CloudFlare review and how I reduced my bounce rate 94%

Just weeks ago CloudFlare debuted at TechCrunch Disrupt and was the runner up. How they didn’t win amazes me.

I was skeptical at first. I mean, a service that requires you to change your DNS hosting is asking a lot. It didn’t take long for me to realize how beneficial the 5-minute setup was.

Before I go on, let me make it clear that I’m not using any affiliate links here. I genuinely love CloudFlare’s technology and believe that everyone should be informed.

CloudFlare claims to be a performance and security enhancement. At the DNS level, they intercept traffic and weed out the bad requests (i.e. spammers, unwanted bots, malicious IPs, and more) before they even get to your server.

Additionally, they can insert Google Analytics tracking code for you. They can obfuscate email addresses to protect them from spammers. They cache your site’s resources around the world and are able to return the files directly from the cache server that is closest to the visitor. I could go on, but you get the point … they do a lot to speed your site up and keep it safe.

The proof is undeniable. Take at look.

CloudFlare Review

  • Average pageviews per visit rose from 2.06 to 3.98 – 90% better!
  • Bounce rate fell from 61.98% to 3.57% – 94% better!

I also had an average of 48 spam comments before CloudFlare … after, an average of 6.

These stats don’t mean much without an explanation. It should be obvious though. CloudFlare stops all the bad traffic from ever getting to my site, so Analytics is only tracking quality traffic. So my human visitors click around the site more and don’t leave immediately as opposed to bots and spammers. These stats don’t prove their experience is any better — although it should be. The increase in pageviews may also be due to the better load time due to the caching.

With all of the crappy traffic excluded my traffic stats are a much more accurate representation of my authentic visitors.

CloudFlare Review

The image above shows the traffic that tried going to my site. You can see that 81% of my traffic was from non-threatening, real users like yourself. The threats were weeded out and not allowed into my site.

There are some caveats to CloudFlare. Because they are brand new there have been a few hiccups. For instance, I’ve experienced an increase in downtime because they’ve made some hardware changes since their number of customers blew up after the TechCrunch coverage. Its minutes worth of downtime so I’m not worried. With time their DNS hosting will improve.

By now you might be expecting to pay a pretty penny. You’d be dead wrong. In fact, the basic services are free. You can pay for advanced security and real-time stats, but all the benefits I’ve seen are achievable through a free account. Really, give CloudFlare a shot. I wouldn’t have spent this much time doing a CloudFlare review if it wasn’t worth it.

27 thoughts on “CloudFlare review and how I reduced my bounce rate 94%”

    1. Unique visitors didn’t change according to Analytics — they don’t count spiders, and according to CloudFlare there were only 51 unique threats that didn’t make it through.

        1. Not quite. While that may contribute a tiny bit, the real reason is this quote, “CloudFlare stops all the bad traffic from ever getting to my site, so Analytics is only tracking quality traffic.” Instead of bots and spammers getting counted in my stats, they aren’t even let into my site. So the bounce rate now reflects that of actual good, human visitors.

          1. So pardon me if I’m a little confused. 51 unique threats with 24 visits contributed to 94% of your bounce rate?

            That’s also a number I’m confused about – how do 24 visits come from 51 uniques?

            I’m trying to understand, because a 4% bounce rate seems to be more some kind of technical issue than a real statistic. Even internal corporate sites, with only interested humans, tend to be several multiples higher than that.

  1. MD… its 51 unique “threats” that were stopped.

    Normally those 51 visitors would count as traffic and unique visitors on the site, they would also contribute to high bounce rate… but CloudFlare is stopping them from even reaching his site, so they are not counted on Google analytics and such. (but CloudFlare keeps track of the info)

  2. I’m sorry, but I have to help you out of this dream: Your bouncerate MIGHT have dropped and your pageviews MIGHT have risen. But that’s not what you can conclude from the analytics data above because it’s flawed.

    Apparently, around the same time you installed the CDN, a new GA code was added to your website. Either you added this manually, you added the “Google Analytics for WordPress by Yoast v4.06” plugin to your site, or maybe you added the GA code to CoudFlare which neglects to remove the original code.

    Anyway, you have two GA codes in your website which I spotted by looking at the source code of your site. Remove one of them and your bouncerate will be alive and kicking again, trust me. Let it run for a couple of weeks and then compare it to the ‘Before CDN’ era and it’ll look much more realistic.

