# Build vs Buy: When to Use Off-the-Shelf and When to Build Custom
A founder called me after getting a $180,000 development proposal. The proposal was for a customer support ticketing system - their team needed a way to manage support requests from their B2B customers.
I spent two minutes on the internet. Zendesk starts at $55 per agent per month. Freshdesk has a free plan. Intercom starts at $39 per month. For a team of five support agents, any of these cost under $500 per month - $6,000 per year.
The $180,000 custom build would take 18 months to pay for itself in subscription costs. And that assumed the custom build worked perfectly, had zero maintenance costs, and never needed updates.
This is the most expensive version of a very common mistake: building custom when buying is obviously better.
The Framework: When Does Custom Make Sense?
Custom software creates a competitive advantage only when it does one of three things that off-the-shelf software cannot: implements a unique business process that defines your competitive advantage, integrates with your specific systems in a way that off-the-shelf software can't support, or provides an experience so differentiated that users make decisions based on it.
Everything else is overhead. Support ticketing, CRM, analytics, email marketing, HR software, accounting - these are not your competitive advantage. The way your product delivers value to users is your competitive advantage. Build for that. Buy for everything else.
The Real Costs of Building Custom
When founders compare "build vs buy," they usually compare sticker price: "Salesforce costs $150 per user per month, and our custom CRM would cost $80,000 to build - that's only 44 months of Salesforce!" This calculation is wrong because it ignores the real costs of custom software.
Development cost is just the beginning. Every custom system needs:
Maintenance. Software degrades over time as dependencies update, security vulnerabilities are discovered, and browsers or operating systems change. Budget 20-30% of the initial development cost per year for maintenance. An $80,000 custom system costs $16,000-24,000 per year to maintain even if you never add a feature.
Feature development. Off-the-shelf products get features from their entire customer base's funding. Your custom system gets features only from your budget. When Salesforce adds AI-powered forecasting, all their customers get it. When you want that feature in your custom system, you pay to build it.
Support burden. Someone on your team becomes the de facto administrator of your custom system. That's time not spent on your actual business.
Vendor risk elimination. Off-the-shelf tools have support teams. When something breaks at 2am, Stripe has an incident response team. When your custom payment integration breaks at 2am, you're calling a developer who is asleep.
I mapped these real costs for a client who had built a custom project management tool for their internal team. The initial build cost $45,000. By year three, total cost (maintenance, two major feature updates, one developer hire for ongoing support) was $140,000. A comparable Asana plan for their team would have cost $28,000 over the same period.
Categories Where Buy Almost Always Wins
Authentication and user management: Auth0, Clerk, and Supabase Auth handle login, SSO, MFA, and session management for $0-500 per month. Custom authentication is a security minefield and takes weeks to build properly. Buy.
Payment processing: Stripe, Braintree, and Adyen are mature, compliant, and deeply documented. Building custom payment processing requires PCI compliance, fraud protection, dispute handling, and support for dozens of payment methods. The "build" option is not really an option for most startups. Buy.
Email sending: Sendgrid, Postmark, and Amazon SES handle delivery, bounce management, and compliance for pennies per email. The infrastructure complexity of reliable email delivery is enormous. Buy.
Customer support: Zendesk, Intercom, Freshdesk. Unless your support workflow is genuinely unique to your business model, buy.
Analytics: Mixpanel, Amplitude, PostHog. Building custom analytics that can answer the questions you don't know to ask yet is very hard. These tools have years of product development behind their query interfaces. Buy.
Categories Where Build Might Win
Your core product differentiator. If your business IS the software - if your unique algorithm, workflow, or user experience is what customers pay for - build that. That's your moat.
Deep integrations with your existing custom systems. If you're a company with a proprietary database or workflow that no off-the-shelf tool integrates with, custom integration or custom tooling might be necessary.
When off-the-shelf costs become prohibitive at scale. At very high scale, off-the-shelf per-unit pricing can exceed the cost of building and maintaining a custom solution. This is a real calculation - but it's almost never relevant at startup stage, and often less expensive than founders think even at scale.
When off-the-shelf tools have genuinely unacceptable limitations. Some products have unique requirements that existing tools simply can't support - specific regulatory environments, unusual data models, unique user experiences. In these cases, build is the right answer because buy doesn't actually solve the problem.
A Decision Process
Before committing to build, do this:
Spend two hours researching existing solutions. Not a quick Google search - actually test three to five tools that might fit your use case.
Calculate the real buy cost: subscription cost x 36 months + integration/setup time.
Calculate the real build cost: initial development + (25% of development cost x 3 years for maintenance) + internal support time.
Evaluate whether the custom solution provides a competitive advantage that the off-the-shelf solution cannot.
If buy is cheaper over 3 years and doesn't provide a disadvantage, buy. If build provides a real competitive advantage that justifies the premium, build.
The customer support ticketing system founder didn't build custom. He implemented Intercom in a week and spent the $180,000 on product development instead. His product shipped two features that directly drove $400,000 in new ARR in the following year. The $180,000 custom support system would have shipped zero revenue.
That's the build vs buy calculus in practice.
Book a 30-minute call: https://calendly.com/alpsf/zoom-with-aleksandr