by Dave Kerr
To stay competitive and relevant, enterprises need to be able to innovate quickly by designing new products and services and digitizing processes. This requires an agile and responsive IT function. The top agenda item for most IT leaders today is modernizing infrastructure and building the platform that supports a growing volume of high-velocity digital projects.
However, as companies invest in the hardware and software that will make IT the enabler of innovation, they face the age-old challenge: where to place their technology bets. From enterprise chatbot solutions to container orchestration platforms and API gateways, IT executives are faced with technology choices that they may have to live with for years. As veteran IT leaders know, the new shiny thing that looks like the answer to your prayers can turn out to be a technological dead end or only an interim solution.
In our work helping clients build their next-generation IT platforms, we’ve helped companies integrate new tools with great success. However, it’s the near misses and cases where things go wrong that have often taught the most valuable lessons.
Here are six of the key things to bear in mind when evaluating shiny new things.
1. Get the right tool for the job
Remember when SOA (service-oriented architecture) was the buzzword? Organizations scrambled to adopt increasingly sophisticated and feature-rich solutions that would allow them to realize its benefits. But in many cases they wound up purchasing incredibly overcomplicated ESB (enterprise service bus) solutions that over time made managing the software environment harder. When we work on implementation projects, particularly when we’re doing the developing and engineering, we often scale back these systems, because they've become ‘black boxes’ of complexity and difficult to maintain.
Lesson: Look beyond the bells and whistles and focus on what your organization needs. When looking at new enterprise tools, be aware of the extraneous features that aren’t what you really need. Focus on the features that are your core requirements.
2. Automation is a must
When you see powerful features in solutions you are considering, always ask how automatable or scriptable they are. No matter how impressive a feature looks in a well-rehearsed demo, if it cannot be automated and controlled programmatically by a development or DevOps team, it will be a bottleneck rather than an accelerator.
Lesson: Strong engineering teams have to automate to support high-velocity projects. If it can’t be automated, it might indicate the software is not as mature as it seems and might turn into shelfware that just takes up space in your shop.
3. Empower your engineers to make decisions
Give your engineers the time and resources to investigate potential solutions. They will be administering the solution day in and day out and will have to support teams that work with the solution. So let the engineers look under the hood before you decide to buy.
We often encourage teams to run ‘spikes,’ which are very short, focused development efforts using two or three engineers, maybe for a week at a time. The goal is to quickly learn about a solution to see if it really fits. These efforts can generate a huge amount of knowledge for the team in a short time.
Lesson: A few spikes led by engineering teams to assess the capabilities of solutions will not only give you deeper insight into how effective each solution really is, but will also give your team the chance to develop real expertise in evaluation, which will be invaluable for future purchasing decisions and implementations.
4. Think MVP
When you think about what you might want over the course of five years, you naturally think big—the scope is huge. But if you think about what you need for the next year, or even the next few months, the requirements are much simpler.
In the agile approach, we focus on the MVP (minimum viable product). We build what’s needed as quickly as possible, bring it to users, learn, and iterate. Applying this thinking to IT is critical; it’s more important to be able to change quickly than to make all the right decisions up front.
Lesson: The best architecture is evolutionary. It pays to look for solutions that can be easily updated and iterated. It’s hard to anticipate the requirements of the future, so flexibility and adaptability are paramount.
5. Know how to use open-source solutions
Enterprises often assume open-source solutions are not viable, usually due to concerns about lack of dedicated support. However, in our experience a simple, open-source solution backed by a large community can often be maintained more easily by your team than a proprietary ‘black box’ solution. Truly digital organizations trust their engineers, who can work with the open-source community and ecosystem to keep your solutions truly cutting edge. In many projects we have a healthy mixture of open-source frameworks and enterprise tools.
Lesson: Include open-source solutions in your considerations, and have your engineers assess them. The results will often surprise you.
6. Think people
Truly digital organizations are empowered by a core team of talented technologists. The shiny new things are just a few of the many tools the team will use to build your products and services. Give a great team the best tools and they’ll do amazing things. Give the team a fighter jet when all they need is a socket wrench and you will soon hear the sound of the crash.
Lesson: The ability to innovate is not bought off the shelf. Keep these lessons in mind when thinking about your next generation of systems, and make sure your strategy to excel is oriented around people, not tools.
Dave Kerr is a senior digital expert in McKinsey’s Singapore office.