The decision in four sentences
If a reputable product already does 80 percent of the job and you can live with the other 20 percent, buy it. If staff time is going into fighting the product, or into Excel beside it, you are past that point. Do not build Tally. Do not buy a “flexible platform” that needs a year of customisation you will not own.
Buy these without a meeting
GST-compliant accounts, email, payroll, card payments, a generic CRM if your pipeline is simple. Building those from scratch is how custom software got a bad name. Pay the subscription. Integrate. Spend attention on the job that is actually yours.
A storefront that is really a catalogue and a checkout is usually a product too — Shopify, WooCommerce, or a small custom front on top of a payment gateway. You are not a payments company. Do not become one by accident.
Build these, or keep paying the workaround
Job cards that do not match any package’s “project” module. Branch stock that is almost a separate company. Dealer portals. Billing that mixes retainers, parts and call-outs. The spreadsheet that already runs the warehouse. Those are the middle. A focused MVP of that middle, talking to Tally, is usually cheaper over three years than another round of package customisation you will not own.
The test is rude and useful: if you stopped paying the vendor tomorrow, could you still change a rate and export your data? If the answer is no, and the process is why customers stay, you are leasing your operation. Sometimes that is fine. Write it down. Often it is how you fund a rewrite in year four.
The hybrid most firms should pick
Tally or a SaaS ledger stays. The public website stays. You build the portal, the dispatch list, the thing customers feel. That is not indecision. It is refusing to rebuild commodity software and refusing to pretend a theme can do roles and audit trails.
Hybrid fails when nobody names the system of record. Two places to enter an order is how you spend the next year reconciling. Pick one. Integrate the other. The comparison table and the five-year sheet are on custom software versus off-the-shelf.
Signs you are already past “just buy a product”
- Staff keep a shadow spreadsheet “because the system doesn’t do it”.
- The last partner quoted a unique module list that looks like a second product.
- You have changed how you invoice to match the package, and customers noticed.
- Two branches run the same product two different ways, and head office lives in Excel.
Any two of those and you should price a first slice, not another licence. How we estimate that slice is in how much custom software costs. Who should build it is in how to choose a software development company.
A five-minute test you can run this week
List the jobs that make you money. Circle the ones a stranger could buy a product for tomorrow. Those are buy. Circle the ones that live in a spreadsheet because “the software doesn’t do it”. Those are build, or they are a product you have not found — spend a day looking before you commission. What remains is hybrid: product for the circle, custom for the spreadsheet, one system of record named in writing.
If everything is circled “build”, you are about to rebuild Tally. If nothing is circled “build” and staff still live in Excel, you have not been honest about the spreadsheet. Do the test with the person who types, not only with the person who signs.
A week of deciding, not a year of arguing
Monday: list the jobs. Tuesday: sit with the person who types and mark buy / build / already-a-spreadsheet. Wednesday: write one sentence for the first live slice if anything was “build”. Thursday: send that sentence and the one-page brief to one product vendor and one custom shop — same brief. Friday: put the two replies in a table. You will know more than a quarter of workshops usually produces.
If both replies are slides, your brief was vague. Tighten it. If the product vendor can show the job working in a sandbox and staff say they can live with the 20 percent, buy. If they show a module that almost works and staff already describe the workbook they would keep, price the slice. That is the whole decision. The longer comparison is custom versus off-the-shelf when you need the table in a meeting.
Do not rebuild accounts, email, or payments in that week. Those are products. Do not buy a “flexible platform” that needs a year of work you will not own. If you cannot name the slice by Friday, buy discovery, not software. That is the useful version of this question. The rest is a longer meeting you can schedule after the table exists.
A jobbing shop versus a trader — same question, two answers
The trader with standard GST, standard stock and a willingness to change a form should buy a product and integrate Tally or the bank. Custom there is vanity. The jobbing shop with its own bill of materials, hold-and-release rules, and customers who buy the exception should not spend another year forcing that into a “manufacturing app”. Price an MVP of jobs and stock for one bay, books still in Tally. That is build. The public website is still buy.
If you are both — trade at the counter, jobbing in the back — name two systems of record or you will reconcile forever. Counter sales can live in a product. The bay cannot. Hybrid is the grown-up answer, not a compromise. Who builds the bay is then how you choose a company, not whether to buy Outlook.
What “buy discovery” looks like in practice
A paid week: sit with the people who type, photograph the workbooks, write the first-slice sentence, list the systems that must stay, and return a scope you can argue with. That is cheaper than a platform login you will not open. It is also how you find out you should have bought a product. Either outcome is a result. A free “we will study your process” that ends in a 12-module quote is a sales cycle, not discovery.
If after that week you still cannot name the slice, you are not ready to spend on software. You are ready to pick a system of record and an owner. Do that. Then come back. Building into an argument is how custom projects get a bad name they did not earn. Discovery that ends in “we need to think” is still cheaper than a build that ends in the same sentence in month four. Pay for the week. Do not pay build rates for the think. If they will not sell the week alone, they are not doing discovery.