Automation, if done right, can add a ton of value to an IT organization. Benefits include speed, quality, and improved productivity. But only a few IT organizations have been able to truly reap the benefits of it. Why is that? What is it that these successful companies do differently while others continue to struggle?
Here are 5 of the most common mistakes we see when organizations try to implement automation.
1. Automating Infrequent Activities
Automation works best when a manual task is well defined, predictable, and highly repetitive. The first two dimensions are key in determining if an activity can be automated. The third helps determine if it is worth automating. Yet when organizations decide to automate, they focus on the first two and rarely consider the third.
For example, automating code builds is a popular activity to automate. It clearly meets the first two dimensions. But if the project plan only calls for monthly builds…what is the point? At best all you have done is saved the developers a few hours of work. However, if the same automation is used for projects that require daily / weekly code builds. Now you are literally eliminating hours of work. Translate that to annual effort saved and you end up moving the needle.
2. Automating In Silos
Hand offs are the silent killers of efficiency. However, most IT organizations are arranged as a series of hand offs (silos) – from requirement to development to testing to operations. And like most silos, the focus is inward, resulting in optimized silos but sub optimized end to end process. No wonder it takes weeks to get a new code through all the way to production with, in many cases, sub-par quality.
The true value of automation kicks in when you stich various activities into a single process, that minimizes / eliminates hand offs, and orchestrates it across silos. In essence, you are not just automating the activities but also automating the hand offs. This in essence, is the heart of the entire DevOps movement. To break down the silos, re-engineer, and automate the process.
3. Forgetting To Reengineer the Process
If you look at the data, most processes follow the classic 80/20 rule. 20% of the process time is spent on performing actual “work”, 80% of the time is wasted on motion. Case in point…it used to take us days (sometime weeks) to open up a simple IP port. The actual time to perform the work takes only about 5 – 10 mins, however getting the work to the SME literally took days (creating tickets, routing, approvals, finding SMEs etc.).
Despite this fact, many organizations spend countless hours finding, agonizing, and optimizing the “perfect” tool (hint: there is’nt one) and spend little to no time on process. In essence, if all you to is focus on the tool, then all you have done is optimize the 20%. No wonder organizations scratch their heads when they put in automation and don’t get the expected results. Automation and process reengineering are like two sides of a coin. If you are truly looking to transform, you cannot have one without another.
4. Not Planning For The Long Tail Of Automation
Pssst…let me tell you a secret. Creating automation scripts is easy…. after all it is the sexy part of automation. The real heavy lifting begins once you have created it and now you have to constantly tweak and maintain it.
Applications and its components, for better or for worse, are constantly changing. Which means that the scripts have to be constantly tweaked. Yet few consider the after math. In fact, our clients have so many scripts that they simply don’t know what to do with it. A better solution is to plan for the long tail ahead of time, set up a governance model that is responsible for updating, maintaining, and pruning the scripts and in turn protecting the investment that they have already made.
5. Fear That Automation Will Take Their Work Away
Fact – many of our clients are leery about automation. Not because they feel it will not work, but are afraid that it will work too well. And that success will eventually lead to good people losing their jobs.
There are three key issues with this line of thinking. First, the assumption is simply incorrect. While automation will eliminate certain types of tasks it also creates new type of work. After all you need someone to manage the automation. Second, the genie is already out of the bottle. Automation is here to stay. Ignoring it will only make catching up to it harder. And third, not implementing automation may have the exact opposite effect and speed up unemployment. Look at Netflix (poster child for automation) vs. BlockBuster. A better strategy is to start small now, while the industry is still trying to figure it out, and try to get ahead of the curve.
Mustafa Kapadia is the Service Line Leader for IBM’s DevOps practice, a business advisory practice focused on helping large enterprises transform their software & application delivery. He has over 17+ years of experience in the tech. space, both as a service provider and a management consultant.
His expertise is in DevOps, cloud, lean, application development, outsourcing, offshoring etc. and has published several articles on DevOps. He lives in San Francisco Bay Area, is an avid blogger (when time permits), a speaker, and advisor to start ups.
You can follow Mustafa on LinkedIn: https://www.linkedin.com/in/kapadiamustafa and Twitter: https://twitter.com/mkapadiatweets