An initiative's Overview shows two cards, and they answer two different questions.

The first card is a forecast. It never changes unless someone edits the project's inputs. The second card is a measurement. It grows each month as more real data arrives.

Important exception: two rows in the second card — Simple payback and Discounted payback — do not follow the "to date" rule. Both sides of those two rows cover the whole project life. Section 8 explains why, and the reason is a good one, but it is an exception to the card's heading.

Every formula in this article is worked through with the same project example, so you can follow one set of numbers from beginning to end. These are the project's recorded inputs.

Jargon — baseline: the baseline is the "before" picture — what the site was emitting or spending before the project changed anything. Every saving is measured as the difference from this. If the baseline is wrong, every saving figure is wrong, which is why section 10 flags incomplete baselines as a risk.

Jargon — capital cost (capex): capital cost, often shortened to capex, is the one-off money spent up front to build or install something. It is separate from the ongoing running costs. In this model the entire capital cost is treated as being spent in the project's first month.

Jargon — useful life: the useful life is how many years the equipment is expected to keep working and producing savings. We do not count any savings after the useful life ends. If a project has not repaid its cost by then, we report that it does not pay back rather than assuming it will continue.

Almost every mistake in this kind of calculation comes from being unclear about when something starts. There are two separate things to understand about timing: when things happen, and how much a dollar is worth at each point on that same timeline.

When things happen. For the worked example:

The build period produces no savings by design, and that is why payback is measured from the day the money leaves the bank rather than from the day savings begin — the two years spent building are two years of waiting to be repaid.

What a dollar is worth at each point. Money arriving later is worth less to us than money arriving now, so before adding anything up we reduce future amounts. This is called discounting, and section 5 explains it fully.

The reduction does not begin immediately. There is a cushion of twelve months at the start, during which every dollar counts in full. Only after that does the reduction begin, and it grows the further into the future you go. In the worked example (a 12% discount rate, valued in January-2024 money): every dollar from January to December 2024 counts in full; the reduction starts in January 2025; $1 saved in January 2026 (the first savings month) counts as $0.88; $1 saved in December 2028 (the last savings month) counts as $0.64.

What this means for this particular project: the cushion covers January to December 2024, and this project does not start saving anything until January 2026 — well past it. So every dollar this project ever saves is reduced before it is counted — the cushion gives it no benefit at all. That is a direct consequence of spending two years building. A project that started saving within its first year would have some of its savings counted in full.

These four rules decide which months are counted. Each one is stated as precisely as possible, because being off by a single month changes the answer.

The completion date is 31 December 2025. Savings are counted from 1 January 2026, not from December 2025.

Why: December 2025 is a month in which construction was still finishing. Counting it would credit the project with a month of savings it could not have produced.

"Today" is August 2026. The measurement window ends on 31 July 2026. August is excluded.

Why: a month that is only part-way through has only part of its bills in. Including it would make every figure look worse than reality, and would make the whole table drift downward through the month and jump back up on the 1st. Excluding it keeps the table stable and honest.

This rule applies to the payback figures. If the running total stands at −$8,000 at the end of one month and +$3,000 at the end of the next, the project broke even during that second month. That second month is included in the count.

Why: the alternative would report the project as not yet paid back in a month where it demonstrably was. We do not estimate a part-month; we count whole months only, matching the reference spreadsheet this model was built against.

Rule 3 with this project's real numbers:

Simple payback is therefore 46 months, or 3 years 10 months. The project did not become profitable at the end of month 46 — it crossed zero at some point during it — but we report whole months, so month 46 is the answer.

The benefit period ends December 2028. If a project has not repaid its cost by then, we display "Does not pay back" rather than projecting further savings.

Why: the equipment's expected working life is the honest limit of what we can claim. Extending past it would be inventing savings.

What "Does not pay back" means precisely: it means the cost is not recovered within the equipment's useful life. It does not mean the project can never pay back. If the equipment lasts longer than expected, or savings improve, the picture changes. The phrase is scoped to the modelled life and nothing beyond it.

Jargon — discounting: discounting is the practice of treating money received in the future as worth less than money in hand today. If you have $100 now, you could invest it; $100 arriving in three years cannot have been invested during those three years, so it is worth less to you now. Discounting converts future amounts into today's equivalent so they can be fairly added together.

The discount rate is taken from the first of these that has been set:

Where 12% comes from: It is a fallback, not a recommendation. Any company that has a formal cost of capital should set its own rate, and any initiative with a different risk profile should set one on the initiative itself.

