Lessons to be learnt from the Argos website failure

Date: 29th October 2014
Author: Deri Jones

Let’s be honest and not pretend that websites failing is a shock (the BBC just wrote: Argos faces glitches after website update ) – they always have and they always will.  Even the biggest and best can never be 100% up for 100% of users indefinitely; it’s just the nature of the complexity of the technology and rapid pace it is being extended (it was the Nationwide Bank just last week: ‘online banking glitch‘.

But having said that, it’s also true to say that Argos could have done better – they should be aiming for 100% ( so should we all), and getting closer than they did.

There are lessons to be learnt, in terms of website management:  both management of the technology and management of problems.

You’ll notice above that I used the management word, in both of the  two headings where Argos could have done better to prevent their website failure being so deep.
Argos website crash 2014

Because ultimately, it is down to the management of every organisation, to ensure that the people and processes are in place to ensure that the business can deliver, and to avoid failure.  Every company spends sums of money on the various forms of insurance to cover fire, flood and being sued: and for those with customers outside the UK the insurance effect of currency hedging is used.

So what lessons can be learnt, to reduce the risk of following the management of Argos in allowing technology failure to impact customers?

Load Testing:  aka: ‘Proper Planning and Preparation Prevents Piss Poor Performance’

Excuse my french*, but that phrase is apparently official British Military wisdom. (* note to non-UK readers: see Wiki explanation of ‘excuse my french‘ )

And as tweeted by  Ruth Attwood @lady_rheena , the question is about Load Testing before this major website change:

Argos_website_fail__did_they_Load_test_adequately_asks_Ruth_AttwoodNow we haven’t been privy to the load testing planning at Argos for this big new launch.

But the Lesson to be Learnt No 1:  is that as eCommerce sites are becoming more complex and feature-rich, and as the visitor base uses a variety of devices from Android to iPad to laptop -there is no doubt that Load Testing is becoming more difficult, if a truly realistic load test is to be delivered.

  • Action item:  eCommerce managers need to be much more involved in speccing the realism of Load testing, then ever before:  don’t just delegate to tech teams and wash your hands of it!

After all, in Argos case:

so it was doubly worth-while to have tried to get this launch right first time.

Lesson No 2:  be ready for it to go wrong

Rather than a website that can’t even load a homepage, it is good to show a holding page:    but the Argos holding page was sadly, obviously not true – as twiterati pointed out, at 3am in the morning it is unlikely Argos is experiencing the ‘high volume of users’ their holding page suggests!

More importantly, it’s about planning insurance:  have Argos ever considered any clever ‘safety nets’ -maybe one like this:

  • a separate site: that is just a catalogue site:  no bells and whistles.  No order taking:  100% independent of the main site
  • but it does allow users to ‘request an order’: i.e. to pass a message to Argos, that will be processed later, after the problems are over.

Yes, that would cost a bit of money. But is that a good investment, if it prevents the black-hole of over 10 hours of total downtime impacting users and your brand?

That’s a decision only you and your FD can decide!  Hint:  first estimate the cost in wasted marketing budget and brand value, that those 10 hours cause!

Lesson 3 – Keep the customer informed

Right now, just hours after the problem cleared, there’s no mention of it on the Argos homepage, (though there is a page buried somewhere I’m told)

A little banner or ribbon, with a link to a Sorry page – that would be courteous.

Likewise the official Twitter page, could have been changed: to add a Sorry message to the page itself (not just a  tweet that gets pushed down over time)

Or even, if you have creative marketing types:   turn it into a positive:  add  some social networking to it, run a  competition to find the most funny or embarrassing story of someone who wanted to buy on Argos but couldn’t:  with the prize of say giving them back double what they finally spent; and a £500 cheque to the charity of their choice.

Lesson 4:  Start the dialogue with your FD and Marketing Director

As your next big website change is coming…when…so soon?    Better start preparing now, and first you need the budget and sign off!

 

PS – Some web pages are easier to serve than others – Argos ought to care about that too

Almost certainly this had nothing to do with the Argos website crash this week: but looking at their site today;  they are missing a well known site optimisation tip – that would:

  • speed up the user experience  for every visitor, especially mobile users;
  • and also  reduce the workload on their servers per visitor.

Simply, their homepage has 30 (thirty) separate javascript files needed in the background to build it!  – it’s possible that became worse since the new site launch and somehow didn’t get spotted.   That’s a nicer explanation, than that it;s been that bad for a long time.

How bad is 30 javascript files? I did a quick run through a bunch of the big UK retailers we work with here at SciVisum, and beyond:  the good guys are down at 8 or less:  sadly quite a few are higher but still only half as high as the Argos figure.

PPS : the wave of Argos Twitter apologies hits from 10am this morning

There are hundreds of personal apologies, at @ArgosHelpers egargos_crash_the_Twitter_apologies_24_hours_later

 

Top