Skip to main content

Buy or Build SaaS Software

In the age of AI Coding should you pay recurring expense for a product or build it yourself?

Purpose: Evolve a framework for understanding costs related to buying/renting operations software at a business and evaluating whether to build the software internally using AI.

The Debate​

AI is changing the calculation for whether to build or buy software:

  • AI reduces development time and cost, making in-house solutions more viable than before
  • However, important considerations remain: maintenance costs, support needs, opportunity costs, and the technical debt that accumulates over time
  • The value of established SaaS tools extends beyond initial creation to include ongoing improvements, industry expertise, and data advantages

Predictions​

Rather than spelling the end of SaaS, AI appears to be transforming it:

  1. Enhanced Capabilities: AI is improving SaaS applications through personalization, automation, predictive analytics, improved customer service, and better security
  2. Hybrid Approaches: The future likely involves a mix of custom AI solutions for core operations and specialized SaaS for non-critical functions
  3. SaaS Evolution: Established vendors are adapting by incorporating AI capabilities into their platforms

SaaS Evaluation​

Costs and Requirements

  1. Identify Business Needs
  2. Evaluate Options
  3. Cost-Benefit Analysis
  4. Compatibility and Usability
  5. Vendor Factors

Identify Business Needs​

  • Clearly define the core functions and features required by your business
  • Understand the specific tasks the software needs to perform and problems to solve

Evaluate Options​

  • Compare features offered by potential software solutions with business needs
  • Determine which features are included in base licenses vs. extra costs
  • Analyze actual usage patterns to avoid paying for unused features/capacity
  • Consider customization and integration capabilities and costs

Cost-Benefit Analysis​

  • Calculate total cost of ownership including upfront, recurring and hidden costs
  • Evaluate productivity gains, cost savings and business value enabled by the software
  • Compare costs and benefits of different options using matrices, charts, ROI analysis
  • Factor in long-term benefits like enhanced efficiency that can outweigh upfront costs

Compatibility and Usability​

  • Ensure new software integrates with existing systems to avoid custom development
  • Evaluate ease of use with existing workflows, processes and practices
  • Test options with end users to validate usability

Vendor Factors​

  • Research vendor reputation, financial stability, support quality and SLAs
  • For critical operations software, prefer established vendors in your industry
  • Understand vendor roadmaps and future development plans

In-House Development​

Evaluating In-House Development with AI

  1. Proprietary Data Footprint
  2. AI Capabilities
  3. Costs and Timelines
  4. Risks and Dependencies
  5. Hybrid Approach

Data Footprint​

Data is the new oil.

  • Where do you collect unique data?
  • What processes for developing valuable insights?
  • What domain secrets need to be protected?

AI Capabilities​

  • Evaluate maturity of AI coding agents to autonomously develop needed software
  • Determine if AI can deliver full functionality and quality required by the business
  • Test ability of AI pricing agents to accurately model costs of software options

Costs and Timelines​

  • Estimate costs to engineer and train AI agents to develop the software in-house
  • Project timeline for AI development vs. implementing an existing software solution
  • Factor in ongoing maintenance and enhancement costs of AI-developed software

Risks and Dependencies​

  • Assess risks if AI is unable to fully deliver working software as expected
  • Determine any dependencies on external data, APIs or technologies for AI development
  • Evaluate compliance of AI-generated code with any industry regulations and audits

Hybrid Approach​

  • Explore using AI to develop non-core software components to reduce costs
  • Evaluate integrating AI agents with expert human developers for critical components
  • Continuously monitor and refine AI agents based on feedback and business needs

Innovators​