This week's MIT Sloan CIO Symposium brings together MIT academics and CIOs to discuss how IT leaders can overcome some of the common hurdles to
IT automation best practices: 7 keys to long-term success
These automation best practices – on teamwork, tools, and more – will help you automate wisely
The term "automation" can make everything it encompasses sound a little too easy. Consider the variant automatic – as in “automatic results” or “automatic success.”
Truth is, as with just about any other major technology initiative, there are no guarantees. Simply giving a team a new automation tool isn’t a turnkey proposition, as we noted in a recent post. Without the proper alignment with people and processes, it may be a setup for failure.
As Red Hat Chief Technology Strategist E.G. Nadhan recently noted, "Automation is an effective trigger to revisit existing processes and determine their applicability in today’s market. While it is true that repeatable processes warrant automation, it is important to make sure they’re being executed the right way. Automating the wrong processes proliferates chaos."
[ Read Nadhan's full article: 6 ways automation will impact the enterprise in 2018. ]
Automation success depends on a variety of factors, so you want to learn from your peers' experiences. Consider these expert recommendations for making your own automation implementation successful for the long haul.
1. Get buy-in from your team
Of course, this advice applies to any significant shift in IT strategy – say, a move to hybrid cloud or adoption of microservices and containers. As Ned Bellavance, director of cloud strategy at Anexinet, says, “Nobody likes having a solution foisted upon them.”
This may be particularly true with automation, which sometimes carries negative connotations. That need not be the case, especially if you’re explaining the strategy and how it will benefit IT and the greater business, both at an organizational and individual job level. Moreover, you’ll want buy-in on tooling: “Foist-and-retreat” isn’t a successful leadership tactic.
“Although you may already have some tools in mind, listen closely to what your team suggests. They are the ones who will be using it the most, and they won’t use it if they don't believe it can do the job,” Bellavance says. “Some other team members may have misgivings about automation and the amount of work it will take. Put their fears to rest by identifying and making some quick wins that simplify the team’s life.”
2. Don’t underestimate resource needs
Speaking of “automatic,” there’s a common misconception that “automation = resource savings.” Indeed, cost savings of all kinds – not just financial – are one of the potential benefits of increasing automation. Just don’t make the mistake of thinking automation is a set-it-and-forget-it proposition.
You’ll need to plan appropriately for the ongoing operations and maintenance of your automated processes – especially when changes are required.
“To be truly successful, managers must keep in mind that the majority of automation’s costs come when there is a required change to the automation pattern,” says Wolfgang Platz, founder and chief strategy officer at Tricentis. “Therefore, one of the primary considerations to be successful in any automation project is to ensure that the ability to maintain and rapidly change any automation in place can be executed efficiently.”
Platz points out that while some automation platforms enable users to make changes to adapt to new business requirements, this doesn’t necessarily cover every future scenario. Consider, for instance, the potential need for resources that can write homegrown automation scripts.
“When specialized labor around scripting is required, managers need a pure understanding of the total change-costs overhead,” Platz says.
3. Choose the right tools for your organization
Buy-in and resource planning lead to the next key success factor: Picking the right automation tools for your organization. Automating on a toolchain that your people don’t want to use, or that is unnecessarily complex, or that simply doesn’t match your needs – all of these will create problems that impede success.
“When selecting an ideal automation toolbox, I would focus on three key attributes: flexibility, simplicity, usability,” Bellavance advises. “Any automation tool should be flexible enough to cover 90% of your use cases.”
From setup and deployment to long-term ease of use, Bellavance stresses the need for each of those attributes in your automation tools for long-term results.
“Any solution with too steep of a learning curve will be a turnoff for most of your team, who are already busy trying to get regular work done,” he says.
4. Automate processes as part of a long-term plan
In our recent post on getting started with automation, CloudBees director of product management James Dumay noted the importance of beginning incrementally rather than trying to tackle your entire software development life cycle (SDLC) or a similarly large workflow all at once.
Your step-by-step approach to automation should be part of a larger long-term plan to achieve your goals.
A key part of that long-term plan will be accounting for ongoing operations and optimization. Don’t fall for the misconception that automating a process means you’ll never have to think about it again, or that it won’t require human intervention.
“Automation, like any other spoke in the SDLC, is a feedback loop requiring monitoring and maintenance,” Bellavance says. “Circumstances will change, the environment will mature, and new requirements will arrive on a regular basis. Those conditions will require the maintenance and updating of your automation framework.”
5. Consider a dedicated automation lead
Mark Kirstein, VP, products at BitTitan, notes that if automation efforts are simply bolted on to someone’s existing job description, you might not get the best results. Sure, if you’re investing significantly in automation, it’s likely to touch multiple existing roles. But that long-term strategy we just covered above might be best served by someone specifically responsible for its execution.
“Dedicate someone to lead automation efforts – and make sure they have both the expertise and the time to execute,” Kirstein recommends. “Successful implementation requires businesses have a grasp on the amount of information required for automation. This includes all inputs and outputs, variables, documented workflows, and identifying business processes where automation makes the most sense.”
6. Avoid the Rube Goldberg effect
Automation done right should simplify processes and reduce low-skill, repeatable work – to free up your team for projects that make the most of its talents.
[ Read ServiceNow CIO Chris Bedi's related story, How automation helped my IT team make time for innovation. ]
Automation done wrong will create as much complexity – if not more – than it eliminates, which kind of defeats the purpose.
“Try to avoid the Rube Goldberg effect when an automation solution becomes so complex that no one but the original author has a chance of understanding how it all works together,” Anexinet’s Bellavance says, noting the importance of strong, clear documentation for automated processes. “Try to create processes that are atomic, reusable, and simple on their own.”
7. Recognize where automation might not fit
While IT automation is most certainly expanding, being successful with it also requires a sharp eye for processes or services where it may not make the best sense in your organization. BitTitan’s Kirstein points to services that require significant customization or one-off deliveries, whether to different business units, partners, or external customers, as a good example.
“[Automation’s] best use cases are repeatable, yet time-intensive, processes with predictable outcomes that IT teams can build and run without constant attention, and [that] can be replicated for other customers,” he says. “Where customization is key, automation is not.”
Want more wisdom like this, IT leaders? Sign up for our weekly email newsletter.