That scorecard seems greatly delayed. It took over a week to recognize some changes that I had made even though the page speed test was coming back fine. I'd pay attention to any messages on the speed test page if they are practical to deal with but the actual score takes time to adjust. The Server Response time of individual pages should be a good indicator of what google is actually seeing until the score catches up.

I wouldn't pay attention to the scorecard. It is actually fairly useless.

I agree, the score card is fairly useless BUT, the actual page speed improvements Google suggests are quite relevant and can offer some detailed insights to responsive issues people might not be aware of. There are definitely better tools for studying page speed generally but for those who aren't well schooled in the subject, the tools associated with site health can provide some useful information. It also provides insight into what G is looking at and what it wants.

I second Ember. The scorecard is inconsistent, inaccurate, and yeah .. useless.

I second Ember. The scorecard is inconsistent, inaccurate, and yeah .. useless.

Do you mean useless like an opinion with no supporting information? Useless like the inconsistent, often out of sync Adsense stats you stare at all day long. Jeez, am I really defending Google or am I just incapable of comprehending your detailed analysis on the subject? Just asking.

Inconsistent? There's definitely a lag.

Inaccurate? By who's determination? On who's scale of accuracy? Yes, I don't pay much attention to the actual score either but the most useful parts of that scorecard are the tools related to page speed and your opinions about usefulness or lack thereof are neither qualified (as in with supporting information) or considerate of the OP's experience.

I will offer a useful comment which is that there are a great many tools available to help analyze page speed and I'll just mention YSlow for an example. This still doesn't address the fact that the tools G created for page analysis (connected to the scorecard) can provide insight into what Google thinks is important and what Google thinks is probably what Google bases it's decisions on. Just guessing here.

@Pjman Whether you think the tools are useless or not is up to you. There is definitely a big delay in how the actual scorecard reacts (just like Adsense stats or the stats in webmaster tools) but even comprehending why Google put that there in the first place can offer a bit of insight into the strategy they're taking and that could be worth something in itself. Many webmasters still have no idea how well their site actually responds or how that response time impacts SEO and user interaction so you're ahead of the game just asking the question in my opinion.

I mean useless as in of no use.

You're making a mountain out of a molehill.

Perhaps but I'll contend that information is only useless to the person who has no use for it. Others may have a different experience.

I would still check a couple of things like gzip, keepalive, initcwnd etc. - since it dropped right after host change. Also, check if the listed 'analyzed pages' have changed, image/content heavy pages that gained recent popularity may drag your score down.

'analyzed pages' may also contain files such as Javascript files referenced in the head of the document or css files that can essentially delay page loading and Google is starting to get sensitive to those files and the delays they can cause (or maybe they always have and just now have a way of telling us about it). If G is looking at different pages than before it may be finding things like that and is trying to tell you about them. Also, G will ding you for not using expires headers and other issues. Since you changed servers, it's very possible that static images are not being served with an expires header and G will lower your score for that. Again, the score is what it is. It's the best practices the score is trying to encourage that I'd pay attention to.

I figured it out, thanks to you guys for pointing out what I should look for.

My new hosts default didn't Leverage Browser Caching.

I poped this in the root htaccess file and I'm rocking again.

<FilesMatch "\.(ico|pdf|jpg|jpeg|png|gif|html|htm|xml|txt|xsl)$">
Header set Cache-Control "max-age=31536050"
</FilesMatch>

Cache-Control is the HTTP/1.1 way of doing things which is a completely acceptable solution to that issue. It's not like that little ding on your scorecard was the next great government crisis looming on the horizon but on a very busy site, little things like that can make a huge difference in how server resources are consumed and how the user experiences the site. Glad you found what you were looking for.

Scorecard sux :(

1. I have my servers in US
2. I moved the static content on CDN, with multiple locations.
3. I have CDN for DNS!

So what? My Scorecard dropped like a rock!

The scorecard in general is full of errors. I've seen fluctuations on things, it tells me I have G+ buttons but I don't like it and don't have G+ anywhere.

Don't believe the scorecard, use other tools. gtmetrix is worthwhile...

I've seen fluctuations on things, it tells me I have G+ buttons but I don't like it and don't have G+ anywhere.