Jargon — valuation anchor: the anchor is the month whose money everything is expressed in. Saying "$243,010" is incomplete without saying which year's dollars. The anchor answers that.

On the initiative detail page, the anchor is the project's own start month — January 2024 in this example. This is shown on the card as "Valued at Jan 2024".

Why the start month: this page is about one project's own economics, and the question being asked is "was this a good investment decision?". That decision was made at the start, so the start month is the fair reference point. It also means the number does not move: opening this page today and again in six months shows the same figure, which makes it usable as a record.

The Scenario Planner uses a different anchor — on purpose. The Scenario Planner values everything in today's money instead. That is because it compares and adds up many initiatives at once, and initiatives that started in different years cannot be added together unless they are all converted to one common point in time. This means the same initiative can legitimately show a different value on the two pages. Both pages state their anchor on screen so the difference is visible rather than surprising.

The reduction does not start straight away. The anchor month itself, plus the 11 months either side of it, are counted in full — every dollar in that period is worth a whole dollar. We call this the cushion. Beyond it, the reduction begins and grows with distance.

For this project the anchor is January 2024, so January to December 2024 is the cushion and the reduction begins in January 2025. The rule works in both directions: a month 11 months before the anchor would also be counted in full.

The adjustment factor:

Why there is a cushion at all: Amounts inside roughly a year of the anchor are treated as near enough to now to need no adjustment.

Why the −11 in the formula: the countdown starts at the edge of the cushion, not at the anchor. A month falling one month past the cushion is reduced by one month's worth, not a full year's. Without this, values would drop sharply the moment they crossed the boundary instead of easing down.

Worked example of the factor (anchor January 2024, discount rate 12%):

No figure on the page is assumed or invented. Each one is either typed in by a user, read from company settings, or measured from recorded data.

A partial baseline is used exactly as recorded. If only 6 of the 12 baseline months have data, we use that 6-month total as recorded. We do not scale it up to estimate a full year.

Why: scaling up would be inventing emissions that were never measured. Using the data as recorded means a thin baseline understates savings rather than overstating them — the safer direction to be wrong in.

What this means for you: when the baseline year is incomplete, a note appears beneath the table saying so, and the savings figures should be read as a floor — at least this much, possibly more.

These are the seven tiles in the Expected Total Project Impact card. All seven use planned inputs and cover the whole project life.

How much pollution will this project avoid in total?

Why it is that way: the reduction percentage is applied to what the site actually emitted, then held constant for every year of the useful life. We assume the equipment performs the same each year, because we have no basis for assuming it degrades at any particular rate.

Jargon — tCO2e: tonnes of carbon dioxide equivalent. Different greenhouse gases warm the planet by different amounts, so they are all converted into the equivalent amount of carbon dioxide to give one comparable number.

How much money will this project save in total?

Why it is that way: this is a simple total, with no discounting applied. It answers "how many dollars in total", not "what is that worth today". The discounted version of the same question is NPV, below.

What will it cost to build?

Why it is that way: no calculation is applied. It is shown alongside the savings so the size of the investment is visible next to the size of the return.

After accounting for the fact that future money is worth less than money today, is this project ahead or behind?

Jargon — NPV (net present value): take every dollar the project spends and every dollar it saves, adjust each one for when it happens, and add them all up. A positive NPV means the project creates value. A negative NPV means it destroys value, at the discount rate being used.

Every month gets a different factor — here is why. The Adjustment Factor is the share of a dollar we count for money arriving in a particular month. A factor of 1.0000 means we count the whole dollar. A factor of 0.8845 means we count 88.45 cents of it. January 2026 is 24 months after January 2024; December 2028 is 59 months after it. The further away, the longer that money has been out of reach, so the smaller its factor. There is no single factor for the project — there are 60, one per month.

Worked — every distinct step, with the middle months summarised:

Only six rows are shown above. The full calculation has 36 saving months, each with its own factor sliding gradually from 0.8845 down to 0.6355. The four shown are simply the first, the last, and two anniversaries in between, so you can see the direction of travel.

A useful sense-check: the 36 saving months are worth $1,650,000 in plain dollars ($45,833.33 × 36). After adjustment they count as $1,243,014. The difference of $406,986 is what the 12% discount rate costs this project over its life. That is also why NPV ($243,010) is so much smaller than the plain Cost savings tile ($1.65M) minus the capex ($1M), which would suggest $650,000. Both are correct; they answer different questions.

