How to deal with legacy systems, the Achilles’ Heel of digital transformation

7 issues to keep in mind while dealing with legacy systems.

Companies that have been executing digital programs over the years have quickly realized that their investments in shiny new mobile apps and web interfaces have not moved the needle. In order to make ground-breaking changes within a business – digital or not – they need to transform the business and not just experiment with technology. Unfortunately, many businesses today are hard-coded in technology investments of the past: legacy systems – containing the data that they need to make decisions today and drive insights.

Now more than ever, reliance on legacy systems is one of the biggest hurdles in the digital transformation journey. The technology industry smells blood and is full of advice: move everything to the cloud, invest in platforms, replace legacy systems with SaaS software, start a legacy modernization program, and create the next generation architecture.

The biggest problem with this advice is that it puts technology first. Here are seven things to keep in mind while dealing with legacy systems:

1. Do not modernize systems ahead of the curve

In the last 6 to 7 years, many legacy modernization programs across the industry have failed and large system implementation projects have been woefully underperforming. Recently, a large American chocolate manufacturer tried to replace its legacy systems for production and distribution with a large ERP/SaaS solution. The project that was promised in April was delivered in July, because the new implementation did not match the previous system in performance or functionality. This ultimately caused the business to miss their key commitments for Halloween, with sales dropping by 12 percent and revenues by 18.9 percent in that quarter.  Large programs aiming to get the application layer ready for digital are bound to fail as they chase elusive outcomes and focus on technology upgrades. It’s important to proceed with care and focus on what matters most when modernizing your systems.

2. For reuse, expose the business logic within the systems

Completely replacing or changing legacy systems is a risky endeavor. Last year, retail companies (Metro and Otto) in Europe, instead adopted a unique approach: building wrappers around legacy systems. These wrappers bring the required business logic, which is implemented deep within the legacy systems, to the surface, retaining the stability of the legacy system while also addressing the difficulty in breaking these monolith systems. New streaming technologies, like Kafka, are now making the vision of these wrappers a reality. This pattern, named “The Strangler Pattern” by Martin Fowler, has helped multiple companies reuse the logic sitting within their mainframes, without undertaking costly mainframe modernization programs.

3. Change legacy systems during the building of business capabilities

Instead of spending energy on modernizing a legacy system, invest your resources in achieving a business outcome. These business outcomes should ultimately lead to the evolving framework of new business capabilities. Use the knowledge required to implement the solution and migrate an existing legacy system towards a concrete result. The team that chases a new digital business outcome should include members who know the legacy systems inside and out, as new solutions and implementations will use the business logic already within legacy systems.

4. Build a digital backbone for service architectures

As business solutions become encapsulated within micro-services, companies will need to replace their old enterprise services bus with modern, event-driven backbones using technology like Apache Kafka. Such backbones will enable organizations to blend newly implemented solutions with exposed business logic from the legacy solutions. This new set of micro-services can now be used to flexibly implement any change in business processes during the digital transformation and eases the pressure away from modernizing legacy systems overnight.

5. Manage your technology portfolio like products

It is equally important to be aware of the technology estate that has grown over the years. Bring product management thinking into the enterprise system with portfolio management. Develop a better understanding of the technology asset portfolio by plotting the large systems alongside a matrix of the systems associated risks vs its expected lifetime. The size of each bubble should indicate the business value that the system is currently supporting. This, combined with a health check of the system, gives a first glance visual indication of the problem areas in an enterprise system portfolio.

6. SaaS is not a silver bullet

SaaS offerings on the market can be a distraction. We have been lured in by packaged software before – earlier it was ERP, now we have SaaS, and while packaged software as a service, by definition, is not wrong to implement, it should not be the first decision made. If you believe that SaaS software comes with best-in-class functions, then you’ll need to buy the system and tailor your organization to comply with the system. If you think you will change the features, or believe that the functionality touches your core competency, then err towards implementing your own solution. Remember, a SaaS decision translates into the immediate creation of new legacy systems.

7. Take an evolutionary view towards legacy modernization

Business outcomes should always take priority, not systems – modern or legacy. And this should be the lens through which all investment decisions should be made. Large software upgrades that introduce blackout periods where functionality can’t be introduced are not helpful to anyone. Instead, we can expose the business logic hidden away in our legacy systems, and simultaneously retire the most problem-ridden systems over time. This evolutionary approach to legacy modernization will ensure that technology spends are distributed wisely.

Above all, move the focus away from systems. Pivot instead to an organization that can fluidly achieve local and global business outcomes with solutions based on partner ecosystems. Even if the goal is to create a modern system, we should remember that today’s modern system is tomorrow’s legacy.

This article is published as part of the IDG Contributor Network. Want to Join?

NEW! Download the Fall 2018 digital issue of CIO