Build vs buy software: when owning an internal system beats renting it
Buy when the process is standard, the subscription is small and a product already fits how you work. Build when you pay a large recurring fee for a narrow system, when the product forces workarounds, or when the data has to stay on your own servers. The test is a 5-year comparison: total subscription and vendor fees against a one-time build plus yearly maintenance.

The decision is about recurring cost, not features
Most build vs buy discussions compare feature lists. The more useful comparison is financial and covers five years. A subscription is a cost that never ends and usually rises. A build is a cost that is paid once, followed by a smaller yearly cost to keep it running.
Neither is better in general. A company that builds its own payroll engine has made a mistake. A company that pays a five-figure subscription every year for a portal that shows news, a phone directory and six forms may also have made one.
This post gives a way to decide per system. It belongs to our writing on internal systems, the category of software a company can own instead of renting.
What each option costs in practice
The cost of buying
The subscription line is only part of it. A full count includes:
- The licence or subscription, per user or per company
- Vendor support and maintenance contracts
- Implementation days billed by the vendor or an integrator
- Paid change requests for anything the product does not do
- Integration work to connect it to your directory, ERP or HR system
- Price increases at renewal
- The cost of leaving: export, migration and retraining
The cost of building
- Scoping and design
- Development, testing and data migration
- Hosting or server capacity
- Maintenance: bug fixes, dependency and security updates, small changes
- A named internal owner who decides priorities
- The risk that the project slips or the requirements were wrong
The build side has more uncertainty at the start. The buy side has more uncertainty at year three, when the renewal quote arrives.
One company’s figures
These are the subscription costs of one industrial company with about 400 users, where Shai ran IT for four years. They are one company’s figures and not a market statistic. Dollar values are rough conversions at about 3.2 NIS per dollar.
| System | What the company paid | Rough dollar value |
|---|---|---|
| Off-the-shelf organizational portal | About 100,000 NIS a year (about 70,000 product, 30,000 support) | About $31,000 a year |
| HR system for managing CVs | About 8,000 NIS a month | About $2,500 a month |
| Employee travel management through an integrator | 8,000 to 12,000 NIS a month | About $2,500 to $3,750 a month |
Those three lines alone come to roughly 292,000 to 340,000 NIS a year, or about $91,000 to $106,000.
Looking at the wider set of rented internal systems and vendor services at that company, Shai presented an estimate to its CEO: replacing rented internal systems with systems the company owns could save about 800,000 NIS a year, roughly $250,000. That number is an estimate presented to management. It is not a measured result, and it should be read that way.
The point of the figures is the shape. None of these systems is exotic. Each does a narrow job for a few hundred people, and each carries a recurring cost that would pay for a custom build in one to three years.
A worked 5-year comparison (example)
The numbers below are a hypothetical example to show the arithmetic. They are not a quote, and the maintenance figure is an assumption you should replace with your own.
Case A: a large subscription for a narrow system. A portal subscription of $31,000 a year against a custom build of $40,000 with $6,000 a year for hosting and maintenance.
| Year | Rent (cumulative) | Own (cumulative) |
|---|---|---|
| 1 | $31,000 | $46,000 |
| 2 | $62,000 | $52,000 |
| 3 | $93,000 | $58,000 |
| 4 | $124,000 | $64,000 |
| 5 | $155,000 | $70,000 |
Break-even: $40,000 divided by the yearly difference of $25,000 ($31,000 minus $6,000) is 1.6 years. Over five years, owning costs $85,000 less. The example assumes no price increase on the subscription and no major rebuild of the custom system. A renewal increase moves the result further toward owning. A large change in requirements moves it back.
Case B: a small subscription. The same build against a product that costs $6,000 a year.
| Rent | Own | |
|---|---|---|
| 5-year total | $30,000 | $70,000 |
Here buying wins by $40,000, and it would still win after ten years. No amount of preference for ownership changes that.
The formula is short: break-even in years equals the build cost divided by (yearly rent minus yearly cost of owning). If the result is under three years and the process is stable, building deserves a serious look. If it is over five, buy.
When buying is the right call
We build software for a living, and we still tell companies to buy in these cases:
- The process is the same everywhere. Email, accounting, payroll, tax reporting, video meetings. A vendor serving thousands of customers will do it better and cheaper.
- Regulation changes the rules. Payroll and tax products are updated when the law changes. Owning that obligation is expensive.
- The subscription is small. If the break-even is longer than five years, the arithmetic is settled.
- Nobody will own it. A custom system without an internal owner decays. If no one will decide priorities and approve changes, buy.
- You need it next month. A product can be live in weeks. A build that is rushed is a build that fails.
- The product is the industry standard. If partners, auditors or new hires expect a specific system, the standard has value on its own.
When owning is the better choice
- The recurring fee is large relative to the job. The figures above are typical of this pattern.
- The product forces workarounds. When staff keep spreadsheets next to the system, you pay for the product and the workaround.
- Every change is a paid change request. A system you own is changed on your schedule.
- The data has to stay on your servers. A custom system can be installed on your own infrastructure.
- The process is how you compete. A quoting flow, a production schedule or a supplier process that is specific to you should not be squeezed into a generic product.
- Per-user pricing punishes growth. Owned software costs the same at 400 users and at 800.
A third option sits between the two: configure a platform you already pay for. If the company is on Microsoft 365, SharePoint and Power Automate cover many internal needs with no new subscription. See our post on how to choose intranet software for that three-way choice applied to portals.
Which internal systems are usually candidates
The systems that come up most often share a profile: a few hundred internal users, a narrow process, a high recurring fee.
- An employee portal with news, directory, forms and documents
- A vendor portal for supplier registration, documents and invoices
- An applicant tracking system for CVs, stages and interview feedback
- Travel management software for requests, approvals and expense reports
- Employee onboarding automation that creates accounts and tasks when someone joins or leaves
- Wider business process automation for approvals that still run on email and spreadsheets
At the same industrial company, Shai built the onboarding and offboarding automation as a fixed-price project of 65,000 NIS plus VAT, about $20,000. That was that project’s price there. Our published pricing is below.
A build vs buy checklist
- List every internal system with a recurring fee, including support contracts and integrator retainers
- For each, write the 5-year total at current prices
- Note who uses it, how many people, and which features are used in practice
- Mark the systems where staff keep a spreadsheet beside the product
- Get a fixed-price estimate for building the features in use, not the full product
- Add a yearly maintenance and hosting figure to the build side
- Calculate break-even in years for each system
- Name the internal owner for anything you plan to build
- Check contract end dates and export rights before deciding the order
- Start with one system, not all of them
Risks of building, and how to reduce them
Building fails in predictable ways. The scope grows, the one developer who understands the system leaves, or the system is never maintained.
The controls are plain. Fix the price per phase so that scope is decided before money is spent. Keep the source code in a repository the company owns. Require documentation and a deployment that another developer can repeat. Run the new system beside the old one before the subscription is cancelled. Replacing an old system in stages, with the data moved and checked at each step, is the core of legacy modernization services.
When the question covers many systems at once, the order matters more than any single decision. Mapping the systems, costs, contracts and dependencies, then sequencing the changes, is what digital transformation consulting is for.
What to do next
Run the checklist on your three most expensive internal subscriptions and calculate the break-even for each. If one comes out under three years, bring the numbers to a scoping call.
Shai does the scoping and the engineering through our business software development service. Projects start at $25,000, at a fixed price per phase, and the company owns the code. If the arithmetic says buy, we will say so.