Custom software vs SaaS: when it makes sense to build

"Should we build or buy?" usually gets asked at the wrong moment — after a year of fighting a tool that almost fits, or before anyone has tried the obvious SaaS product. For most businesses the honest answer is: buy first, and build when you have specific, expensive reasons. This guide is about recognising those reasons.

We build custom software for a living, so weigh our view accordingly. We also tell people to use Shopify, an off-the-shelf HR tool or a spreadsheet when that is the right answer, because a system nobody needed is the worst kind of project to have your name on.

The short version

Signs you have outgrown SaaS

• Staff keep a spreadsheet beside the software to do what it cannot.
• You pay for three tools that do not talk to each other, and someone re-types data between them.
• Per-user fees mean you ration logins, so the people who should use it do not.
• A rule that matters — a pricing formula, an approval, a compliance step — is done by hand because the tool cannot express it.
• You have asked the vendor for the same feature for a year.
• Your data is hard to get out, and you worry about what happens if the vendor changes its price or closes.

Two or three of these is normal friction. Five or six means the tool is costing you more than its invoice.

A three-year cost comparison

Compare the total over three years, not the first invoice. An illustrative example for an operations system with 40 users — your numbers will differ:

In this example custom is cheaper by year three — but change the inputs and it flips. At 10 users the SaaS licence falls to $9,000 over three years, and custom cannot compete. Run your own numbers: expected headcount, the price rise at renewal (ask what it was last year), and the cost of workarounds.

Price the time, too. If two people spend an hour a day re-typing and working around a tool, that is around 500 hours a year — often the biggest number in the comparison, and the one nobody puts on the spreadsheet.

Questions to put to the SaaS vendor first

Before deciding to build, give the tool you have a fair last chance. Ask the vendor, in writing:

• Can the rule we do by hand be configured, scripted or handled by an add-on — and who has done it before?
• Is the feature we keep asking for on the roadmap, with a date?
• What did prices do at the last two renewals, and is there a cap on future rises?
• Is there an API, so we can build a small piece around the tool instead of replacing it?
• How do we export everything, and in what format, if we leave?

Clear yeses to the first two mean you probably should not build. A no to the last one is a risk whatever you decide.

What building really involves

• Ownership. Someone in the business decides what the software should do next. Without that person, custom systems drift.
• Maintenance. Security updates, backups and monitoring never stop; budget for a support plan — ours start at ₹20,000 / $350 a month — or an in-house developer.
• Time. A focused system takes four to eight weeks to build and about a month to settle in.
• Risk. A bad developer can leave you with code nobody else can maintain. Insist on mainstream technology, documentation, and the source code and accounts in your name — the defences are in our guide to outsourcing software development.

The middle routes

It is rarely all or nothing.

Extend the SaaS

Keep the SaaS at the core and build a small app around it: a field app that feeds your CRM, or a customer portal that reads from your accounting software. It is cheap, and you keep the vendor's maintenance.

Start from a product, then customise

Begin with an existing product configured for you, and build only where you differ. Several of our own products can be set up this way: CoreStack People for HR and payroll, SitePulse for site attendance, and CalibIQ for calibration labs, which is in development. Configuration takes days; extensions are quoted like any other build.

Build only the process that is yours

Buy SaaS for everything standard — accounting, email, chat — and build only the process that is genuinely yours: the way you quote, schedule, inspect or deliver. That is where custom software pays back fastest.

When building is a mistake

• You have not properly tried the leading SaaS tool for the job.
• The process still changes every month; you would be building on sand.
• Nobody inside the business can own the system.
• The only reason is to avoid a monthly fee, and the fee is small.
• You want to build a product to sell but have not yet confirmed anyone will pay — prototype and sell first.

A decision checklist

• We have used the best SaaS option for a few months, or know precisely why it cannot work.
• We can list the rules or integrations SaaS cannot handle.
• The three-year total, including staff time, favours building.
• Someone inside owns the system.
• We will own the code, the data and the accounts.
• We are starting with one process, not the whole business.

Tick five or six and building is probably right. Tick two or three, stay on SaaS and look again in a year.

What custom costs with us

Our custom business software starts at ₹5,00,000 / $7,000 for a focused system, built in four to eight weeks, with a fixed price, a demo every week and the code handed to you. For wider ranges, see how much custom software costs in India; for ERP in particular, custom ERP vs off-the-shelf.

Get a straight answer

Tell us which tool you use today and what it cannot do — on WhatsApp or at krishan.joshi@corestacktechnology.com. We will tell you honestly whether to stay, extend or build, and if building makes sense, send a free design of the core screens within 48 hours.

Frequently asked questions