    Sorry :)

    1. Its possible that you’re right. However, CloudFlare explicitly says, “Note: You do not have to remove the GA code from your site.” I can see how this might bloat the numbers, but not how it would drop the bounce rate.

      In either case, there’s no use for both scripts, so I’ve removed one.

  3. Every time a user calls on the GA script, that counts as one ‘hit’. Bouncerate is defined as percentage of users leaving your website after one hit. If every page contains two GA scipts, then every user that completely loads the page will count as 2 hits. This means there will almost be no users with only 1 hi -> thus almost no users that ‘bounce’ -> thus a very low bounce rate.

  4. I was involved with CF in the beta testing and I too was skeptical. While there are some false positives (like listing my site’s IP as a botnet), it has drastically reduced the wasted bandwidth by all the fly-by-night blog/web aggregators that scour the internet, but don’t have anything positive to show for it.

    The thing people need to realize is this: If your IP is caught by Project Honey Pot and the site you visit is hosted with Cloudflare, you’ll be directed to a confirmation page, where you have to confirm you’re human and then enter a CAPTCHA; you’re then showed why you were sent to that page. So if these are legitimate visitors, they can still view the site and fix the issue that got them flagged. But, of the thousands of hits CF caught from IPs, NONE completed the CAPTCHA (the admin page will notify you if they do), so it’s fair to say it’s blocking the crap and not harming legitimate traffic.

    CF has helped me A LOT with the music leechers on my netlabel, like the MP3 sites that link to another site; most are flagged as malicious and I see it daily in my threat control panel.

    I love CF and use it on all my sites, especially those with WordPress, given the hacking and spamming done to those installs.

  5. i have no idea if this is related or not but i’ve tried your contact form a couple of times in the last day and i can’t get past the human validator. the human validator code seems like it’s caching. right now no matter what browser i use or if i refresh the page the validator characters don’t change.
    the contact form won’t let me submit.

    is there another way to contact you?

  6. Personally, I would not use something like this until it gets out of Beta. The traffic it is asking to do the CAPTCHA could be legitimate traffic, but most people are not going to do the CAPTCHA just to see the site. You could be losing traffic because of this Cloudflare being in Beta and not working up to par.

  7. I agree with that last comment. I installed CF, then the next day went to look at my site, and I MYSELF got told I was a potential risk and had to fill in the captcha. So, I’m a risk to my own site? I think it is for sure stopping a certain amount of legitimate traffic. My traffic has dropped a lot (and earnings), and also my rankings have dropped too. But maybe that was a coincidence. Is CF bad for SEO? I thought Google liked fast sites.

  8. Another possible reason for the extreme drop in bounce rate (and increase in page views) is that the pro version prefetches other pages on the site when a visitor first lands, causing additional page loads for each visitor.

    So, even if the visitor only sees one page, they might load up several others in preparation.

    Just another possibility.

  9. I have CF on my photo site, which really gets a performance boost due to all the images.

    I have the security level dialed down to “Low” after seeing a visitor drop off. Traffic has picked up again as a result. Google Project Honeypot is flawed in many ways — the net is that it has a very high false positive rate. That will damage your site’s traffic since most vistors aren’t going to complete the validation page.

    Finally, if you have two copies of the GA Javascript on a page, it won’t count as two visits. GA appears to be smart enough to detect two instances of the .js in the same page.

    In summary, I love CF. It has reduced page load time and bounce rate, and pages/visit.

  10. I tested pages with two GA javascripts on one page. I can tell you it definitely is NOT smart enough. It won’t count double ubique visitors but it will definitely count double pageviews and hence (almost) erase your bouncerate.

  11. Every time a user calls on the GA script, that counts as one ‘hit’. Bouncerate is defined as percentage of users leaving your website after one hit. If every page contains two GA scipts, then every user that completely loads the page will count as 2 hits.

  12. well for me i don’t see a change in my site as of yet but i do see load time change but not much. but the reason i still use it is i see my site no longer have thoses 10 or 15 secound down times anymore
    oh and to all thoses running a blog with them use super cache not w3 total. that way your whole site gets backed up not just the text.

  13. BTW, I also like the fact the it can help to use Google Analytics on every page
    including CGI generated ones…

  14. I know this is an old post but I just found this out today and figured I’d share for anyone else reading. Your bounce rate probably didn’t drop that significantly by switching to CloudFlare. Check your source and look to see if your Google Analytics code is installed multiple times. My bounce rate dropped to 3% and sure enough my tracking code is listed in the source 3 times. 

Leave a Reply

Your email address will not be published. Required fields are marked *