The Question Is Not Build or Buy. It Is Who Owns the Capability in Five Years.
Most software investment debates get framed as a price comparison: the SaaS subscription against a development team. That misses the real decision. You are choosing who controls a capability your business depends on, what that control costs over the life of the system, and how quickly you can change direction.
There are three options, not two: build in house, buy SaaS, or build with an external engineering partner and own the result. Each has a cost curve that looks very different in year one than in year five. This is the framework we use with finance leaders.
Total Cost of Ownership Over Three to Five Years
Year one cost is the least useful number here. SaaS looks cheap because there is no build cost. In house looks expensive because you hire before you ship. The picture flips by year three.
SaaS cost grows with headcount, with feature tiers you are pushed into, and with renewal increases that compound. Add integration, data migration and admin time.
In house build cost is salaries, benefits, recruiting, tooling, cloud and management. It is largely fixed: payroll continues whether or not there is a roadmap to fill it. Run cost after launch is typically 15% to 25% of build cost per year.
Partner build cost is a defined project fee plus a variable run budget. You pay for the capacity you use, scale down after launch, and own the code outright. There is no seat pricing and no license fee escalating with growth.
An illustrative worked example
Consider a mid market distributor with 180 users who need an order management portal. The numbers are illustrative, not a quote.
- SaaS: $95 per user per month, rising to $110 and then $125 at renewals. With headcount growing to 240 users, five year licenses approach $1.4 million, plus roughly $150,000 of integration and admin. Total: about $1.55 million, and you own nothing.
- In house: Four engineers and a product manager at a fully loaded US cost of about $900,000 per year for a nine month build, then two for maintenance. Five year total: about $2.1 million, after a six month hiring delay.
- Partner: A scoped build at roughly $220,000 over four months using senior offshore engineers with AI coding agents, then $60,000 per year to run. Five year total: about $460,000, with the code and IP owned by the distributor.
The ratios change with your specifics, so run your own numbers through our build vs buy calculator and in house vs offshore TCO calculator.
The Hidden Costs of SaaS
SaaS is right for commodity capabilities. The problem is the costs that never appear on the pricing page.
- Per seat pricing scales with headcount, not value. Cost rises every time you hire, whether or not the new hire uses the tool.
- Integration is on you. Connecting to your ERP, CRM and data warehouse is rarely included, and vendor APIs change on their schedule.
- Customization has a ceiling. When your process diverges from the vendor's model, you change your process or bolt on workarounds nobody maintains.
- Renewal increases compound. Once your data and workflows live in the platform, your negotiating position is weak, and vendors know it.
- Data export is harder than import. Getting out cleanly can take months, which is lock in by definition, and the vendor can be acquired or sunset features you depend on.
The Hidden Costs of Building In House
A permanent team is the most expensive way to own a capability unless software is your core product.
- Hiring time. Recruiting senior engineers in the US, UK or Australia routinely takes three to six months per role.
- Fully loaded cost. Salary is typically 65% to 75% of true cost once benefits, payroll taxes, tooling, cloud and office are added.
- Management overhead. Engineers need leadership, product management and a delivery process, which often adds 30% to team cost.
- Attrition. Every departure costs months of productivity, and the hiring cycle starts again.
- Opportunity cost. Executive attention spent building an engineering function is not spent on the business. And a build needs eight people while run mode needs two, so a permanent team is always the wrong size.
Where the Partner Model Fits
A partner makes sense when the capability is a differentiator (so SaaS will not fit) but software is not your core business (so a permanent team is the wrong shape). The financial properties:
- Variable cost. Scale from a full build team to a small run team without severance.
- Speed to start. A scoped engagement begins in weeks, not quarters.
- Senior architecture without a permanent hire. You get an experienced architect for the decisions that matter, without carrying that salary indefinitely.
- Ownership. Under a properly written contract you own the code, IP and deployment, and can bring it in house later.
The risks are vendor selection, time zones and quality control. Our engagement guide and comparison page cover how to mitigate them.
How AI Coding Agents Have Moved the Build Cost Curve
This is what makes the 2026 decision different from the 2023 decision. In the builds we have scoped this year, features a traditional team would have planned as a quarter of work are consistently delivered in a few weeks by senior engineers directing AI coding agents, with tests written alongside the code and human review of every change.
The consequence for a CFO: the cost of building has fallen faster than the cost of buying. A capability that was correctly a "buy" three years ago, because building it would have cost more than a decade of subscriptions, may now pay back in eighteen months as a build. The break even point has moved toward building anything that differentiates you.
When building costs a third of what it did, the question stops being "can we afford to build" and becomes "can we afford to run our differentiator on someone else's roadmap."
