ago|reply
[+] [-] myraahio|5 years ago|reply
[+] [-] JohnStrangeII|5 years ago|reply
[+] [-] myraahio|5 years ago|reply
[+] [-] KaoruAoiShiho|5 years ago|reply
[+] [-] heikkilevanto|5 years ago|reply
[+] [-] modinfo|5 years ago|reply
Remove the space before the question mark.
In English, it is always an error. There should be no space between a sentence and its ending punctuation, whether that's a period, a question mark, or an exclamation mark. There should also be no space before a colon, semicolon, or comma.
[+] [-] myraahio|5 years ago|reply
[+] [-] lionkor|5 years ago|reply
This whole thing seems to me like it isn't any AI by the common understanding, but just a few hardcoded scores and criteria.
Now, thats probably wrong, and the AI is super neat, but I think it doesn't look very good when it does that.
I dont think the chances are high that so many different websites all score exactly 72%. Not 71, not 73, 72.
Its weird. Now the whole site looks nice and all, I saw the papers referenced, too.
But again, when I read that it's "97% accurate", I wonder how that is even measured? I understand that there are measures of how easy to use a site is, etc. So I wonder what cases those 3% were, where your AI was "wrong"? Did you hand-review them, or what?
[+] [-] doomlaser|5 years ago|reply
> Your website is better than 72% of millions+ websites analysed by Visual MIND AI
for: http://doomlaser.com
What's it actually doing under the hood? Just taking a screen grab of the first few hundred vertical pixels x the full width? What's the actual resolution that it's grabbing, and how does the algorithm work and come up with scores? Is the system backed by some sort of data from a representative group of users?
[+] [-] niknetniko|5 years ago|reply
> Visual Mind is an AI engine specifically designed for understanding and scoring visual appearance of a website. Visual Mind has analyzed over a million websites to achieve an accuracy rate of over 97%.
How does accuracy work in a project like this, where the result is subjective?
[+] [-] nf__85|5 years ago|reply
These suggestions need to be improved if they are going to be the output of this tool.
Better images? Define better. It doesn’t ask what the images or site are trying to achieve. Maybe for their intended purpose these images are the best.
And making it more dense? Again, why?
If you’re going to give design feedback it might be valuable to consider what valuable design feedback looks like. There are books on this.
[+] [-] ajani|5 years ago|reply
So
https://myraah.io/index.php/visualmind
netflix.com
primevideo.com
ALL score an exact 72%.
Not convincing.
[+] [-] sillysaurusx|5 years ago|reply
[+] [-] insickness|5 years ago|reply
[+] [-] mtm7|5 years ago|reply
Input: https://mtm.dev
Output: 38% of users are expected to like your site
[+] [-] pkolaczk|5 years ago|reply
https://pkolaczk.github.io/
[+] [-] aliceryhl|5 years ago|reply
https://ryhl.io/ https://myraah.io/index.php/visualmind/report/aHR0cDovL3J5aG...
[+] [-] myraahio|5 years ago|reply
[+] [-] vbezhenar|5 years ago|reply
https://vbezhenar.com/host-configuration.html
[+] [-] lulzx|5 years ago|reply
[+] [-] hipplec|5 years ago|reply
[+] [-] tgv|5 years ago|reply
[+] [-] LordAtlas|5 years ago|reply
[+] [-] thekeele|5 years ago|reply
This is an interesting concept and another way to get feedback besides asking friends / co-workers.
Would like to see the "Tips" section expanded in future iterations.
Website: https://keele.codes/ Report: https://myraah.io/index.php/visualmind/report/aHR0cDovL2tlZW...
[+] [-] rancar2|5 years ago|reply
96% of users are expected to like your site. https://myraah.io/index.php/visualmind/report/aHR0cDovL0Nhcm...
We use Bootstrap at CareDash with larger hero images on landers with clear calls to action. This structure tends to perform well on A/B Tests too as it leads the user to the action meant to be taken.
[+] [-] forgotmypw17|5 years ago|reply
And I can't even save the images, because they are just all the same screenshot, base64-embedded in the HTML and then altered with JavaScript-based filters! What is that!?