The original question was about "site health" in the scorecard. Just because there are several features lumped into the overall scorecard doesn't mean they have anything at all to do with each other. I'm pretty sure Google+ scores and Revenue Optimization scores are addressing something entirely different than server and site optimization which is what the "site health" section of the score card is about.

I'm not recommending you use the site health aspect of the scorecard for a primary analysis tool but the fact remains that is has alerted people to deficiencies in server and/or page configuration as evidenced by the OP's discovery regarding caching so it does serve a purpose. Advanced webmasters may find little new in what the site health/page speed reports are saying but a novice user might actually discover something about web development that they never knew about.

"One person's trash is another person's treasure" and "everything is relative."

<FilesMatch "\.(ico|pdf|jpg|jpeg|png|gif|html|htm|xml|txt|xsl)$">
Header set Cache-Control "max-age=31536050"
</FilesMatch>

An expire header way in the future can help too.

That Site Health Scorecard reminds me of the Windows Experience Index (WEI). It's really a useless and meaningless tool. Too bad there's now way to remove it. I guess I'll keep on ignoring it.

An expire header way in the future can help too.

Yes, the two actually work together.

It's really a useless and meaningless tool.

Just like a book sitting on a shelf gathering dust is useless and meaningless. I agree!

BTW, what resource did you use to learn that an Expires Header or Cache-Control are useful tools for a website developer? Is that resource also useless?

This morning a new category popped up - Multi-site - something to do with how sites look on different platforms and screen sizes. Then it disappeared later in the day.

I did some work recently trying to improve my scorecard. Revenue Optimisation goes up and down every day without me doing anything, however, when I look at the detailed items everything looks good. I therefore don't know what it is complaining about. Previously it was complaining about not having text/image enabled for all ads when I did have text/image enabled for all ads.

For Site Health it doesn't like page speed on certain pages because it tells me I can compress jpg images more than they are compressed already. Previously, it didn't like my non-asynch javascript and some aspects of my external css.

I fixed these problems and did see an actual increase in speed when loading pages, but because it still doesn't like my jpg files I am still getting a low score. After fixing some of the problems I thought I might at least get a higher score than before.

For Google+ I get a low score. I recently added the Google+ button to a lot more pages, but I still haven't done it for all pages. Despite my efforts, my score didn't increase even by one dot. Maybe it will only be happy when I have the button on every single page?

I wouldn't say that the scorecard is completely useless and meaningless because some of the improvement suggestions, especially for page speed, are quite relevant. However, after taking actions the results are not always what I would expect.

I'm told they are going to be releasing a much improved version of the Site Health tool in the first quarter of 2014.

Thanks for the heads up, Netmeg. I am not a professional developer - just a hobbyist - and I found some of the suggestions useful. Some improvements would be welcome, as would linking to some less technical resources. Technically, I'm not completely incompetent but I have limits. Many of the developer resources that the Scorecard directs you to are obviously meant for people with high technical knowledge and ability.

I find their Page Speed advice to be somewhat helpful: I got caching and gzip set up (things I had delayed taking care of), moved some js out of the head - but I tried their advice about moving the css link to <noscript> tags past the end of the closing </html> tag and maybe it works on Chrome, but I put it back where it belongs because it messed up my layout and did not load in FF so I'm sure IE would never find it. In the new UI, they rate page speed for desktop and Mobile, still don't know why it scores higher on Mobile, but it is useful to me to see what they are looking for.

You should get a second and third opinion. There are other services that provide suggestions on improving your site's speed. WebPageTest .org is one, another is GTMetrix. Personally I use both of those and Google's Page Speed test and get suggestions from them all, they don't match up 100% on ratings and tips. Caching and gzip are two biggies though that everyone should do if they can.

Is there a Dummie's Guide to setting up caching and gzip anywhere?

Not that I know of. If you are running WordPress, you might want to look at a caching plugin such as W3 Total Cache (more features but harder to configure) or WP Super Cache.

No WordPress - just handcrafted HTML, CSS and Javascript running on a Linux server. I will continue to read up on these features, but so far I have found that most articles are quite technical.

.htaccess file entries will do what you want.

Thanks Swanny007.