This does not mean build everything. Payroll, email and accounting remain commodity purchases. It means the boundary between commodity and differentiator deserves a fresh look.
The Decision Framework
Run each capability through six questions. The pattern of answers points to an option.
- Differentiator or commodity? If customers choose you partly because of how this works, it belongs under your control. If nobody outside would notice a change, buy it.
- How many users, growing how fast? Per seat SaaS gets expensive at scale. Above a few hundred users, build economics usually win.
- How much customization? Configuring within the vendor's model means buy. Reshaping the workflow itself means build.
- How sensitive is the data? Regulated or competitively sensitive data pushes toward a build you control.
- What is the time horizon? Two years favors SaaS. A decade favors owning it.
- How deep is the integration? If this system must talk to five others and hold the master record, integration alone can exceed the build cost.
| Criterion | Buy (SaaS) | Build in house | Build with a partner |
|---|---|---|---|
| Time to first value | Days to weeks | Six to twelve months with hiring | Weeks to start, months to launch |
| Five year cost profile | Grows with seats and renewals | High fixed payroll, slow to scale down | Defined build fee, variable run cost |
| Customization | Limited to vendor model | Unlimited | Unlimited |
| Ownership of code and IP | None | Full | Full, if contracted correctly |
| Data control and residency | Vendor's terms and regions | Your choice | Your choice |
| Management burden | Low | High (hiring, leadership, retention) | Moderate (product ownership, reviews) |
| Lock in risk | High | Low | Low with owned code and docs |
Structuring the Financial Case
Board approval depends on how the numbers are framed.
Capex versus opex. SaaS subscriptions are operating expense. Custom development that creates a long lived asset can often be capitalized and amortized, depending on jurisdiction and accounting policy. Confirm treatment with your auditors early.
Payback period. Plot cumulative cost for each option and find the month where the build line crosses below the SaaS line. In the illustrative example, the partner build pays back in under two years. If payback sits beyond your planning horizon, buy.
The cost of doing nothing. Every quarter on spreadsheets and manual process costs errors, staff time and lost deals. Put a number on it. It is usually the largest line in the model and the one most often left out. Our B2B software ROI calculator surfaces it.
Questions to Ask for Each Option
Before you buy
- What has the price done at each renewal for existing customers in the last three years?
- What does a full data export look like, and how long does it take?
Before you build in house
- How long will it realistically take to hire the first three senior engineers?
- Who leads the team, and what does that person cost?
Before you build with a partner
- Does the contract assign us full ownership of code, IP and infrastructure accounts?
- What is the automated test coverage, and who reviews AI generated code before it ships?
- Can we see a written scope first? Our project cost calculator is a starting point.
How RG INSYS Fits the Partner Option
RG INSYS builds custom software and SaaS platforms for mid market companies in the US, UK, Canada, Australia and the UAE, using senior engineers paired with AI coding agents. That model delivers roughly 3x faster at about 60% lower cost than onshore rates, with 80% or more automated test coverage and human review of every change. Clients own everything we build.
If you are weighing a build against a subscription, we will help you make the case honestly, including when buying wins. See our new product development and SaaS development services, review transparent pricing including dedicated teams from $5,000 per month, or contact us with the capability you are considering and we will return a written scope within 48 hours.
Frequently asked questions
When is buying SaaS clearly the right decision?
When the capability is a commodity that no customer chooses you for, user count is modest, your process fits the vendor's model without workarounds and the time horizon is short. Payroll, accounting, email and most HR tooling fall here. Buying is also right when you need something running next week and the strategic decision can wait.
How has AI changed the build versus buy calculation?
AI coding agents directed by senior engineers have cut build time and cost substantially, so the break even point against a subscription arrives far sooner than it did three years ago. Capabilities that were only affordable as SaaS are now realistic to build and own, which pushes more differentiating functions toward build while commodity functions remain purchases.
Can custom software development be capitalized?
In many jurisdictions, development that creates a long lived internal use asset can be capitalized and amortized rather than expensed, while SaaS subscriptions are treated as operating expense. The rules depend on your accounting standards and the stage of development. Confirm treatment with your auditors before the project starts, because it changes how the investment appears on the income statement.
What should a partner contract include to protect ownership?
Explicit assignment of all code, documentation and intellectual property to you on payment, infrastructure and source control accounts registered in your name, a documented handover process, no proprietary frameworks you cannot maintain without the vendor, and clear terms for scaling the team down. If a partner resists any of these, treat it as a warning sign.
Make the build vs buy case with real numbers
Share the capability you are weighing and we will help you model the five year cost of building, buying or partnering, including the cases where buying wins. You will have a written scope within 48 hours.
Book a Free Consultation