Thailand's fuel price formula has just had assumed costs removed, and the same pattern sits in IT budgets. This article covers how to check lines that rest on assumptions, an assumption register, contracts that adjust to actuals, and when not to cut a reserve.
On 28 October 2026, Thailand's Minister of Energy signed an order removing three cost items from the formula used to calculate refinery-gate fuel prices: freight, insurance and transit loss. According to the ministry, these were calculated on the assumption that the fuel was shipped from Singapore, and they had been in the formula since 2004. The order takes effect on 29 October 2026.
We are not energy experts, and this article does not judge whether the policy is right or wrong. What makes the story interesting for people who manage IT budgets and contracts is the shape of the problem. A cost line rests on an assumption set on day one, sits inside a formula everyone keeps using, and for a long time nobody compares it with reality. The same thing happens in IT budgets. We will describe where, how to check, and when removing an assumed cost becomes dangerous.
What happened to the fuel price formula, and how far we have read
According to Thai PBS and The Standard, reporting on 28 October 2026 and citing the ministry's statement, the order cancels the freight, insurance and transit loss charges calculated on a hypothetical route, cuts the oil reserve allowance to two thirds of its former level to match actual reserve volumes, and is estimated to lower refinery-gate prices by 0.0875 to 0.3548 baht per litre, with B7 diesel falling the most, by 0.3548 baht per litre. The refinery-gate price, as EPPO defines it, is the price of finished fuel before taxes, funds and the marketing margin.
We should be plain that these figures are the ministry's own words, repeated by the press. We have not seen the order or the Energy Policy Administration Committee's resolution as official documents. The coverage we read does not explain how the 22 years is counted and gives no share for how much fuel is refined domestically. The reduction is at the refinery gate, not the pump. The Standard says the ministry will discuss with the Oil Fund how to pass it on to retail prices, with no timeline, and we cannot confirm how far retail prices will fall. A representative of the Thailand Consumers Council also proposed that other items deserve scrutiny, which is that person's opinion (Thailand Consumers Council). So we do not use this article to judge energy policy. We use the facts above only as an example of a pattern.
What assumed costs look like in an IT budget
In IT budgets and contracts generally, some lines do not come from receipts. They come from assumptions set when the budget was first drawn up. The examples below are common ones, and they are our own summary, not statistics.
User counts or licenses set from the estimate on the day of approval, still paid at that number although actual use has changed
Reserved cloud capacity sized to an expected peak that was never compared with the peak that actually occurred
Project contingency set as a fixed proportion of contract value, not tied to the real risk of that work
Annual maintenance fees charged as a proportion of the software price, not from work the provider actually does
Storage or backup space calculated from a growth rate assumed on the design date
None of these is wrong by default, and many had good reasons on the day they were set. The problem arises when nobody goes back to ask whether the assumption is still true, and the line becomes a cost paid again every year. We wrote about the gap between the figure proposed at approval and the actual bill in our article on why cloud savings fall short. This article looks at a different group: lines that never reflected anything real from the start but are still being charged.
Why assumed costs last longer than you would think
As far as we can see, assumed costs persist for four reasons.
There was a reason when it was set. An assumption tied to the market or standard of the day may have been correct. We have not checked the history of the fuel formula, so we do not say it was unreasonable when set.
It sits inside a formula nobody opens. The final figure was approved, and people see only the total, not the components.
Nobody owns the question of whether it is still true. Everyone is responsible for using the number, and nobody for reviewing it.
The neighboring numbers are consistent. The total is within budget and quotes can be compared, so nothing prompts anyone to open the formula.
In our view the third reason matters most, because once someone owns the question, the other three can mostly be fixed.
How to check: an assumption register
The method we recommend is a one-page assumption register. Go through every line of the budget or contract that is a formula rather than a receipt, and for each line record six things.
The line name and the amount paid per year
The assumption behind it, such as user count, peak or growth rate
Where the assumption came from, who set it and when
The latest measured actual, and what it was measured from
The gap between the assumption and the actual
The owner who must answer, and the date of the next review
The fourth field is the one many organizations cannot fill, because nobody has ever measured it. That is useful information in itself: it means the line is still a pure assumption with no real figure to confirm it. Reserved cloud capacity is an easy one to check, because monitoring already records actual usage. How to plan capacity from the actual peak is in our article on capacity planning.
An example: checking one line
Take the line for an enterprise software license paid annually by number of users. The figure on the quote is the number of users estimated at purchase. We check in this order.
