The short answer: buy by default, build where the money actually is
For a team of ten or fewer, off-the-shelf software wins on five year cost almost every time, often by three times or more. At twenty five users on a Rs. 1,200 per seat product, a Rs. 6,50,000 custom build usually breaks even in year three and saves money after that. Those two sentences settle the question for most Indian businesses.
The decision is not really about software. It is about how many people will touch the system, how long you intend to keep it, and whether the workflow in question is a commodity or the thing your customers actually pay you for. Write down those three answers and the choice mostly makes itself.
The real cost comparison over three and five years
A subscription looks cheap in month one and a custom build looks frightening. That is a timing illusion, not a cost fact. Below is the same twenty five user operations tool costed both ways. All figures exclude 18 percent GST, and the subscription assumes an 8 percent annual price increase, which is ordinary for business software.
| Cost head | Off-the-shelf, 25 users | Custom build |
|---|---|---|
| Year 1 | Rs. 5,10,000 (Rs. 3,60,000 licences plus Rs. 1,50,000 setup, migration, training) | Rs. 7,00,000 (Rs. 6,50,000 build plus Rs. 50,000 hosting and monitoring) |
| Year 2 | Rs. 3,90,000 | Rs. 1,80,000 maintenance and hosting |
| Year 3 | Rs. 4,20,000 | Rs. 1,90,000 |
| Three year total | Rs. 13,20,000 | Rs. 10,70,000 |
| Years 4 and 5 | Rs. 9,45,000 | Rs. 4,10,000 |
| Five year total | Rs. 22,65,000 | Rs. 14,80,000 |
| Same model at 6 users | About Rs. 5,50,000 over five years | Still about Rs. 14,80,000 |
What the subscription actually costs by year five
The licence line is the obvious part. The parts buyers forget are the one time setup, data migration and training in year one, the annual price uplift, the higher plan tier you are pushed onto the moment you want an API or single sign on, and the seats you keep adding as headcount grows. Per user pricing is a tax on growth. Every person you hire raises the bill, whether that person lives in the product all day or opens it twice a month.
What the custom build actually costs after handover
The build is a one time number. What follows is not. Budget 15 to 20 percent of the build cost every year for framework and library updates, security patches, bug fixes, browser and OS changes, and small enhancements as the business shifts, plus hosting and backups on top. A build quoted at Rs. 6,50,000 with no maintenance plan is not a Rs. 6,50,000 project, it is a Rs. 6,50,000 project plus an unfunded liability. Our breakdown of software and website maintenance costs in India is a fair planning baseline, and the build side price bands are here.
Where the two lines cross, and what moves the crossing point
Three variables move the break even point far more than the quality of anyone’s sales deck.
- User count. Subscription cost scales with people. Custom cost does not. At six users the same build never pays back inside five years, as the last row of the table shows. At sixty users it pays back inside eighteen months.
- How long you keep it. A build amortised over two years is expensive. Over seven years it is cheap. If you cannot commit to keeping the process broadly stable for five years, do not build it.
- Integration depth. Every connector, middleware licence and sync tool you bolt onto a SaaS to make it talk to your other systems is real money and real fragility. Three of those and the arithmetic often flips on its own.
Run the table with your own seat count and your own tenure before you talk to any vendor, including us. If the numbers say buy, buy.
What off-the-shelf genuinely wins on
Somebody else carries maintenance, security and compliance
This is the biggest and most underrated advantage. Patching, uptime, backups, penetration testing, certifications and the never ending work of keeping dependencies current all sit on the vendor’s payroll, not yours. When a critical vulnerability lands on a Saturday night, their team wakes up. With custom software, that call is yours to fund and yours to answer, forever.
The vendor has already solved problems you have not hit yet
A mature product carries a decade of edge cases you do not know exist: audit trails, role permissions, bulk imports, rate limits, duplicate detection, e-invoicing quirks, the reconciliation screen your accountant will ask for in month seven. You also start on Monday instead of in month four, and delay has a cost that never appears on a quote. Momentum is a real benefit and buying it is a legitimate business decision.
What custom software genuinely wins on
It matches how you actually work, with no per user tax
Off-the-shelf tools encode the vendor’s idea of your process. If your process is genuinely different, and in a differentiated business it usually is, your team pays the gap in spreadsheets, copy paste and WhatsApp follow ups. Fifteen minutes of lost time per person per day across thirty people works out to roughly one full time salary a year, quietly. Custom removes that gap, and adding your thirty first user costs nothing.
You own the asset, the data and the exit
You hold the source code, the repository and the database. Nobody can raise your price by 40 percent at renewal, sunset the feature your operations depend on, or move your data to a region your enterprise client refuses to accept. You can also integrate on your terms with Razorpay, your ERP, your logistics partner or your accounting stack without waiting for a roadmap vote.
The middle path most Indian companies should take
The real answer for most businesses in the Rs. 1,00,000 to Rs. 10,00,000 range is not either or. It is buy the commodity, build the edge. Use standard products for the functions that are the same in every company, and build custom only for the one or two workflows that are actually your competitive advantage. This keeps custom scope small, which is the single best predictor of a project that ships.
| Function | Verdict | Reason |
|---|---|---|
| Accounting, GST returns, e-invoicing | Buy | Rules change every year and no customer chooses you for your ledger |
| Email, storage, payroll, HR records | Buy | Pure commodity, low per user cost, cheap to switch later |
| CRM for a standard sales motion | Buy | A pipeline is a pipeline, and building one wins you nothing |
| Your quoting, pricing or estimation logic | Build | This is where your margin lives and no product will model it |
| Core operations flow: dispatch, job cards, production, service routing | Build | Your sequence is yours, and workarounds here cost hours daily |
| Customer facing portal or mobile app | Build | It carries your brand and your customer data |
| Reporting across everything | Build thin | A small data store plus dashboards beats forcing one vendor to own it all |
The two traps, one on each side
Trap one: over-customising a SaaS until upgrades break
This is the expensive one, and almost nobody sees it coming. You buy a product, then pay a partner Rs. 2,00,000 to bend it into your process with scripts, custom objects and workflow hacks. It works. Then the vendor ships a major version and half of your customisation breaks. Now you are paying subscription fees plus a rework bill every upgrade cycle, and you own none of it. If you find yourself spending more on customising a product than the product costs, you have already chosen to build, just in the worst possible way. Take the hint and build properly.
Trap two: building custom with no maintenance budget
The mirror image. A company approves a Rs. 5,00,000 build, skips the annual maintenance contract to save money, and two years later the software runs on an unsupported framework, the original developer is unreachable and a rebuild costs more than the original. Custom software is a living asset. If the annual maintenance line does not exist in the budget, the project is not funded, and you should buy instead.
How to run an evaluation that is actually fair
Most buy versus build decisions are decided by whoever presented last. Run it properly instead.
- Write the ten workflows first, in your own words, before any demo. Mark each one commodity or edge. This document, not the vendor deck, is the evaluation.
- Cost both routes over five years, not one. Include setup, integrations, seat growth, price uplift, maintenance and hosting. Use the table above as your template.
- Run a real pilot. Take one live process, one real team of five to eight people, and thirty days of actual data. Not a sandbox and not a demo. Off-the-shelf pilots are usually free. For custom, pay for a small paid discovery or a working prototype of the single hardest screen, typically Rs. 50,000 to Rs. 1,50,000, and treat it as insurance on a Rs. 6,00,000 decision.
- Count the workarounds the pilot team invented. Every export to Excel, every parallel WhatsApp group, every manual re-entry is a gap in the product. Three or more in a core workflow is a build signal.
- Ask both vendors the same exit question. How do I leave, what do I take with me, and what does that cost?
Not sure whether to buy or build?
Send us your workflow list and your current licence spend. We will cost both routes over five years and tell you honestly if off-the-shelf wins, because a project that should never have been built helps nobody.
Ownership, GST and exit terms for both routes
Both routes attract 18 percent GST in India. Software development services are taxed at 18 percent, and so are SaaS subscriptions, including those billed from abroad, where a GST registered Indian buyer generally accounts for the tax under reverse charge. If you are registered, that GST is usually available as input tax credit, so compare quotes on the base amount and let your CA confirm the treatment. The details are covered in our note on GST on software and website development services.
On custom work, get these in writing before you pay anything: intellectual property assigns to you on final payment, in a named clause; the repository sits in your organisation account from day one, not the vendor’s; credentials for hosting, domain, database and third party services are registered in your company’s name; the contract lists milestones with payment tied to each; change requests are priced by a stated hourly or daily rate rather than negotiated in anger; and handover includes deployment documentation, environment setup notes and a working staging environment. Ask where the data is hosted, because enterprise and BFSI clients increasingly require Indian data residency.
On SaaS, the equivalent questions are: can I export everything including attachments and audit history in a usable format, what notice period applies on a price increase, how long is my data retained after cancellation, and what is the contracted uptime and support response. A vendor who cannot answer the export question in one sentence is a vendor you cannot leave.
When you should not buy custom software yet
Plainly, and against our own commercial interest, do not commission a custom build if any of these is true.
- Your process is still changing every month. Build software around a stable process, not an experiment. Use a cheap tool until the process settles.
- Fewer than about ten users and no growth plan. The arithmetic will not close. Buy.
- Nobody internally owns the project. Custom software needs one decision maker who can answer questions weekly. Without that owner, scope drifts and the budget doubles.
- You cannot fund 15 to 20 percent a year for maintenance. Then you cannot afford the software, only the invoice.
- A standard product covers 80 percent or more of your workflow with no ugly workarounds. That is a good fit. Take it.
- The real problem is process, not tooling. Automating a broken process just produces broken output faster.
A decision you can make this week
List your ten workflows. Mark each commodity or edge. Buy every commodity item without sentiment. For the one or two edge workflows, run the five year cost table with your real seat count, then run a thirty day pilot on the leading option. If off-the-shelf covers the edge workflow without three workarounds, take it and move on. If it does not, scope the smallest possible custom build around that one workflow and integrate it with the products you already bought.
Most companies that regret this decision did not pick the wrong side, they picked the right side for the wrong scope: they built everything, or they bought everything and then spent two years forcing it to fit. The middle path is less dramatic and it is almost always correct.
Our team is based in Pune and works with founders, directors and IT heads across India, and we quote both routes honestly, including the ones where we do not get the build. If you want that comparison on your own numbers, send the workflow list on WhatsApp at +91 86691 17894.
Also read: our guide to custom software development for Indian businesses and the price bands for building a SaaS MVP in India.
