Custom or off the shelf: deciding it properly

Almost every build versus buy article reaches the same conclusion. Off the shelf is cheaper and faster, custom fits better, choose according to budget. It is true, it is useless, and it is usually written about a market that files different taxes than you do.

The decision is more decidable than that, and it does not start with price.

Start with one question

Is the process you are trying to systematise the thing you compete on?

If you run a distribution business and your advantage is that you get orders out faster than anyone in your region, then the way you pick, pack and dispatch is not overhead. It is the business. Software that forces you into a generic version of that workflow does not save you money. It removes the reason customers buy from you.

If you are trying to systematise payroll, or bookkeeping, or email, then no. Nobody wins on payroll. Buy it, configure it, and spend nothing further thinking about it.

Most operations are a mix, and that mix is the real answer: buy the commodity parts, build the part you actually compete on, and make them talk to each other.

The Philippine specifics that generic advice skips

Four things routinely decide this question locally, and none appear in advice written elsewhere.

Statutory obligations are not a configuration option. BIR filing formats and official receipt requirements, SSS, PhilHealth and Pag-IBIG contributions, the documentation an LGU or DTI process expects. A global platform may let you attach these as custom fields. It will not produce what the agency actually wants, and the gap gets filled by somebody re-keying data into a separate file every month. That monthly file is a permanent cost, and it belongs in the comparison.

Local payment rails. GCash, Maya, over the counter, bank transfer, cheque, and terms extended to long-standing customers on trust. International platforms model card and direct debit well and the rest poorly. If reconciliation is where your month-end disappears, this is why.

Branches, and the day they diverge. Multi-branch operations do not fail on features. They fail when two branches record the same event differently and nobody notices for three weeks. Off the shelf software enforces one way of recording. Whether that is a benefit or a disaster depends on whether your branches can actually be made to work identically, which is an operational question and not a software one.

Connectivity and the people using it. A warehouse with unreliable signal, staff working from phones rather than desktops, a supervisor who will use Viber and will not use a web portal. These constrain what will actually get used, and a tool that is not used is not cheaper. It is just unused.

The comparison people get wrong

Off the shelf looks cheaper because the licence fee is visible and the workarounds are not.

Count the workarounds. The export that gets massaged in Excel every week. The second system kept because the first one cannot do one specific thing. The person whose job is partly to move data between two tools. The report that three people build separately because the standard one is close but wrong. These are real recurring costs, usually paid in staff time, and they rarely make it onto the comparison sheet because no invoice arrives for them.

Then count what the subscription actually gets used for. Broad research on software spending repeatedly finds that a large share of purchased licences sit idle and a large share of features go untouched. You are not buying a feature list. You are buying the handful of features your team opens, and the rest is paying for someone else's roadmap.

That does not mean build. Often the honest total still favours buying, and a provider worth hiring will tell you so. It means the comparison has to include the workarounds, or it is not a comparison.

A decision procedure

Work through it in this order. Stop at the first clear answer.

1. Is this a commodity process? Payroll, accounting, email, storage. Buy it. Stop.
2. Does a credible local product already handle your statutory and payment reality? If yes, buy it, and accept the process changes it imposes. Stop.
3. Can your process be changed to fit a global product without losing what you compete on? If yes, buy it and change the process deliberately, as a project with an owner. Stop.
4. Is the misfit concentrated in one or two workflows? Then buy the commodity layer and build only the part that misfits, connected to it. This is the answer more often than either extreme.
5. Is the whole operation the thing you compete on? Build, and scope it against a study of how the work actually moves.

Step four is the one most people skip. The choice is presented as all or nothing, and it almost never is.

What both roads have in common

Whichever way it goes, you need to know how the work actually moves before you can judge fit. Buying without that produces a subscription plus a spreadsheet. Building without it produces an expensive version of the same confusion.

That study is worth doing on its own terms, and it is worth doing before anyone quotes you a price for anything.