Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


ServerBear - Detailed Look at Stats Since Launch
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

ServerBear - Detailed Look at Stats Since Launch

serverbearserverbear Member
edited August 2012 in General

Hey guys, so ServerBear has been "launched" officially for about a month now. We've been around for 2 months (but we spent a month quietly testing).

We've met a heap of awesome people here & thought we'd do a huge update post so you can see how we're performing, what we've been up to (from a development POV) & have a bit of a look behind the scenes.

Warning, the post is quite long:

http://blog.serverbear.com/post/30226186538/aug-2012

Big thanks from the both of us for your support & feedback. You guys rock!

«13

Comments

  • Looking awesome, keep up the good work. :-)

  • TazTaz Member

    WHT (as it’s known for short) is similar to Digital Point in that there’s a lot of uneducated people that just reduce the quality of this forum to mush. You’ll see more posts about people complaining about downtime, or too many pointless filler posts from hosts trying to inflate their post count.

    This made me Lough. Nice post though mate! Best of luck.

  • You made $3k off 5k visits? Then again, web hosts scrub data like crazy or just flat out refuse to pay. I'd be interested in how much you actually make from the referrals.

  • @subigo said: You made $3k off 5k visits

    That's assuming that we had 100% coverage with affiliate programs. Which we don't.

    We might call out non-paying hosts each month. We'll see.

  • @serverbear said: We might call out non-paying hosts each month. We'll see.

    You should. I'm still owed thousands of dollars from Hostgator, Lunarpages, and others from a few years back.

  • prometeusprometeus Member, Host Rep

    I appreciate the openess and transparency attitude. Good luck for the coming months and years :)

  • zserozsero Member
    edited August 2012

    @serverbear, there is still no way to look at the actual benchmark on your site! That would be the only reason I visit your site, to actually look at the benchmark results.

    Also, generally I find your site extremely hard to use. Please understand it as a positive comment for making things better.
    For example, someone wants to look for the fastest IOPS host in Europe for under $7 a month.
    1. serverbear.com
    2. why do we have unixbench and IO split at the first place? It should all be based around plans, unixbench and IO are just two properties for each plan. And then why only 2? What about network speed of total bandwith? These all should be just properties of each host.
    3. benchmark vs compare? Again, what's this? Going to the same table layout, what do they mean, and why are they separate?
    4. OK, let's say I go to benchmark / IO. I tick HDD type. Wait.
    5. HDD type: totally stupid, you'll never know the underlying HW setup of a VPS, and it doesn't matter at all.
    6. OK, look at IO. No, not what I want, a simple caching makes lot of hosts having huge performance values, while not mirroring the real world performance. The most important for a LEB server is Random Read IOPS, as in ioping -RD.
    7. OK, no way to get what I want from that table, let's look at the actual results and compare them myself. No way to do it. How can I actually look at the results? After clicking for a dozen times, all I get is an annoying little popup window saying: buy this plan, 2 benchmarks. OK, 2 benchmarks, just show me! No, not gonna go. All I get is this place:
    http://serverbear.com/9796/urpad#view-benchmarks
    8. Try clicking on any benchmark: buy plan, 2 benchmarks. No way, OK, I close the window, relax, have a feeling of never going back to serverbear.

    All I just wanted to know is that what host has the best IOPS in Europe <$7. Took me 15 minutes and got annoyed by a stupid layout. Please redesign the UI.

  • zserozsero Member
    edited August 2012

    and: which is an incredible click through rate
    no, it's not, its all happening because of your stupid UI makes everyone go to the host instead of showing the results. 99% of those clicks actually wanted to go to the results but the most unintuitive UI just directed them out of your site, to the host's site.

    Trust me, 99% of your visitors are looking for the results, not for the host's site. The referral link should be only a link on a plan with clearly stating that
    - it's an external link
    - you are referrals on that link

    Not following these policies will give you a bad feeling to start with, not something you can change easily in the future.

  • @serverbear said: We might call out non-paying hosts each month. We'll see.

    You know how we love added drama around here, so yes, please do this! Great blog post, nice to see how everyone is benefiting from your efforts.

  • serverbearserverbear Member
    edited August 2012

    @zsero Thanks for the honest feedback, that's what we love! I sense some frustration, we're not totally happy with the layout either but it's challenging when there's so much data.

    Best way to do what you want is to click IO Benchmarks then choose the filtering options on the left. I just put in a ticket to get IOPS put into the filtering options. We'll probably do a more intuitive "plan finder" type thing, but we've been time limited on getting stuff done.

