They've gone back to a simple SEO tool that gives you actionable items to resolve. Feedback?
www.pearanalytics.com
- Merge "Things you need to work on" and "full report". Put errors/suggestions on top.
- Auto-expand the top most report field or make it more obvious that those are links.
- Add re-analyze site button. Typing the URL in over and over again is tedious.
- "You have too many meta keywords." It says I have more than 10 phrases but I only have 9. Off by 1 in the code?
- Change "domain is too new" to yellow instead of red since there's nothing that can be done about it.
- "Good! You your page has a meta description." has a broken img link in the report text. Ditto with "This page has a title tag."
- "The CSS on this page isn't valid." The only invalid part is foo-border-radius. I think that's typical.
Pretty useful. Caught that my HTML wasn't valid anymore due to some changes I made recently.
It also convinced me to add a sitemap to my site even though it's only a 1 page weather webapp.
We've marked the reason why we need the email address, if it let you see the report without an email address, it's a bug :).. we'll fix it.
I've merged the work on and full reports together.
We're looking into the too many meta keywords issue, I'll fix that soon it's not a off by one issue, but I'm calculating words, instead of phrases.
We're working on display issues :). I'm a former sysadmin who's now doing UI :)
I need to get the images to show up in the copy better.
We're running your CSS against Jigsaw and just outputting the result.
One of the things that's on the way is "importance", so that the list of things to fix is displayed in order of importance. This would mean the css errors would be fairly low on the list.
I really wouldn't do that but it's not my choice to make :)
> I've merged the work on and full reports together.
I saw that. Makes it a ton better.
> We're running your CSS against Jigsaw and just outputting the result.
I figured as much. Not a big deal. It might be nice to include the link in the report.
One thing you might want to consider is building the report in real time rather than doing the status bar and then asking the user to click for the report.
You can do something like the following w/jQuery:
$.ajax({
url: '/api/v1/',
data: { cmd: "report_getdata", id: reportid },
timeout: 5000,
success: function(data) {
// remove everything but the table header
$("#reportinfo").find("tr:gt(0)").remove();
var obj = jQuery.parseJSON(data);
jQuery.each(obj, function(i, ddata) {
$("#reportinfo").last().append("<tr><td colspan=7>JSON report data row</td></tr>");
});
}
});On each report item I'm going to have a "re-check" button so you can make a change and just analyze that one part. Shouldn't be too hard to make the report in real time.
2) If they really do demand my email address to try their service, they should let me see a screenshot, at least, so I can get a sense of whether or not I want to risk giving them my email address.
I think that this errs on the side of giving too little information to make an informed choice about wanting to use their service. There are many sites that will, without requiring an email address, do some basic SEO analysis for me. Without knowing what their site does, I was not enthusiastic enough to actually try their product.
I'm still new to layout/jquery etc so I'll have the screenshot up in a few hours based on the feedback of other posters.
Was it a deliberate design decision to make it easy to access other people's reports (by just changing the number in the URL)? Obviously anybody could run any website through the tool, so there's not really anything private about it, but it still feels kind of weird. I've just had a fun 5 minutes seeing what everyone else has been running through it, which I'm not sure they'd all be happy about.
Possibly you could use a long random string instead?
Ooh fun indeed :) Yeah use a random string or cookie auth.
But, seriously I did think about using some hash or form of uuid, but long term plan is to let people make accounts.. for trending, and alerts..
but first, we're working on accuracy, then speed, and then usability.