CSPs Are Driving Out SI and Custom Code Overspend in Software

If you dig down into the annual reports from communications service providers (CSPs), you’ll see billions of dollars spent on systems integrators (SI), external management consultants and software consultants. A large chunk of that is being spent on making sure software systems play along with one another and ensuring that processes perform in a relatively robust way. But one of the major problems with this expense is that it is never ending, with systems requiring constant customization by expert software engineers to meet the business needs of the operator.

However, systemic changes in the way the leading vendors develop, operate and maintain their software and services mean that the need for post-deployment work by a third party has been dropping off significantly.

A Shift Away From Major Software Releases and Piecemeal Upgrades

A fundamental shift has already happened in the way we approach telecom software development, and under this microservices and Agile/DevOps methodology, we’ll see the end of large, annual software releases and mid-sized, incremental fixes and upgrades. Instead, we’ll see a constant release strategy where best practices are derived from analysis of live systems data and fed back into the design of those systems and processes to optimize their performance.

With this constant improvement, patches for new assets or amendments for new products and services are brought down to essentially nil. For applications, vendors are also automating their deployments and moving to cross-functional teams that manage their own applications, which eliminates the dependency on external system integrators. In addition to this, the move towards cloud infrastructure, platform and SaaS means some CSPs are building more of their own capabilities where it will allow them to differentiate on digital application delivery. All of these are traditional SI areas of business.

The Impact on CSP Profit Margins

This fundamental shift will have various impacts on the overheads of SI services, custom coding and consultant fees. In the short term, there will be an uptick in operational efficiency, which will have a positive effect on the company but is notoriously difficult to measure tangibly. The mid-term effect is that the feedback cycles built into modular, microservices-based, DevOps software architectures will optimize themselves rapidly and phase out the need for expensive custom work. The long-term goal is that these costs will be eliminated altogether and that the platform approach becomes like an operating system for CSP/DSP business.

The flipside is that these eliminated costs may be incurred elsewhere in a complex multivendor environment: an eventuality which many operators have foreseen in their digital transformation strategies. To counter this, we are again seeing bigger deals and partnerships where CSPs are entrusting their entire architectural transformation to one experienced vendor/service provider. In this way, contracts often reflect a results-based success-sharing scheme by which the vendors are remunerated on a sliding scale depending on their ability to deliver business goals for the operator.