Open the quote and the contract, and find what the billed user count actually is: accounts created, users who actually log in, or a number agreed in advance.
Ask for a usage report from the provider's system, or pull one from your own identity system, to get the number of users who actually used the software in the latest period.
Compare the two numbers. If they are close, the assumption still holds. If they are far apart, check whether the cause is accounts without owners, staff who have left, or genuinely reduced use.
Take the result to the provider before the renewal round, with the real figures and how you obtained them in hand.
One caution is that a gap does not always mean the provider is charging wrongly. Some contracts set a minimum deliberately, and users who rarely log in may still need entitlements for periodic work. What this check gives you is information that gives the negotiation a basis. It does not mean the amount will always come down.
Warning signs that an assumed cost is hiding
When there is no time to go through every line, we use these signs to pick which lines to check first.
A suspiciously round number that has not changed across several contract rounds, even though the business has
Nobody on the team can explain where the number came from beyond "it is standard" or "we have always done it this way"
The vendor cannot show the formula or the source of the number, but the total is presented as one lump
The line is calculated as a proportion of other lines, not from measurable workload or usage
The monitoring your organization already has has never been used to compare against the line
These signs do not prove a line is wrong. They only say it has never been checked, and checking takes little effort compared with paying again every year for several more contract rounds.
Writing contracts that can adjust to actuals
Once you know which lines are assumptions, the next question is whether the contract lets you adjust them. We see three things worth putting in a contract. They are our suggestions, not legal requirements.
Adjust to actuals on a cycle, for example reviewing user counts and capacity each contract period and adjusting the amount to what is actually used
A right to inspect, so the organization can ask to see where the figures on a quote come from, not only the total
A written review date, rather than a verbal promise
We wrote about conditions like these in our article on conditions to put in a TOR, and about reading annual maintenance contracts in our article on MA and SLA contracts.
When an assumed cost is removed, do not cut everything
One observation from the fuel formula case. According to the reports, the oil reserve allowance was reduced to two thirds to match actual volumes, not cut to zero. We see that as a principle that applies to IT budgets: adjust to what is real, do not delete every line that is an assumption.
Some lines that look like assumptions are working as real buffers. Spare capacity held for a peak that has not yet occurred, or a project reserve for risk, are examples. If you cut them to the average use on the day you measured, the system may fail on the day load rises. The test we use is that a good buffer has a reason you can write down for what risk it protects against, and someone who knows when it was actually drawn on. An assumed cost without that reason is the line to question.
The limits we should state plainly
This comparison has limits. A fuel price formula is a state price-setting mechanism with stakeholders and rules very different from an organization's budget. We are not saying the two situations are the same, and we are not judging that the old formula was unreasonable from the start.
The other limit is that if your organization has few systems and few contracts, a full assumption register may be more than you need. One sheet of paper that says what assumptions the three largest sums rest on, and who checks them, gives almost the same benefit, because a register larger than the job becomes a document nobody updates.
An objection we often hear
The objection we hear most is that these numbers were already approved. It is a fair question, and approval is something you should have. But approval answers whether the number was reasonable on that day. It does not answer whether it is still reasonable today. The two questions need different owners on different dates.
Another is that the vendor prices at market standard anyway. That may be true, and asking where a number comes from is not an accusation. A vendor pricing in good faith can say what assumption a number rests on. A vendor who cannot is also information the organization should have before signing.
Where to start
Pick the three largest lines in the IT budget or in recurring contracts.
Ask what assumption each rests on, who set it and when.
Find the actuals to compare, from monitoring, provider invoices or usage reports.
Name an owner and a review date for each line, as a person and a date, not a department.
Add adjust-to-actuals terms and an inspection right to the next contract that renews.
Summary
If you take only one thing from this article, pick the single largest line in your IT budget and ask what assumption the figure rests on, and when anyone last compared it with the latest actual. If nobody can answer, that line is where the check starts.
If your organization is still scoping the work and would like someone to talk through the problem first, you can reach Superdev at 088-983-9386 or [email protected]. Our office is in Bangkapi, Bangkok. There is no deadline, you do not need complete documents to start, and if it turns out the timing is not right to build a new system, we will say so plainly, because starting a project at the wrong time costs more than waiting one more quarter.
FAQ: Frequently Asked Questions about This Article
A collection of questions and answers to help you better understand the content of this article.
Tags: