Buy software when the capability is a commodity that many businesses need in much the same way, such as payroll, accounting or email. Build when the capability is a genuine source of competitive advantage and no product fits without heavy compromise. In between, the best answer is often a hybrid: buy a strong core platform, then build the differentiating layer on top.
That rule of thumb is simple, but applying it is not. Growing businesses face build vs buy decisions constantly, from customer portals and quoting tools to internal operations systems, and a wrong call can lock in years of cost or slow the business down. This article sets out a practical framework for making the decision deliberately.
What “build” and “buy” actually mean today
The choice is rarely binary. Most real options sit on a spectrum:
| Option | What it looks like | Typical fit |
|---|---|---|
| Buy off the shelf | Subscribe to a SaaS product and use it largely as designed | Standard processes where “good enough” really is good enough |
| Buy and configure | Adapt a product using its settings, workflows, fields and permissions | Common processes with some business-specific rules |
| Buy and extend | Use a platform’s APIs, apps or extension points to add custom features | A strong core product with a few important gaps |
| Assemble | Connect several products and services with custom integration and a thin custom layer | Workflows that cross several systems |
| Build custom | Design and engineer your own application, usually on cloud services and open-source components | Differentiating capabilities, or needs no product serves well |
Even “build” today rarely means writing everything from scratch. Custom applications typically rely on managed cloud infrastructure, open-source frameworks and third-party services for things like payments, authentication and messaging. The real question is where you want to own the design and the code, and where you are happy to rent someone else’s.
Start with the key question: does this differentiate you?
The most useful single test comes from software author Martin Fowler, who distinguishes between utility and strategic systems. His test is whether the way you perform a function is a crucial part of what makes you better than the competition. If it is, the software supporting it “needs to be as good as you can make it.” If it is not, he argues, “the obvious thing is to go with the package.” He also cautions that only a few projects are genuinely strategic, suggesting the split “may be more like 95/5.”
Ask yourself honestly:
- Would customers notice, and care, if this capability worked exactly like a competitor’s?
- Is the way we do this part of why customers choose us or stay with us?
- Does this capability embody knowledge, data or a process that competitors cannot easily copy?
Payroll, bookkeeping, email and most HR administration are classic utilities. A pricing engine that reflects hard-won market knowledge, a customer experience that sets you apart or an operational workflow that lets you deliver faster than rivals may well be strategic.
This matters because the strategic stakes can be high. In a 2022 McKinsey survey, one of three practices associated with top performers was “the development of proprietary assets, such as AI, data, and software, rather than a reliance on off-the-shelf tools”. That is not an argument to build everything; it is an argument to build where it differentiates.
Six factors that should shape the decision
Once you know whether a capability is utility or strategic, weigh six further factors.
1. Fit with your requirements
List your must-have requirements, then test candidate products against them with real scenarios, not feature checklists. A product that covers most of your needs out of the box may still be the wrong choice if the missing part is the part that matters most.
2. Total cost of ownership
Compare costs over several years, not just the first. Include licenses or subscriptions, implementation, integration, data migration, training, internal staff time, hosting, support, security and upgrades. Custom software has ongoing costs too: it needs maintenance, security patching and people who understand it.
3. Time to value
How soon do you need the capability? Buying usually gets a standard capability live faster. Building usually takes longer before the first release, but can deliver exactly what you need once it lands.
4. Control and flexibility
When you buy, the vendor controls the roadmap, pricing and release schedule. When you build, you control them, and you also carry the responsibility.
5. Integration and data
Consider how the solution will exchange data with your other systems, and whether you can get your data out in a usable form if you leave.
6. Capability and capacity
Building requires product management, design, engineering, testing and operations skills, either in-house or through a partner. Be realistic about whether you can sustain that effort after launch, not just during the initial project.
A scoring framework you can use
To make the decision transparent, score each viable option from one to five on the factors that matter most to you, and weight them according to your priorities.
| Factor | Favors buying when… | Favors building when… |
|---|---|---|
| Differentiation | The capability is a utility; customers do not care how you do it | The capability is part of why customers choose you |
| Requirements fit | A product meets your must-haves with configuration | Critical requirements need heavy workarounds or are missing |
| Total cost | Subscription costs stay reasonable as you scale | Per-user or usage pricing grows faster than the value delivered |
| Time to value | You need a standard capability quickly | You can phase delivery and the payoff justifies the wait |
| Control | The vendor’s roadmap aligns with your direction | You need to change the capability often and on your own schedule |
| Integration | The product has mature APIs and connectors for your systems | Your workflow spans systems in ways no product supports |
| Capability | You lack capacity to own and maintain software | You have, or can reliably access, product and engineering capacity |
Scores are a tool for discussion, not a formula that makes the decision for you. The value comes from forcing leaders to state their assumptions, especially about cost, risk and strategic importance, so they can be tested.
How to compare total cost of ownership fairly
Build vs buy comparisons often go wrong because the two sides are costed differently. The buy option is priced at the subscription alone; the build option is priced at the initial development alone. Neither is the real cost.
For a fair comparison, model a multi-year view, often three to five years, covering:
- Upfront costs: implementation or development, integration, data migration, testing and training.
- Recurring costs: subscriptions and licenses, hosting, third-party services, support and monitoring.
- Change costs: ongoing enhancements, vendor-driven upgrades and the internal effort to absorb them.
- People costs: product ownership, administration and the time your staff spend on the system.
