Investors
A differentshape of economics.
Villas and restaurants are good businesses. The difference is not whether they are good, it is what you have to buy every time you want more revenue. This page opens that comparison, assumptions included.
Every figure on this page is a model, not survey data and not our own results.
The booksare the product.
The short version
Physical assets grow on capital.Software grows on distribution.
A villa adds revenue by buying another room. A restaurant adds revenue by opening another outlet. Both need large capital up front, and that capital is hard to pull back if the bet was wrong.
Software adds revenue by finding the next customer. That spend can be stopped next month if the numbers look bad. This does not make it easier, it makes the risk a different shape.
01 · How we think
A system of record,not a helper app.
Three things taken from how Palantir thinks about data and how Anduril thinks about owning the product, then fitted to the size of a corner shop.
Whoever holds the official record does not get replaced
A helper app is easy to drop. A system holding a business's numbered receipts, stock, debts, payroll, and tax filings becomes its official record. Leaving it means leaving your own history behind.
One map per business
Products, customers, suppliers, per-customer prices, IOUs, staff, and assets are not stored as separate tables but as one connected map. A question like which product is quietly losing money can only be answered when the map connects.
We sit with the people using it
Features come from sitting in someone's shop, watching them mistype, and fixing it. Not from a meeting. Every correction they make comes back as data for measuring which model actually understands how they talk.
02 · Capital
Capital to addRp100M a month.
What has to be spent before that extra revenue exists at all.
- Four-room villa / Forty-seat restaurant
- Subscription software
+Show the numbers
| Business | Capital | Shape of it |
|---|---|---|
| Four-room villa | Rp2.5 billion | One more villa, land plus build |
| Forty-seat restaurant | Rp800 million | One more outlet, lease plus fit-out plus equipment |
| Subscription software | Rp253 million | 505 Pro customers, through acquisition spend |
What the chart does not show: the first two bars buy assets that stay bought. The third is acquisition spend that can be stopped next month, and if it is stopped, the customers already won keep paying.
03 · Cost structure
Where every millionin revenue goes.
The same four buckets for all three, so the columns can actually be stacked against each other.
- Cost to serve
- Fixed operations
- Winning customers
- Left for the owner
+Show the numbers
| Bucket | Villa | Restaurant | Software |
|---|---|---|---|
| Cost to serve | 22% | 40% | 20% |
| Fixed operations | 28% | 40% | 40% |
| Winning customers | 18% | 10% | 25% |
| Left for the owner | 32% | 10% | 15% |
The restaurant looks heaviest because ingredients and shift labour rise with every additional guest. The villa has far more room, and gets more still when the OTA commission disappears. That is exactly the job Dewata Tech does.
04 · The ceiling
One of themhits a wall.
Demand doubles. One line follows it, the other stops at the number of rooms.
+Show the numbers
| Demand | Physical asset | Software |
|---|---|---|
| 0 | 0 | 0 |
| 50 | 50 | 50 |
| 100 | 100 | 100 |
| 150 | 100 | 150 |
| 200 | 100 | 200 |
A full villa cannot take a fifth guest in a fourth room, whatever that guest is willing to pay. Getting past the line means buying another villa, which returns you to the first chart. Software has no wall of that kind. Its wall is distribution, and a distribution wall can be pushed without buying land.
05 · The fair part
Where villas andrestaurants win.
Without this section the three above it do not deserve to be believed.
The demand already exists
A villa in Bali does not have to convince anyone that staying somewhere is necessary. Software has to convince people that bookkeeping is necessary, which is far harder work.
The asset holds value
Land and buildings are worth something even if the business fails. Failed software is worth nothing unless the data gets used.
Cash arrives in month one
Open the villa, guests pay. Subscription software takes time before customer count covers the fixed cost.
Banks understand the collateral
A bank knows what to do with property. A bank does not necessarily know what to do with recurring subscription revenue.
Assumptions
Where the numbers come from.
These are illustrative models, not surveys. Change an assumption and the numbers change with it.
Four-room villa in Bali
- Average rate Rp1,500,000 a night at 60 percent occupancy
- Roughly Rp108 million of monthly revenue
- Average OTA commission of 18 percent on rooms booked through an OTA
- Rp2.5 billion to build one more villa, land included
Forty-seat restaurant
- Two table turns a day at 60 percent occupancy
- Average spend of Rp85,000 per guest
- Ingredients 33 percent, labour 25 percent, rent 12 percent
- Rp800 million for one more outlet, including lease deposit and equipment
Subscription software
- Public pricing: Pro at Rp198,000 a month, Team at Rp600,000 a month
- Blended revenue of Rp250,000 per customer per month after plan mix and annual discount
- Rp500,000 to acquire one customer
- Roughly 20 percent of the price to serve one additional customer
What is deliberately not on this page: AI cost per message, the margin floor, and today's paying customer count. Those are internal, and we bring them to the meeting.
Legal note
- This page is company information. It is not an investment offer and not a solicitation to buy securities.
- We make no public offering of securities. Public offerings in Indonesia are governed by capital markets law and supervised by OJK, and we are not registered for one.
- The figures here are projections built on assumptions printed in the open, not a promise of results.
Want the real numbers instead of a model?
Measured traction, acquisition cost, and retention we open up in the conversation.