Azure customers feel the heat with Datacentre “temperature issues”

Microsoft Azure has suffered yet another outage to their European cloud estate citing ‘an underlying temperature issue’. This time round, the cloud behemoth outperformed itself on June 19th by beating 2018’s January record of 10 hours (read about that here) by a full hour.

Azure stated on their status history page that the issue lasted 10 hours 46 minutes. Customers on the other hand were still voicing their issues well after the official 4.30am resolution, particularly on Twitter. Whoever is right, 2018 is proving to be another testing year for Azure’s customer service and support desks. But could much of the misery have been avoided with one simple change?

Out of curiosity, I went to the Azure status page to find an update in search of some form of statement to allay concerns. There was none. What could it tell me that I couldn’t work out for myself? There is an outage and access is down. As it turned out, not much more than that.

Like many, I then turned to Twitter to see what Azure and their customers were saying in real time. Hopefully something more helpful? Not so. Nearly every other post relating to the outage was a request for an update. Both @Azure and @AzureSupport offered little except to check the status page. The hours ticked by. The requests for updates continued. Check the status page. The requests got more animated. We’re sorry, check the status page. There’s a pattern forming here…

When everything is working perfectly, we neither care nor wonder what our services are actually doing, just so long as they are working correctly. We are only interested when the machine stops. It’s at that point where good customer service earns its keep. But what is good customer service? When things are going wrong, it’s only human to want to know what is going on. And when you’re paying, doubly so.

Reading what I could from the various and colourful commentaries, it was clear to me that much of the anger was down to the lack of meaningful information being communicated by Azure. Sometimes, words on a screen are just not enough. I don’t doubt that Microsoft were working very hard to fix things from the moment it all started going sideways. It’s almost as if the bigger the supplier, the more likely they will struggle to communicate coherently on a one-to-one basis. But this is just speculation, right? I decided to contact some of those stuck in the middle of this problem and ask.

Surprisingly, I found that those who were affected by the outage were more disparaging of the support (or lack thereof) than they were of the product. Here is one example of what most were saying:

“What I will say is that Azure status updates were generally poor, and more or less, said the same for hours. Their social media team simply pointed me back at the status page. It was clear to me that they didn’t really know what was going on for a while, and so it made internal updates equally difficult.”

This highlights four important lessons about customer service:

  • communicate regularly and often
  • share what you can, when you can
  • always be honest
  • don’t hide behind messages and feeds

It’s ironic that in these heady days of digital and cloud automation, perhaps a real person on the end of the phone keeping you informed could have made all the difference to the customer experience. No, things wouldn’t have been fixed faster. But the placebo effect of the personal reassurance that you do matter, someone does care and that you are being kept ‘in the loop’ goes a very long way until the fix is resolved.

When IT platforms and applications go down and affect the services to your customers, all the finger pointing begins and ends with the IT department.  IT needs to be able to provide answers on the problem and fixes quickly so that this information can be relayed on to customers. When your Cloud service provider isn’t forthcoming with information, what do you tell your customers to reassure them? How can you prevent irreversible damage to your revenues and brand reputation?

Poor communication from your Cloud Service Provider has a severely negative knock-on effect to your customers – something they won’t quickly forget. Likewise, determining when to trigger a DR plan or simply setting customer expectations would also be a lot easier.

Will Microsoft Azure change their approach to customer support because of this outage and outrage? Probably not. Yet it is a salutary lesson to more humble service providers out there, that if you want to differentiate your customer experience, a real support person is something you won’t find very often – if ever – from the big boys. Dare is say that at ONI, we still prefer to have real people involved at all stages with our clients.

If you’d like to know more about our data centre operations, our Business Assured approach or arrange a personal tour of our data centre to see things for yourself, contact us at getintouch@oni.co.uk.