Methodology
The arithmetic behind the payback period
Two models you can already read, one number you supply yourself, and a division. This page prints every step, so a figure that looks like research is visible as what it is.
This calculator has no model of its own. The problem half is the cost of chaos model and the fix half is the ERP cost model, both taken unchanged, and what happens here is one subtraction and one division. The single figure it introduces — the share of the problem a new system removes — is not ours either. It is a field you fill in, and the page redraws the answer at three other values of it so that no single result looks authorised.
Outputs
Every figure the calculator prints, and the arithmetic behind it
Nine figures, in the order the calculator produces them. Nothing else is involved at any point.
| Figure | How it is calculated |
|---|---|
| Recovered each year | annual cost of the problem × recovery share ÷ 100 |
| Year-one investment, low | the estimator’s low total × your licence figure |
| Year-one investment, high | the estimator’s high total × your licence figure |
| What still runs each year, low | (1 + weight × 0.15) × your licence figure |
| What still runs each year, high | (1 + weight × 0.28) × your licence figure |
| Net saving, best case | recovered − what still runs, low |
| Net saving, worst case | recovered − what still runs, high |
| Shortest payback | investment low ÷ (best-case net saving ÷ 12) |
| Longest payback | investment high ÷ (worst-case net saving ÷ 12) |
The headline, written out in full
- shortest = year-one investment, low ÷ (best-case net saving ÷ 12)
- longest = year-one investment, high ÷ (worst-case net saving ÷ 12)
- Both are rounded up to a whole month, because a payback that lands mid-month has not landed.
Every money figure is rounded before it is used, not after — to the nearest 5,000 below 100,000, the nearest 10,000 from 100,000 up to 500,000, and the nearest 25,000 at 500,000 and above, with low ends rounded down and high ends rounded up. That is why the workings printed under the result add up exactly as shown. A sum that is out by one because the display rounded it is a sum nobody trusts twice.
The assumption
The one number nobody can source
The recovery share is the percentage of the annual problem cost that you believe a new system actually removes. It is the only figure in the whole calculator that is neither yours by observation nor published on another page, and we do not supply it. Any number we printed there would be one we made up.
The field opens at 50 because a field has to open at something, and that is the whole of the reasoning. It is a placeholder, not a finding. It is clamped at 100 at the top, because a share above that would claim the system removes more than the problem costs, which is not pessimism worth protecting against — it is arithmetic that cannot be true.
Whatever you type, the page also redraws the entire result at 25%, 50% and 75% in a strip under the headline. Three answers rather than one, deliberately: a single payback figure reads as a finding, and three of them read as what this actually is — an answer that moves entirely with a number nobody can source. The distance between the three is the real output of the tool.
There are two places the calculator refuses to answer. Without a licence figure it prints no money at all — no investment, no payback, only a prompt — because the licence is the unit the whole ERP model is denominated in. And the worst case decides whether there is a payback at all: if the recovered saving does not clear the higher of the two recurring figures, the page says it does not pay back rather than printing the lower bound on its own. Printing half a range is advertising the half we happen to like.
The two halves
What each borrowed model contributes
Both models are imported unchanged, so this page can never arrive at a different problem cost or a different implementation range from the pages that specialise in each. What crosses between them is small.
The problem half
- Six of the cost of chaos fields, on the same defaults that calculator opens with.
- Days to close the month is not asked here and is passed as zero. It feeds the close-cycle narrative on the other page and nothing in this one's arithmetic.
- The length of a full-time week is fixed and not asked either. It only ever produces the full-time equivalent, and this page never prints that.
- One figure crosses over: the annual cost of the problem.
The fix half
- All seven ERP estimator answers, on the estimator's own published weights.
- Four multiples cross over: the low and high first-year totals, and the low and high figures that still renew after year one.
- The licence figure you supply is what turns those multiples into money. It is yours, from your own vendor's quote, and there is no price of ours anywhere in it.
Bounds
Where the model stops
Six things it does not do. Each one is a place where a real payback period would be longer, shorter, or simply not the right question.
- There is no discounting and no time value of money. A dirham saved in the third year is counted as a dirham saved today.
- The investment is treated as paid up front and the saving as arriving in twelve equal monthly parts. Real programmes pay in stages and save nothing at all until go-live.
- The two ends are corners, not scenarios. The shortest pairs the cheapest implementation with the cheapest support; the longest pairs the dearest with the dearest. A real programme lands somewhere inside the box, and not necessarily on the line between the two corners.
- There is no ramp and no learning curve. Nothing here models the months when the new system is live and the saving has not arrived yet, which is every implementation there has ever been.
- A payback period is not a business case. It says when the money comes back, not whether the programme was worth doing, and not whether it will work — most of what goes wrong is not in the price.
- The close cycle is not in this calculator at all. Days to close is passed as zero, so nothing about the month-end window reaches the answer, even though it is usually the reason the conversation started.
Payback Calculator
Argue with a step, not with the months
Every step is on this page so that it can be disputed. Tell us which one is wrong for your programme and we will tell you what we would put in its place.
Monday to Friday, 9:00 AM – 6:00 PM (GST)
Prefer we call you?
Leave your WhatsApp number and we will reach out.