    2/3: They are leaderboards for UnixBench & IO, we don't have one for "network" because we're not happy with using the download speed from Cachefly. We're working on a network efficiency score. Compare allows you to compare plans without the benchmarks, remember less than 10% of the plans on the site actually have benchmarks. Some people want to compare by features too - plus you can use the filtering to add on what extra things you want to see.

    4/5: HDD Type is needed for us to create SSD specific categories/leaderboards.

    6/7: You want to filter by something that isn't implemented, I understand the frustration. We'll get it in ASAP. Ideally you should see a table ranked by the thing you want to filter.

    8: When you say you want to see results, you want a link to the full benchmark? i.e. http://serverbear.com/benchmark/2012/07/22/nhvwlr22zS4yz60O As stated in the blog post we're working on implementing that & a proper page for each plan with everything you need to compare. At the moment we summarise the results, lots of people don't want anyone looking at their full benchmarks.

    Regarding the clickthrough rates, we've added the tooltip so that people know they are going to be taken to the site to "buy this plan" when they do.

    Still lots of work to do, thanks for taking the time to post. I think some of the things you want will the fixed shortly, feel free to chat to me on IRC if you have any more ideas!

  • InfinityInfinity Member, Host Rep

    Hmm, apparently my ISP's DNS doesn't like Tumblr.

  • @Infinity: who do you have?

  • InfinityInfinity Member, Host Rep

    @HalfEatenPie said: @Infinity: who do you have?

    I'm sure you've seen me ranting everywhere, Virgin Media.

  • @Zen said: Screw corporations, get satellite - $5000/mo

    and get 2 second pings

  • Submitted a VPS6 XEN-512 benchmark and a correction to the locations!

    @Infinity said: I'm sure you've seen me ranting everywhere, Virgin Media.

    kinda off topic but doesn't VM filter a ton of websites?

  • @Zen said: Oh god super off topic mode :(

    sorry :(!

  • @serverbear don't know if this was already suggested, but you should have all prices in the same currency. It's quiet hard to compare this way.

  • InfinityInfinity Member, Host Rep

    Zen, they're not resellers, they just have BT's old infrastructure, at least here around London. Those green cabinets used to belong to BT.

  • @gsrdgrdghd Yeah, we're aware of this. We'll allow you to set a default currency at some point, the only thing there is that it might annoy people to visit a site (that we've shown a USD price for) then see the currency is Euros.

  • InfinityInfinity Member, Host Rep
    edited August 2012

    @Jack said: @Infinity Didn't it used to be Telewest/NTL down there?

    Yes, some of them. Same thing applies, I can't remember exactly all the details, somewhere on the VM forum.

  • We added FIO randread & randwrite to the script now:

    http://serverbear.com/benchmark/2012/08/26/DfdWBW8fpdg8Q06w#io

  • serverbearserverbear Member
    edited August 2012

    @zsero Just for you here's what you're after: http://serverbear.com/iops-benchmarks

    You can do country filtering down the left, this data still needs work & we need to work on regions too (i.e. Europe/West Coast, East Coast).

  • NanoG6NanoG6 Member
    edited August 2012

    @serverbear but still no actual - detailed results :(
    like this http://serverbear.com/benchmark/2012/07/22/nhvwlr22zS4yz60O

  • serverbearserverbear Member
    edited August 2012

    @NanoG6 Indeed, there will be soon. We're doing a week sprint in the next month to fix some of those core issues. Right now we're finishing off our uptime stuff first.

    Plus I don't want someone just scraping all the plans & stealing the data.

    Thanked by 1NanoG6
  • Awmusic12635Awmusic12635 Member, Host Rep

    @Serverbear I would recommend adding an area for Web Hosts to add coupons for the coupons page. You mentioned wanting to do more with it, that would allow you to not have to do as much work to make it happen

  • @Fliphost said: I would recommend adding an area for Web Hosts to add coupons for the coupons page.

    We're looking at doing something like this via a WHMCS plugin.

  • It'd be nice to have something closer to real world tests benchmarked - at least something we usually see in typical benchmarks like running web servers, database servers, etc...

  • @concerto49 Yep, easier said than done. Lots of factors at play, plus we'd need permanent access to the boxes (rather than once off tests) to continue to test over a longer period of time.

    We have implemented page responsiveness in our uptime code, so we can look at pings (plus other protocols) or actual page load times. So we could look at variances in page load time of sites hosted on a particular node.

  • Congrats on the growth @serverbear

Sign In or Register to comment.