Your competitor is naked. Care to peek?

You’ve probably read the last case study about Magento Fails, so we had to follow it up with something even more eye-opening that dives even deeper into live e-Commerce performance data.  We tested 52 stores that are in the same industry to see how competitive their speeds are.

 

 

OVERVIEW

 

52 random e-Commerce site were tested ***(not just the home page, but the product pages that are actually accessing the database).***  The actual sites and hosts have been blurred out because this case study is not about acknowledgement or pointing fingers, it’s about measuring metrics that have an impact on the bottom line.  You will notice that Magento sites are either top performers or are sitting at the bottom of the barrel when it comes to speed / profitability.  That is because Magento performance optimization is an art.  If you are new to Magento, please be aware that you will not be able to achieve amazing results by simply adding a few plug-ins into your store while hosting on a shared platform.  Optimizing Magento requires the creativity of a developer’s brain to be lined up in perfect harmony with the brilliant process-driven expertise of a Linux Administrator.  It takes 2 to tango.

 

BOTTOM OF THE BARREL DATA

TEST SITE # 34 – Magento Community Store, Highlighted In Yellow:

The time to first byte (responsiveness of the server) clocked in at over 7.11 seconds.
Google likes to see sub 0.300 second results in this department for consideration in organic rankings.  The load time of site # 34 clocked in at 12.196 seconds.  How many web-visitors will stick around to buy a product there?  According to Kissmetrics, 40% of people surveyed abandon a website that takes longer that 3 seconds to load.  For test site #34, that means about 21,000 of their unique visitors in the month of January abandoned the site because it was too slow.  I wonder how much money they spent to get all that traffic there?  In case you were wondering, they are running Nginx and are doing some caching.

 

TEST SITE # 49 – Volusion Store (IIS 8 – expensive licensing), Highlighted In Yellow:

The time to first byte (responsiveness of the server) clocked in at over 0.583 seconds.  Not too bad, but not amazing.  The load time of site # 49 clocked in at over 33.639 seconds.  Would you wait over 30 seconds for a site to load ?(Keep in mind that there are so many other competitive sites that you can visit by clicking the back button and going back to your Google search).  My thoughts are the same, I probably wouldn’t wait unless it was a unique specialty site of some sort.

A typical business hears from 4% of its dissatisfied customers.  96% of unhappy customers don’t bother to voice their complaint and 91% will never come back.  3 seconds of waiting reduces customer satisfaction by 16%, so if a web page takes longer than 18.75 seconds to load, then the shoppers satisfaction is reduced by 100%.  TEST SITE # 49 took over 33 seconds to load, so is it possible that their visitors were 176% dissatisfied? (Thats a joke.)

How much money is being left on the table / snatched away by competition?  I don’t have that data, but I’m sure it’s not a laughing matter….

 

CREAM OF THE CROP DATA

TEST SITE # 14 – Volusion Store (IIS 6 – very pricey), Highlighted In Green:

Kudos to the people that made it happen, but at what expense?  A time to first byte of 0.552 seconds; not too shabby.  2.383 second page load time?…. very nice.  Don’t let this data completely win you over.  Their pages are only about 600 KB in size with images that are hard to look at and average at best.  Just about any decent platform can deliver a 600 KB page in no time flat.  The speed index (average time at which visible parts of the page are displayed) clocked in at 2661ms (about 2x faster than the competition).  This is particularly useful for comparing experiences of pages against each other (before/after optimizing, my site vs competitor, etc) and should be used in combination with the other metrics (load time, start render, etc) to better understand a site’s performance.  I’ll get into this more in my next case study.

Bottom line:  Yes, fast page load times can be achieved by average sites, but again; it takes 2 to tango so the site design should be appealing enough to convert shoppers while serving pages up as fast visitors expect them to.

 

TEST SITE # 37 – Magento Enterprise, Highlighted In Green:

The numbers speak for themselves.

0.177 time to first byte
2.334 second load time
Fully loaded in 3.876 seconds
Speed index value: 1415 ms & 779 ms on repeat view
Page size: 1,367 KB

While other Magento stores are either running Apache, or Nginx, or LiteSpeed; this Magento Enterprise store is on another level and in a class of its own.  It is running JetRails(TM) with eBoundHost; which is a unique way of stacking Apache, Nginx and Varnish together.  Looking at this data, it seems impossible to make a Magento store run well without caching (Varnish/Redis) and just on Apache or Nginx (or IIS for that matter).

 

THE POWER OF ONE SECOND

So whats the big deal here?  Why all this talk about site speed and performance?  It seems to be a very common pain point with developers and store owners because every second counts.  Just 1 second could mean the difference between a shopper choosing one brand over the other.  What is one second worth to you?

 

CASE STUDIES Expert advice, guidance and solutions from ideas to creation.  Real transformations. Read Case Studies	MANAGED SERVICES Changing platforms is only the beginning.  Value added services keep you competitive. Learn More	PERFORMANCE TEST Out-of-the box isn’t good enough for your investment. Discover your bottleneck. Test My Site