DevOps, WebOps, CloudOps, NoOps. Everyone’s talking about Ops in one way or another. To me, this explosion in ‘Ops’ terms represents a recognition of the key role Ops plays in business today.
Ops has always been important, but with the exponential increase in the number and speed of application roll outs and the volatility of brand reputation in an ‘always on’ society, the role of Ops has become critical to the business landscape. If your Ops fail and even a portion of your services go offline for any reason, brand value erodes very quickly.
Electrical retailer Dixon’s recently fell foul of this, when its website crashed during its annual sale due to a high volume of visitors. Retailers need to be able to prepare for this demand or prepare to lose customers – RightNow’s 2011 customer experience impact report for example shows that 89% of consumers began doing business or purchasing from a competitor following a poor customer experience.
The more important the customer interaction; the more important Ops is. When the BATS (Better Advanced Trading Systems) exchange went public and subsequently suffered a “software glitch”, it was forced to withdraw its IPO completely – an enormously rare and reputation-damaging move.
It’s not just investment banks either – retail banks have let thousands of customers down on numerous occasions when software glitches meant that Citibanks’ payments were going through twice when carried out via an iPad app, or when people were unable to access their accounts through Barclays’ ATMs and HSBC’s outage at the end of 2011 left millions of its customers with no access to services during one of the busiest trading periods of the year.
So the question is why do we define the problem as DevOps and not OpsDev?
We all know that the software lifecycle and certainly development precedes the release to operations. However in terms of cost and business value, Ops trumps Dev.
Development cares about innovation, not operational risk and stability. Development has always been the first to use innovative tools and the last to follow stable multi-function processes. Development will not solve the DevOps problem but operations can and this is because it’s a process problem and operations is more process-centric than development.
For me, in whichever context Ops is talked about, the industry is still getting it back-to-front. It should be OPSdev, OPSweb, OPScloud. Ops isn’t just important; it’s essential. Putting Ops second is a sign that although businesses realise its role, they don’t fully understand the far reaching effect of Ops.
At the recent IDC Retail Insights conference, retail giant Marks & Spencer spoke about its focus on its online presence and the development of consumer-facing technology. A move from a large project approach to regular iterative application releases on a monthly basis had been noted by many retailers at the conference resulting in a high demand for agile development skills.
Marks & Spencer seem to be moving in the right direction – more agile applications for more dynamic customer experiences. This will certainly increase the importance of DevOps for both firms and most likely enhance the processes, policies and systems between Dev and Ops. The question is which function should be the anchor point and take a lead on defining these processes and systems. The answer will determine their success.