- Exit costs: what it would take to migrate away later.
The cost of maintaining software over time is easy to underestimate. When McKinsey surveyed 50 CIOs at large financial services and technology companies in 2020, they estimated that 10 to 20 percent of the technology budget dedicated to new products is diverted to resolving issues related to tech debt. That was a survey of large enterprises, not small businesses, but the lesson carries: if you build, budget to keep the software healthy, not just to launch it.
Work with GeekOzy
Build the software your business actually needs.
Tell us what you're working on. We'll come back with honest thinking on a practical next step.
Start a projectExplore Software & Product EngineeringA useful precedent: the federal government’s order of preference
US federal agencies have a formal version of this decision. Under the Office of Management and Budget’s Federal Source Code Policy, memorandum M-16-21, agencies are directed to follow a three-step analysis: first look for an existing federal solution, then explore whether a commercially available solution can meet the requirements and only then consider custom-developed code, alone or combined with existing code.
The memo also lists factors to weigh throughout, including “performance, total life-cycle cost of ownership, security and privacy protections, interoperability, ability to share or reuse, resources required to later switch vendors, and availability of quality support.” Private businesses are not bound by this policy, but its sequence (reuse what you have, then buy, then build) and its checklist are a sensible default for any organization.
For a business, the equivalent first step is to check what you already own. Many companies pay for platforms with modules or features they have never switched on.
Evaluating vendors when you buy
If the analysis points toward buying, the quality of your vendor evaluation determines whether the decision holds up.
Test real scenarios
Ask shortlisted vendors to demonstrate your actual workflows using sample data you provide, rather than their standard demo. Involve the people who will use the system daily.
Check security and assurance
For software that will handle customer or business data, ask for independent assurance. A SOC 2 report, for example, is an attestation designed to help users of a service organization assess risks, covering the AICPA’s trust services categories of security, availability, processing integrity, confidentiality or privacy.
CISA and the FBI’s Secure by Demand guide suggests questions software buyers can ask manufacturers, including whether they offer standards-based single sign-on without extra charge, whether phishing-resistant multifactor authentication is available by default, whether security logs are included in the baseline product, whether they provide a software bill of materials and whether they publish a vulnerability disclosure policy.
Protect your exit
Before signing, confirm how you can export your data, in what formats and at what cost. Review contract terms on renewals, price changes and termination. Consider these points with your legal counsel, since contract terms and obligations vary.
Getting build decisions right
If the analysis points toward building, reduce the risk by treating the custom software as a product, not a one-off project.
- Start with a narrow first release. Deliver the smallest version that solves a real problem for real users, then iterate. Our guide to scoping and building an MVP covers the same discipline for software products.
- Buy the commodity parts. Use established services for payments, authentication, email delivery, search and hosting, so your effort goes into what differentiates you.
- Name a product owner. Someone in the business should own priorities and outcomes after launch.
- Design for change. Clean interfaces between components make it easier to swap parts later, including replacing a custom module with a product if one emerges.
- Plan for operations. Monitoring, backups, security patching and incident response are part of building, not optional extras.
- Own the code and knowledge. If a partner builds for you, make sure your agreement covers code ownership, documentation and handover.
The hybrid path: buy the core, build the edge
For many growing businesses, the strongest option combines both approaches. A proven platform handles the standard parts, such as accounting, CRM or ecommerce, while a custom layer delivers the features that make you different: a tailored customer portal, a pricing or quoting engine, a specialized workflow or integrations that tie systems together.
This path works best when:
- The core platform has well-documented, stable APIs.
- Custom work sits alongside the platform through its extension points rather than modifying its internals.
- You keep the custom layer small and focused on differentiation.
- You review regularly whether the platform now offers something you built, so you can retire custom code.
The main risk is dependency. If the vendor changes its APIs or pricing, your custom layer may need rework, so keep integrations well documented and isolated.
Warning signs you made the wrong call
Decisions should be revisited as the business grows. Watch for these signals:
- Bought, but should have built: growing piles of spreadsheets and manual workarounds around the product; customizations that break at every upgrade; subscription costs rising faster than the value delivered; the product blocking a strategic initiative.
- Built, but should have bought: most engineering time spent maintaining commodity features; one or two people who are the only ones who understand the system; features falling behind what off-the-shelf products now offer; security updates slipping.
Neither situation is unusual, and neither is permanent. Build vs buy decisions belong on your technology roadmap, with review dates, just like any other investment. Our guide to building a digital transformation roadmap explains how to keep those decisions under regular review.
The bottom line
Buy utilities, build differentiators and consider the hybrid path for everything in between. Make the decision with a fair, multi-year view of total cost, an honest assessment of fit and your capacity to own software, and a clear plan for security and exit. Then revisit it as the business changes.
If you are weighing a build vs buy decision, our Strategy & Consulting team can help structure the analysis, and our Software & Product Engineering team can design and build the custom parts when building is the right call. Start a conversation when you are ready.
Sources
- Martin Fowler: UtilityVsStrategicDichotomy
- McKinsey: Three new mandates for capturing a digital transformation’s full value (2022)
- McKinsey: Tech debt: Reclaiming tech equity (2020)
- OMB Memorandum M-16-21: Federal Source Code Policy
- AICPA & CIMA: SOC 2 – SOC for Service Organizations: Trust Services Criteria
- CISA: Secure by Demand Guide



