You built a budget in January. It is June, and nobody has reopened it. A budget is not there to forecast accurately — it is there to show you, month by month, exactly where reality is diverging from what you expected. It is the variance that informs you, not the forecast.
The essentials in five points
- The budget’s value is in the comparison, not in the accuracy of the forecast.
- A global variance says nothing: you need to know whether it comes from volume, price or mix.
- Compare on a constant scope, or you are measuring a change of scope.
- Not every variance is worth treating: only those that are material and actionable.
- The review is monthly and short: one hour, four questions, one decision per variance kept.
1. Break it down before explaining it
“Turnover is 12% below budget” calls for no action. The same information broken down does call for one, because it points at a cause.
| Variance | What it measures | What it suggests |
|---|---|---|
| Volume | You sold fewer units than planned | Demand, season, stock-out, competition |
| Price | You sold the same volume, more cheaply | Discounts granted, competitive pressure, customer mix |
| Mix | You sold something other than what was planned | Demand shifting towards lower-margin products |
| Purchase cost | Your margin shrinks at an unchanged selling price | Supplier increase, a discount lost, exchange rates |
| Fixed costs | A structural cost has moved | A past decision, indexation, an omission in the budget |
The first three lines are constantly conflated in conversation. Turnover down with stable volume is a price problem; the same turnover with a stable price is a demand problem. The two call for different decisions, and the average does not distinguish them.
2. Compare like with like
Most inexplicable variances come from a badly framed comparison rather than a surprising reality.
- Compare on the same scope: a shop opened in March distorts everything until it is isolated.
- Compare on amounts excluding VAT, never on VAT-inclusive collections.
- Attach costs to their period, or an invoicing lag looks like a real variance.
- Strip out non-recurring items: a disposal, compensation, an exceptional refund.
- Check that the budget itself was not built on a different basis from the accounts.
The third line produces half of all false variances in a small business. A supplier invoice recorded a month late creates a flattering month followed by a disastrous one, while nothing in the activity has changed. Cut-off applies to monthly management too.
3. Filtering: what deserves action
Analysing every variance is the surest way of treating none of them. Two filters cut the list down to what matters.
- The materiality filter: Set a threshold, in amount and in percentage, below which a variance goes uncommented. Without one, the review drowns in movements of a few hundred dirhams.
- The action filter: Ask what you would do differently if you understood this variance. If the answer is “nothing”, the analysis is an exercise rather than a tool.
- Recurrence beats size: A 3% variance recurring for six months matters more than an isolated 15% one: the first is structural, the second is an event.
- Direction matters as much as size: An unexplained favourable variance deserves the same attention as an adverse one. Not knowing why you are winning is as risky as not knowing why you are losing.
The last line is systematically neglected. An exceptionally good month you cannot explain is information lost: either it is repeatable and you need to know how, or it is not, and the following months’ budget rests on an illusion.
An unexplained favourable variance is not good news
This is the reflex that costs most in management: you hunt for the causes of bad months and bank the good ones without looking. But an overshoot you cannot explain means either that you do not know what is working — and so cannot repeat it — or that there is an error in the accounts: a cost not recorded, revenue counted twice, an invoice attached to the wrong period. Good variances deserve the same sentence of explanation as bad ones.
4. The monthly review, in an hour
Regularity beats depth. A short review every month beats a deep analysis every six, because it still leaves time to act.
- Pull the month’s actuals and the year-to-date, against budget.
- Remove at once the variances below the materiality threshold.
- For each variance kept, say whether it comes from volume, price, mix or costs.
- Write one sentence of explanation per variance, with its author and the date.
- Decide: act, watch, or revise the budget — and note which.
- Carry the ones marked “watch” into the next review.
The fifth line is what separates a review from a meeting. Only three outcomes, and one of them is entirely legitimate: revising the budget when a starting assumption has proved wrong beats commenting on the same variance for eight months.
Mistakes to avoid
- Analysing a global variance without splitting volume, price and mix.
- Comparing a scope that changed during the year.
- Mixing net amounts with VAT-inclusive collections.
- Commenting on every variance, with no materiality threshold.
- Ignoring favourable variances because they suit you.
- Repeating the same analysis monthly without ever revising the budget.
Frequently asked questions
What use is a budget if the forecast is wrong?
It locates the variance. A budget’s value is not its accuracy but the fact that it gives a point of comparison: with no reference, a fall in turnover is neither good nor bad, it is simply a number.
How is a turnover variance broken down?
By separating volume, price and mix. Selling fewer units, selling the same volume more cheaply, or selling something other than planned all produce the same global variance and call for three different decisions.
What materiality threshold should be used?
The one that keeps the review manageable: in practice a threshold combining a minimum amount and a percentage. The aim is a handful of commented variances, not an exhaustive list nobody will read.
Should the budget be revised during the year?
Yes, when a starting assumption has proved durably wrong. Continuing to compare against a budget you know is wrong produces the same variance every month, commented identically, with no decision coming out of it.
Does BelloPOS produce these variances?
BelloPOS supplies the actuals: sales by period, by product and by rate from the free Lite licence onward, and analytics with Go. The budget itself and the comparison are kept alongside, in a spreadsheet or with your accountant.
What to take away
A budget is worth whatever the review that follows it is worth. Break down before explaining, compare on a constant scope, filter by materiality and by capacity to act, and close each variance kept with a written decision. One hour a month is enough, provided it is every month.
Sources
The figures and rules quoted above come from these pages, read on the date given in the article.
- Moroccan Tax Administration, 2026 General Tax Code
- Ministry of Economy and Finance, General Code of Accounting Standardisation, read 1 September 2026
Actuals, available without reconstructing them
BelloPOS gives sales by period, by product and by rate from the free Lite licence onward, and analytics from Go: the real half of the comparison, ready every month.
Read next
Other practical guides on the same subject: