Transforming software delivery: how Sodelico optimizes costs and enhances efficiency across projects

About Sodelico

Sodelico (software delivery cost optimizer) is an AI-driven platform designed to optimize every phase of your software development life cycle. Powered by statistical models, it can identify inefficiencies, provide suggestions for improvement, and predict the return on investment for every recommendation in your SDLC.

Sodelico can help your organization to:

• Reduce operational costs by up to 30%
• Improve your delivery timelines by 20%
• Mitigate risks across your entire development pipeline.

Andrei Stotski and Aleksandr Gushchin, experts contributing to Sodelico, shared outcomes from applying it on real projects.

Please keep in mind that some information was changed to protect client confidentiality. All calculations were performed based on the blended rate of EPAM, and it doesn’t reflect the account rates. The data provided solely pertains to pipeline and delivery costs, not the total budget or overall savings for the entire business.

Project A: ERP system

Problem statement

There were inefficiencies in the current QA processes, absence of automation, low maturity of engineering practices, lack of automation CI/CD, and architecture processes.

Proposed solution

EngX team started with the assessment of the client’s engineering maturity and software development lifecycle (SDLC).

The following steps were implemented:

• Developed a transformation plan with a roadmap for process optimization.
• Set up a data-driven delivery approach as a base for EngX transformation.
• Created QA artifacts based on recommendations from EngX in EPAM.
• Optimized the testing pyramid by reducing manual tests.
• Implemented significant changes in the testing pipeline.
• Enhanced collaboration between automation and manual teams.
• Refined EngX sprint planning to boost efficiency through regular syncs and activity monitoring.

Key results

• Transformation cost: decreased by €99K (from a projected €210K to an actual €111K), saving around 315 person days.
• Delivery pipeline cost: decreased by €225K per year (from €580K to €355K), saving around 640 person days.
• ROI of EngX transformation: 215%.
• Time for onboarding newcomers: decreased by 50% (from 16 to 8 months).
• Test automation coverage: increased from 49% to 59%.
• Defect containment efficiency: increased from 60% to 80%.
• Test design coverage: increased from 20% to 95%.

Average velocity has been increased by 20% — from 179 to 210 story points — during the period when a continuous improvement program started.

Project B: recognition portal

Problem statement

There were critical challenges in development processes connected with CI/CD, test automation, and code review.

Proposed solution

EngX team started with the assessment of the client’s engineering maturity and software development lifecycle (SDLC).

The following steps were implemented:

• Introduced cost-saving measures in the pipeline to enhance efficiency and optimize resource use:
  • Increased unit test coverage
  • Increased test automation coverage
  • Set up strict quality gates in the CI/CD pipeline
  • Implemented a data-driven delivery approach
  • Optimized seniority pyramid
• Applied Agile methodologies to improve self-organization, collaboration, and responsiveness to change through effective sprint planning.
• Incorporated code coverage into the Integrated Development Environment (IDE), allowing for early detection of issues and streamlined code reviews.
• Introduced a build verification suite to expedite post-deployment testing resulting in a quicker feedback loop and faster issue resolution.
• Identified areas of improvement to minimize time spent on manual QA tasks enabling teams to focus on more strategic initiatives.
• Streamlined (CI/CD) processes, reducing time-to-market and enhancing software release efficiency.

Key results

• Transformation cost: decreased from a projected €40K to an actual €31.5K for improving
• Delivery pipeline cost: decreased by €54K per year (from €156K to €102K).
• ROI of EngX transformation: 200%.
• Regression suite duration: decreased from 270 minutes to 180 minutes.
• Test automation coverage: increased from 5% to 52%.
• Rework rate: decreased from 15% to 5%.
• Cost of delivery pipeline: decreased by €54K (from €156K to €102K).
• Cost of risk: decreased by €46K in 8 months.

The graph below illustrates the distribution of risks, and the financial investment required over time for risk reduction. €11.2K were invested, and the overall benefits were €36.1K.

Project C: identity user management

Problem statement

The release cycle is not really fast. There were critical challenges in development processes connected with CI/CD, test automation, and code review.

Proposed solution

EngX team started with the assessment of the client’s engineering maturity and software development lifecycle (SDLC).

Assessment conclusion: improving quality and increasing automation helps deliver faster in a real project.

The following steps were implemented:

• Introduced cost-saving measures in the pipeline to enhance efficiency and optimize resource use:
  • Increased unit test coverage
  • Increased test automation coverage
  • Set up strict quality gates in the CI/CD pipeline
  • Implemented a data-driven delivery approach
  • Optimized seniority pyramid
• Applied Agile methodologies to improve self-organization, collaboration, and responsiveness to change through effective sprint planning.
• Incorporated code coverage into the Integrated Development Environment (IDE), allowing for early detection of issues and streamlined code reviews.
• Introduced a build verification suite to expedite post-deployment testing resulting in a quicker feedback loop and faster issue resolution.
• Identified areas of improvement to minimize time spent on manual QA tasks enabling teams to focus on more strategic initiatives.
• Streamlined (CI/CD) processes, reducing time-to-market and enhancing software release efficiency.

Key results

• Invalid defect (May–September): decreased from 10% to 5%.
• Defect containment: remained the same — 100%.
• Average build time in minutes (January–February): decreased from 40–50 minutes to 10–13 minutes.
• Deployment frequency (high envs): increased from 1 per 15 days to 1–2 per 14 days.
• Pipeline success rate based on average for two iterations: increased from 50–60% to 90%.
• Code maintainability rating: remained the same — A.
• Code reliability raring: remained the same — B.
• Code security rating: remained the same — A.
• Cyclomatic complexity: decreased from 20 to 15.3.

Based on metrics, EngX team decreased overall build time and increased the frequency of releases to production.

Conclusion

Andrei Stotski added some words about the limitations in the applicability of Sodelico: “Not every project can be improved. We primarily focus on software development, where pipeline enhancements depend on the team’s maturity. After an assessment, you may find no improvements needed if your maturity is high, which is perfectly fine. If you are interested in minimizing risks and reducing SDLC ownership costs with AI-powered software, book a demo now”.