Open Letter to BMI Baby

Written by Jamie Saturday, 26 November 2011 12:19

Dear BMI Baby,

How are you today?  Now, in terms of bad customer or web experience, I am going to score you a 1/5 - a quite low score, Dutch Rail recently scored a whopping 5.  I base my score on the fact that you said you are going to ring me to help me purchase a ticket and did not make me ring a hotline for 50p a minute.  But, you do get a score of one because your site is not working and this is annoying for your customers.

So, here I am then, on a Saturday morning, slowly waking up and trying to ignore the fact that I have to be in Birmingham, England, on Wednesday night.  Trying to ignore the fact that my working week is not quite over because I have to get my flights sorted.  I finally put the coffee on and, acquiesced to my fate, got the exact dates and times out and starting booking the flight on-line.  Your site is annoying because the purchase process is long.  Along the way you make me manually un-check boxes to receive promotional material, you force me to reject your offer of a car - some assumption, that, that I want to hire a car - etc.  All of this is just about forgivable.  However, what is not forgivable is that when I got to the end of the process your system stopped working.  This is what it said:




I thought, shit!  Then I thought, aw, never mind, these things happen, so I clicked the call me button.

 

 

I did keep the window open, as you asked, but, alas, nothing happened.  The phone didn’t ring, either.

This is absolutley unacceptable to your customers.  I mean, a CGI script?  It’s 2011, not 1991 - have you got any idea how large the memory foot-print will be for a CGI script?  I am no expert, but for a site like yours, you must be creating a lot of processes.  Let me guess: did you buy a shed-load of new machines because someone from tech told you that would speed the site up?

I look forward to your response, BMI Baby.  The site was not working at 11:30 nor at 12:00.  I will try again later.  In the meantime, consider the following, industry accepted practices:

• During non-peak times, deliberately switch off servers and pull plugs out.  This fault injection will help you to find failure points which you can then make fail-safe.
• Write a suite of robotic tests that continuously stress your most important features and, should the tests fail, your team can start to fix your problems immediatley.
• Focus all your efforts on building a continuous delivery platform.  This way, should the site break, you can patch and re-deploy in less than 10 minutes.  You are still down, nearly an hour later - what is this costing you?
• The aim for you should be to find all your own defects so your customer doesn't have to.  The secondary aim is to be able to fix and re-deploy quickly.

Sincerely, Jamie Dobson, Amsterdam, Holland.

< Prev		Next >