Build, Buy, or SaaS? Working Out When a Custom App Makes Sense

One seat of HubSpot Sales Hub Professional costs USD 90 a month on annual billing (Source: HubSpot, 2026). One user on Odoo's Standard plan costs Rp 136,000 a month (Source: Odoo, 2026). At roughly Rp 17,900 to the US dollar on 6 October 2026, the first seat runs to about Rp 58 million over three years, while the second stays under Rp 5 million (Source: Bank Indonesia, 2026). That gap of more than ten times is what decides when a custom app is worth costing out, and when to drop the idea.

A custom app makes financial sense when three years of per-user subscriptions add up to more than the build price plus maintenance. Against per-seat software priced in dollars, that point can arrive with a handful of users. Against software priced in rupiah, it takes dozens. Against standard accounting such as Accurate Online, it practically never arrives.

Three options, three ways the bill grows

Build versus buy is usually framed as two camps. Look at the invoices and there are three patterns. Buying off the shelf means local software billed per company or per outlet: technically you are still subscribing, but the bill barely moves as you hire. Per-seat SaaS works the other way, because every new employee adds to the bill for as long as they stay. Some platforms, Odoo among them, also sell implementation and customisation separately from the subscription (Source: Odoo, 2026). Building your own is paid for up front and then maintained each year, and the number of users does not change the figure.

The second pattern is still the most common. One compilation covering more than 1,000 B2B SaaS companies found about 58% of products still priced per seat, with a median list-price increase of 7.8% a year (Source: Stripo Research, 2026). The data is not Indonesian. The direction still applies: a per-seat bill grows from both sides at once, the number of people and the price per person.

How many users before a custom app pays for itself?

What follows is an illustration, not a quote. The assumptions: a 36-month period; regular list prices on annual billing as of 6 October 2026; an exchange rate of Rp 17,900 to the US dollar (Bank Indonesia's mid transaction rate that day, rounded); no subscription price rises; and VAT (PPN) left out, since both sides carry the same effective 11% rate (Source: Bank Indonesia, 2026; Enforce Advisory, 2026).

Start with the custom side. App maintenance is commonly budgeted at 15–25% of the build cost per year; that is a planning convention among agencies, not a research finding (Source: Edana, 2026). Take the midpoint, 20%. A Rp 150 million app means Rp 30 million a year in maintenance, for a three-year total of Rp 240 million.

Now the subscription side. One Odoo Standard user: Rp 136,000 times 36 months comes to Rp 4,896,000. Divide Rp 240 million by that and you get about 50 users. Below 50 people, subscribing is cheaper. Above that, the gap widens in favour of custom with every hire. The same method, applied to three build prices:

• A Rp 50 million build (three-year total Rp 80 million): breaks even at 17 Odoo Standard users, or 2 HubSpot Sales Hub Professional seats.
• A Rp 150 million build (total Rp 240 million): breaks even at 50 Odoo Standard users, 5 HubSpot seats, or 4 seats of Salesforce Pro Suite at USD 100 per user per month (Source: Salesforce, 2026).
• A Rp 200 million build (total Rp 320 million): breaks even at 66 Odoo Standard users, 6 HubSpot seats, or 5 Salesforce Pro Suite seats.

The Rp 50 million and Rp 200 million figures come from the range one local vendor publishes, for a simple internal app and for the lower bound of a complex operational system respectively (Source: Next IT, 2026). At Agentiq, custom apps start from Rp 15 million (Source: Agentiq, 2026). That is a floor price for a narrow app handling a single workflow, not a replacement for a full ERP or CRM. The only honest claim is this: a small app for the one workflow a team really uses can cost less than the row of seats bought just to get that workflow.

When does the maths never favour building?

Standard accounting is the clearest case. Accurate Online charges Rp 333,000 a month for its base package and Rp 22,200 for each additional user, so 10 users cost Rp 532,800 a month, or Rp 19,180,800 over three years, VAT included. For 25 users, Rp 31,168,800 (Source: Accurate, 2026). No accounting system worth using can be built and maintained for that money. Buy it and move on.

Point of sale is much the same. The Prime plan from majoo costs Rp 999,000 a month for one outlet, about Rp 36 million over three years before VAT, and that figure stays put however many cashiers you have (Source: majoo, 2026). For software of this kind, the case for building has to come from process fit and the manual work it removes. Licence fees will never be the reason.

What this means for your business

Treat those numbers as a rough zone. A licence break-even does not count your staff's time spent specifying requirements and testing the result, the months before the app is usable, or the risk that the process changes midway. List prices are also before negotiation, and first-year promotions make the subscription side lighter than it appears here. The dollar figures also use mid-range tiers. On Salesforce Starter Suite, at USD 25 per user per month (about Rp 16.1 million a seat over three years), the same Rp 150 million build breaks even at about 15 seats rather than 4 (Source: Salesforce, 2026). The sum also only holds if the build covers the workflow you actually use that software for.

The subscription side, in turn, has items that rarely make it into the sum. HubSpot Sales Hub Professional carries a USD 1,500 onboarding fee, paid once up front, roughly Rp 26.9 million (Source: HubSpot, 2026). Every Rp 1,000 move in the exchange rate shifts the cost of one USD 90 seat by Rp 3.24 million over three years, and a business earning in rupiah carries that risk for as long as it subscribes. Price labels are not consistent either: Accurate quotes its prices with VAT included, majoo without.

There is one more item, and it is usually the largest. Jakarta's 2026 provincial minimum wage (UMP) is Rp 5,729,876 a month, which puts one admin's base salary at about Rp 206 million over three years, before the holiday allowance (THR) and BPJS contributions (Source: Katadata Databoks, 2025). In our view, this line should weigh more heavily on the decision to build than the software bill does. When a workflow does not fit generic software, the gap is almost always plugged by a person: someone who spends each day copying orders from WhatsApp and the Shopee dashboard into a spreadsheet, then keying them into the system all over again.

Five steps before you decide

• List every software subscription you pay for: the price per seat, the currency, and whether VAT is included.
• Count the seats actually used for one core workflow, then multiply by 36 months.
• Ask for a build quote for that workflow alone, add three years of maintenance, and put the two figures side by side.
• Add up the admin hours spent copying data between apps, then convert them to rupiah.
• Leave standard accounting and point of sale as off-the-shelf purchases.

If you would like to see what this sum looks like for your business, tell us about your process through our contact page — the Agentiq team will reply with a concrete recommendation, not a template, including when the answer is to keep subscribing.