inparlor.
engineering#custom-software#costs#engineering

Build vs buy: a framework for custom software

The total-cost math behind build versus buy, when SaaS wins, when custom wins, the buy-the-core-build-the-edge rule, and five real scenarios with verdicts.

4 min read·April 22, 2026

Priya Nair

Technical Director · Last updated April 22, 2026

"Should we build this or buy it?" is one of the most expensive questions a company asks, and most people answer it with gut feel. They build things they should have bought, and buy things they should have built. Both mistakes cost six figures and a year you don't get back.

Here's the framework we actually use when a client asks. It's not "build if you're technical, buy if you're not." It's a cost model and a small number of rules that catch most of the bad decisions.

The total-cost math nobody does

When people compare build vs buy, they compare the SaaS subscription against the cost of writing the code. That's the wrong comparison, and it's why so many build decisions go bad.

Buying isn't just the subscription. Building isn't just the initial build.

The full cost of each path

Dimension
Cost of buying
Cost of building
Upfront
Onboarding, data migration, integration work
Discovery, design, the initial build
Ongoing
Per-seat or usage fees that scale with you
Maintenance, bug fixes, dependency upkeep (15-25%/yr of build)
Hidden
Price hikes, lock-in, features you'll never use
The roadmap you now own forever, opportunity cost
Risk
Vendor pivots, gets acquired, or sunsets the product
Key engineer leaves, scope balloons, you under-budget upkeep
Time to value
Days to weeks
Weeks to months

The single number people forget on the build side is maintenance: budget 15 to 25 percent of the original build cost, every year, forever, just to keep custom software alive and current. A $120K build is really a $120K build plus roughly $24K a year. Over five years that's a $216K decision, not a $120K one. Run that math before you commit.

When SaaS wins

Buy when the thing you need is not what makes you money.

  • It's a solved, commodity problem. Email, payroll, CRM, helpdesk, accounting. Thousands of companies need the exact same thing you do. Someone has already built it better than you will, and they amortize the cost across thousands of customers.
  • You'd be reinventing a category. If a mature category with five credible vendors exists, the build will almost always lose on total cost.
  • Speed matters more than fit. You need it next week, not next quarter.
  • The requirements are stable and standard. You don't have weird workflows that no vendor supports.

The test: if a competitor uses the same tool and it gives neither of you an edge, buy it. Spending engineering budget to rebuild a CRM is spending your scarcest resource on a problem your customers will never see.

When custom wins

Build when the software is the edge, or when buying forces you into a shape that costs you more than building would.

  • It's your differentiator. The workflow is the reason customers choose you. The moment you templatize it into someone else's tool, you look like everyone else.
  • No tool fits without painful workarounds. You're paying for three SaaS tools plus a small army of Zapier zaps and a spreadsheet that someone updates by hand. That duct tape has a salary cost, and it breaks.
  • The data is the asset. You want full ownership and control of the data model because it compounds in value over time.
  • Per-seat pricing punishes your growth. At 12 users the SaaS bill is fine. At 400 users it's a tax on every hire, and the build pays for itself.

Buy the core, build the edge

This is the rule that resolves most "but it's not really either" situations, and it's how most good systems actually get built.

You almost never need to build the whole stack. You build the thin layer that's differentiated and buy everything underneath it.

  • Buy auth, billing, email, payments, error tracking, the database itself. These are solved. Rolling your own auth in 2026 is a security liability, not a feature.
  • Build the workflow, data model, and interface that make your product yours.

A modern custom build is mostly integration: Stripe for billing, Clerk or WorkOS for auth, Postgres for data, and your differentiated logic stitched on top. That's why "custom" no longer means "from scratch," and why the build side of the math is cheaper than it was five years ago. The leverage is in being precise about which 20 percent is actually yours.

Five scenarios, with verdicts

Real decisions we've helped make

Dimension
Situation
Verdict
Internal ops dashboard stitching 4 SaaS tools by hand
Team wastes ~10 hrs/week reconciling
Build. The workaround already costs more than the build.
A CRM for a 20-person B2B sales team
Standard pipeline, standard needs
Buy. HubSpot. Don't build a CRM.
A field-service app with a workflow no vendor supports
The workflow is the competitive edge
Build. This is the differentiator.
Payroll for a US company
Compliance-heavy, commodity, high stakes if wrong
Buy. Never build payroll.
Customer portal that must match your exact product data
Off-the-shelf forces an awkward data shape
Build the portal, buy auth + billing underneath.

Notice the pattern: build when the software is the product or when the workaround tax already exceeds the build cost; buy when it's a commodity or compliance-heavy; and split when only part of it is yours.

How to actually decide in an afternoon

Score the decision on four questions before you ever write a line of code or sign a contract:

  1. Is this our differentiator, or plumbing? Plumbing leans buy.
  2. What's the five-year total cost of each path, including 20 percent annual maintenance on the build and per-seat scaling on the SaaS? Run the real number.
  3. What's the cost of the workaround we're doing today? If it's high, the build's payback is faster than it looks.
  4. Can we buy the core and build only the edge? Usually yes, and that's usually the answer.

If you want help running that math on a specific decision, our Custom Software Development practice does build-vs-buy assessments before any build, and a fair number of them end with us telling clients to buy. Browse the cost guides for ballpark ranges, or send us the decision you're stuck on.


Read next:

Work with us

The team that builds this is for hire.

Priya sets architecture and engagement strategy across Inparlor's custom-software work, advising clients on build-vs-buy, scoping, and modernization.If this is the kind of work you need, here's where to start.

Get the next one

One operating note a month. No drip sequences.

Subscribe for one substantive piece a month plus the occasional working playbook we don't publish elsewhere.

Monthly notes on shipping software. No fluff, unsubscribe any time.

Want to talk?

Get a proposal

Send a 1-page brief; we respond in 48 hours with scope, pricing, and the team that would actually run the engagement.