Buy budget allocation by marginal return
Spreads the buying budget across categories in steps: each next step goes where the marginal return (the growth of gross profit per rouble of purchase) is higher. Compares the result with the split by last year’s shares.
At a glance
- What question does it answer?
- How to spread the budget so that the last rouble gives the same profit in all categories.
- What data is needed?
- The buying budget.
- How to read the result?
- Compare the gain with the split by shares: a small gain means the shares are close to the optimum. Check saturated categories especially. Example on the default values: Allocated of the budget: RUB 40 million; Gross profit at the optimal allocation: RUB 25.5 million; Gross profit at last year’s split: RUB 25.4 million; Gain over last year’s split: +RUB 131,168. The figures are illustrative: enter your own to see your result.
Why and how to reason
A rouble of purchase gives diminishing profit, so the budget is split so that the marginal return equalizes. In the example the budget is 40 m RUB and there are three categories: outerwear 18.6, knitwear 12.2, denim 9.2 m RUB; the gross profit is 25.51 m RUB against 25.37 m RUB by last year shares.
When to use it
Apply when spreading the season budget across categories when there is a history of purchases and profit for at least two or three seasons. The budget and weeks of stock by category are counted by otbcat.
How to read the result
A small gain over the shares means the shares are already close to the optimum; check saturated categories especially.
Common mistakes
The parameters are estimated from one season, the category minimum is not limited, MOQ is ignored.
What data is needed?
- The buying budget.
- The parameters a and b of each category from past seasons.
- Category shares and minimums.
Inputs
| Field | Unit | Value in the example | Meaning |
|---|---|---|---|
| Buying budget | ₽ | 40,000,000 | The sum of the purchase at cost for the season that is to be spread. The example value is illustrative. |
| Allocation step | ₽ | 100,000 | The size of a portion of the budget at each step. A smaller step: more precise but slower. Not above the budget. The example value is illustrative. |
| Minimum marginal return (optional) | × | 0 | Roubles of gross profit per rouble of purchase: below this return a rouble is not invested. For example the carrying cost rate per season (carrycost). 0: the whole budget is spent. The example value is illustrative. Optional. Empty: the field is ignored (counted as zero in sums). |
How it is calculated
| Step | Formula |
|---|---|
| Category profit | GM(x) = a × (1 − e^(−x ÷ b)) |
| Marginal return | GM′(x) = (a ÷ b) × e^(−x ÷ b): the profit growth per rouble of purchase |
| Allocation | the budget is split in steps of 100,000 ₽; each step goes to the category with the largest GM′ |
| Comparison | profit of the optimum − profit by last year shares at the same budget |
Worked numbers
Default values from the app; in the interactive version you enter your own.
- Allocated of the budget
- 40 m ₽
- Gross profit at the optimal allocation
- 25.5 m ₽
- Gross profit at last year’s split
- 25.4 m ₽
- Gain over last year’s split
- +131,168 ₽
- Marginal return across categories: maximum, times
- 0.356
- Marginal return across categories: minimum, times
- 0.354
- Unallocated budget remainder
- 0 ₽
- Optimum
- By last year shares
Data as a table
| Measure | Optimum | By last year shares |
|---|---|---|
| Outerwear | 18.6 | 16 |
| Knitwear | 12.2 | 14 |
| Denim | 9.2 | 10 |
Allocation by category
| Category | Budget, optimum | Share | By last year shares | Gross profit, optimum | Marginal return | Purchase to scale b |
|---|---|---|---|---|---|---|
| Outerwear | 18.6 m ₽ | 46.5 % | 16 m ₽ | 10.9 m ₽ | 0.355 | 0.93 |
| Knitwear | 12.2 m ₽ | 30.5 % | 14 m ₽ | 8.46 m ₽ | 0.354 | 1.22 |
| Denim | 9.2 m ₽ | 23 % | 10 m ₽ | 6.15 m ₽ | 0.356 | 1.15 |
The parameters a and b are estimated by you from past seasons and give a hypothesis, not a fact: check it on two or three seasons and set the minimum for the brand core and MOQ.
What to do with the result: shift the budget gradually to the categories with a high marginal return, comparing with the OTB remainder and the availability of sizes; by the end of the allocation the marginal returns equalize.
A typical mistake: estimating a and b from one season, not limiting the category minimum and ignoring MOQ.
Nuances and limits
- Estimating a and b from one season.
- Not limiting the category minimum.
- Ignoring MOQ.