Why the money spent is not adjusted: it leaves the bank in January 2024, which is the anchor month itself — zero months away. It is already in January-2024 dollars, so its factor is 1.0000 and the full $1,000,000 is counted.

Why it is positive here: over the full three years the discounted savings come to slightly more than the million spent. This is the answer to "is it worth doing".

What interest rate is this project effectively paying us?

Jargon — IRR (internal rate of return): the single interest rate at which this project would exactly break even. If the IRR is higher than your cost of borrowing, the project is worth doing. It lets you compare a building retrofit against, for example, a bond or a different project, on one common measure.

How to read it: 22.5% against a 12% discount rate means the project returns roughly double the hurdle it must clear. That is the same conclusion the positive NPV gives, expressed as a rate rather than an amount.

Why we search rather than solve: there is no direct formula for IRR with monthly cash flows. Searching is the standard method and is what spreadsheet software does internally.

When IRR cannot be produced: sometimes no rate makes NPV zero. There are two opposite reasons, and the page shows a different thing for each.

How long until we get our money back?

Why counting starts at the project start, not at first savings: the money leaves the bank in January 2024. The two years spent building are two years of waiting for it back. Measuring from first savings would make a two-year build appear to cost nothing in time, and would give 1 year 10 months instead of 3 years 10 months — a difference of two full years on the same project.

Why whole months: we do not estimate a part-month. The 22nd savings month is the month in which the total crosses zero, and it is counted in full (rule 3, section 4).

How long until we get our money back, once we account for future money being worth less?

Each month's $45,833.33 savings is reduced by that month's own Adjustment Factor before being added, so the running total climbs more slowly than it does for simple payback:

Discounted payback is 52 months, or 4 years 4 months.

Why it is always longer than simple payback: each future dollar counts for less, so it takes more of them to reach the same total. Discounted payback can never be shorter.

Why show both: the gap between the two is the clearest single measure of what the discount rate is costing the project. Six months, in this example.

These are the seven rows in the Project Performance to Date card. They compare what the plan said should have happened by now against what actually happened.

The rows do not all cover the same period. This table states exactly which period each row uses, and it is the single most important thing to understand about this card.

For the four windowed rows, the two columns are built differently but describe the same seven months.

A common misunderstanding worth stating plainly: the Actual figure of $191,667 is already a seven-month figure. It is the difference between two seven-month amounts. It is not a yearly rate, and it must never be multiplied by 7÷12 a second time. The Expected figure is different in kind: it starts life as a yearly rate ($550,000/yr) and does need converting once. That asymmetry is correct, and it is the reason the two columns are calculated by different formulas.

Payback asks one question: when do we get our money back? That is a question about the whole investment, not about the months that happen to have passed.

If we truncated both sides at today, both would read "not yet", because at seven months the plan itself has only banked $320,833 against $1,000,000 spent. The row would compare nothing against nothing, and the benchmark would disappear.

So both sides run the full project life. What differs is only the inputs:

This gives the comparison that is actually useful: you planned to pay back in 3 years 10 months; at the rate this is genuinely saving, it will take 4 years 6 months.

All four payback figures are measured from the same starting point — month 1, when the money was spent. In this example, the actual discounted payback would arrive after the equipment's useful life ends, so we report that it does not pay back rather than projecting beyond that limit.

Why emissions can beat plan while cost savings miss it: these two rows measure different things and can genuinely move in opposite directions. Emissions fell more than expected, but the financial saving depends on energy prices as well as volumes. If prices fell during the period, the same avoided energy is worth less money. This is a real result, not an inconsistency.

These are the only values in the system that are not taken from your data. Each one is listed with where it came from.

An incomplete baseline understates savings. If the baseline year is missing months, the comparison uses less data than a full year and every saving figure comes out lower than the truth. A note appears beneath the table whenever this is the case. Read those figures as a minimum.

Very early figures rest on very little. The actual payback figures project a measured savings rate across several years. Early in a project that rate is based on only a few months of bills. Below three complete months the figure is labelled to say so, but the underlying caution applies for longer than that.

Two payback rows sit in a table headed "to date". As explained in section 8, this is deliberate and the alternative is worse. It is nonetheless the one place where the card's heading is not literally true of every row, and it is worth pointing out when sharing a screenshot of the table.

The same initiative may show different values on different pages. The initiative detail page values money in the project's own start month; the Scenario Planner values it in today's money. Both are correct for their purpose, and both state their basis on screen. When quoting a figure, quote the page it came from.