VELA Dhi Nakhon Phanom Value tracker
Proof of concept · hotel setup is real, rates and bookings simulated
Value tracker · VELA Dhi Nakhon Phanom

Is the engine earning its keep?

Four value lanes, each against a baseline finance froze on 20 October 2026 and with one owner. One AI cost, counted once. Claimed is what the engine says it earns on this year's calendar; bankable is the share finance will sign before the pilot's 8 control weeks are read. Full-year run-rate, simulated.

All the rules, in the Strategy studio
Net value · a year, after AI cost ฿3.35M bankable ฿937K
ROI · contribution ÷ year-1 AI cost 4.9× bankable 2.1×
Cash payback · months of value to repay year 1 2.4 mo bankable 5.7 mo
AI cost · counted once, in full ฿856K year 1 ฿450K build + ฿406K/yr run · ฿406K/yr from year 2

Claimed vs bankable, by lane

BankableClaimed, not yet bankedAI cost
Revenue
฿335K of ฿669K
Cost saved
฿212K of ฿423K
Loss prevented
฿1.25M of ฿3.12M
AI cost, year 1
฿856K

Hours in the cost-avoided lane are not on this chart: hours are never money.

Revenue Revenue manager
฿669K
claimed contribution a year· ฿992K revenue
Bankable ฿335K × 0.5

Room revenue above the static sheet on the 333 nights the sheet would not have sold out, each night replayed over its 120-day booking window. Counted after OTA commission and the THB 380 cost of each room sold.

Room revenue upliftbefore commission and room cost
฿992K
Rooms sold vs the sheetroom nights; fewer rooms at a higher rate cost less to service
+596
Frozen baseline20 Oct 2026
Static rate sheet, next 365 nights: THB 56.18M room revenue at 67.3% occupancy
Source: PMS night audit + 2026 rate sheet

Why × 0.5. The deck's conservative case is about half the base case. Finance banks that half until 8 control weeks show a RevPAR index of 103 or better.

Cost saved F&B manager
฿423K
claimed contribution a year
Bankable ฿212K × 0.5

OTA commission the hotel stops paying as website packages pull room nights direct: the blended rate falls from 9.7% to 9.0% of room revenue. The F&B manager owns the packages that do the pulling.

Engine room revenuethe base the commission applies to
฿60.44M
Commission avoided0.7 points of room revenue
฿423K
Frozen baseline20 Oct 2026
Blended OTA commission 9.7% of room revenue; 60% of room nights via OTAs
Source: Channel manager production report

Why × 0.5. Packages reach the website in phase 3. The channel manager reports direct share monthly, so this lane is measured, not modelled, within a quarter of go-live.

Cost avoided Finance
166 h
hours a year, claimed · not converted to money
Bankable THB 0 × 0.0

Hours the revenue manager no longer spends typing rates, less the morning brief and the approvals the rules still ask for. Shown as hours. Not converted to money: nobody leaves the payroll.

Typing rates and comp checks today
312 h
Morning brief with the engine
−91 h
Approvals (3 a day × 3 min)
−55 h
Frozen baseline20 Oct 2026
Revenue manager: 6 h a week typing rates into extranets and checking competitors by hand; 3 hand-offs per change
Source: RM time study + channel-manager change log

Why × 0.0. Finance counts the hours and keeps them out of the THB column until a cost line actually leaves the P&L. Hours saved are a capacity gain, not cash.

Loss prevented General manager
฿3.12M
claimed contribution a year· ฿3.36M revenue
Bankable ฿1.25M × 0.4

Event nights the static sheet would have sold out weeks early, now priced over the whole window; plus the displaced-guest value the hotel would have given away by accepting groups blind. The GM signs A/A+ nights and every group on a peak date.

Sell-out nights (32)contribution over the static sheet
฿3.03M
Groups turned away or countered (4 of 17)displaced-guest value kept
฿87K
Frozen baseline20 Oct 2026
32 nights the static sheet sells out before arrival, on average 28 days early; 17 group requests a year quoted without a displacement check
Source: PMS sell-out dates + sales group log

Why × 0.4. A loss that did not happen is the hardest claim to prove. Finance banks it only where last year’s PMS shows the same night sold out early, and only after the control weeks.

Move one boundary

These are the lines the GM and revenue manager sign. Move one and the engine re-prices all 365 nights; every lane, the net value and the ROI follow. Changes apply on every screen until you reset.

Auto-publish on A nights
The deck signs: both wait. Only the GM can widen this.
LaneSignedNowMoved by
Revenue฿669K฿669K±0
Cost saved฿423K฿423K±0
Cost avoided166 h166 h±0
Loss prevented฿3.12M฿3.12M±0
Net value, claimed฿3.35M฿3.35M±0
ROI, claimed4.9×4.9×±0

Right now 3 price moves are waiting for the revenue manager, worth ฿66 of contribution if approved tomorrow. The engine has usually finished moving a peak night long before 7 days out, so the window costs clicks, not revenue: a shorter one gives the revenue manager hours back and changes the money by about nothing. The band is where the money moves.

Finance's question, answered

"Which of this can I actually bank, and when?"

What counts as bankable
Only money that beats the pilot's control weeks: the same room types, the same weeks of the year, one priced by the engine and one by the old sheet, read from the PMS night audit and the channel manager, never from the engine's own forecast. Until those 8 weeks are read, the tracker banks a fixed share of each claim: ×0.5 on revenue and cost saved, ×0.4 on loss prevented, ×0 on hours.
When
Phase 3 of the roadmap: auto-publish on C/D dates with control weeks, gate RevPAR index ≥ 103 against control. A lane moves from claimed to bankable the month finance signs its control-week read-out, and the claimed figure is retired at the same time. Nothing is bankable in shadow mode.
What never counts
Hours saved (166 h a year here) stay in hours until a cost line leaves the P&L. A dashboard of steps taken is not value. Revenue the engine would have earned on a night the sheet also filled is not loss prevented; only nights the sheet sells out 2 weeks early qualify.
Why one AI cost
฿450K to build and ฿406K a year to run, charged in full against year one (฿856K) and never spread across the lanes. ROI is contribution ÷ that cost; payback is the months of contribution needed to repay it.
What is not in here yet
The deck's package margin (breakfast, Mekong dinner, festival nights) is claimed on the slide but not in this tracker: the engine ranks packages but does not yet model attach rates, so it waits for POS data. Everything on this screen is synthetic until the hotel's exports arrive.

Baselines frozen 20 October 2026 from the hotels' night-audit reports. Blended commission 9.7%; variable cost THB 380 per room sold.