Price volume mix analysis: why you grew but margin fell

Revenue was up. Units were up more. Gross margin percentage was down. Nobody changed a price list. This combination shows up in board packs constantly, and "it's mix" is the answer that gets muttered without being measured. Price volume mix analysis measures it, and does so in a form that has to tie to the total revenue change exactly, which makes it hard to fudge.

Price volume mix analysis

A decomposition of the change in revenue between two periods into three parts: the effect of selling at different prices, the effect of selling a different total number of units, and the effect of those units being distributed differently across products. Together the three account for the whole change.

State your convention before you calculate anything

There is no single correct decomposition. The cross term — the interaction between a price change and a volume change — has to be allocated somewhere, and different companies allocate it differently. What matters is that you pick one, write it down, and use it every period. Here is the convention used throughout this article:

For each product line i, with prior period 0 and current period 1:

  g = total units in period 1 ÷ total units in period 0

  Price effect  = (P1 − P0) × Q1
  Volume effect = P0 × Q0 × (g − 1)
  Mix effect    = P0 × (Q1 − Q0 × g)

Sum of all three, across all lines = Revenue1 − Revenue0

Why it closes: the price and volume effects for one line collapse algebraically to P1Q1 − P0Q0, because (P1 − P0)Q1 + (Q1 − Q0)P0 = P1Q1 − P0Q0. Splitting that (Q1 − Q0)P0 term into a proportional part (volume) and a residual (mix) leaves the total untouched, since P0·Q0(g − 1) + P0(Q1 − Q0·g) = P0(Q1 − Q0). No plug, no rounding line, no residual bucket.

Worked example: three product lines

A distributor with three lines. Unchanged unit costs, so the margin story below is entirely price and mix.

That gap is the whole story: 15.0% more units produced 3.4% more revenue. Total unit growth g = 92,000 ÷ 80,000 = 1.15. Now compute each effect.

Price effect: (P1 − P0) × Q1

• A: ($48.50 − $50.00) × 44,000 = −$1.50 × 44,000 = −$66,000
• B: ($186.00 − $180.00) × 9,000 = $6.00 × 9,000 = +$54,000
• C: ($12.60 − $12.00) × 39,000 = $0.60 × 39,000 = +$23,400
• Total price effect = −$66,000 + $54,000 + $23,400 = +$11,400

Volume effect: prior revenue × (g − 1)

• A: $2,000,000 × 0.15 = +$300,000
• B: $1,800,000 × 0.15 = +$270,000
• C: $360,000 × 0.15 = +$54,000
• Total volume effect = $4,160,000 × 0.15 = +$624,000

This is the counterfactual: if every line had grown 15% at last year's prices, revenue would have risen $624,000. It did not, because the growth was not proportional.

Mix effect: P0 × (Q1 − Q0 × g)

• A: proportional share = 40,000 × 1.15 = 46,000 units; actual 44,000; shortfall 2,000 × $50 = −$100,000
• B: proportional share = 10,000 × 1.15 = 11,500 units; actual 9,000; shortfall 2,500 × $180 = −$450,000
• C: proportional share = 30,000 × 1.15 = 34,500 units; actual 39,000; surplus 4,500 × $12 = +$54,000
• Total mix effect = −$100,000 − $450,000 + $54,000 = −$496,000

The reconciliation

The one-sentence narrative

The point of the arithmetic is to produce a sentence a non-finance executive can repeat. From the table above: revenue grew $139,400 because 15% unit growth added $624,000, offset by $496,000 of mix as 1,000 Premium units at $180 were replaced by 9,000 Accessory units at $12, with net pricing roughly flat at $11,400.

Note what that sentence does. It names the driver, gives the number, and identifies the physical event behind it. It does not say "unfavourable mix variance". This kind of construction is the backbone of a defensible budget vs actual variance analysis too — the arithmetic is the same, with budget in place of the prior period.

Extending it to gross profit

Revenue mix is interesting. Margin mix is what the CFO wants. Add unchanged unit costs — A $32.00, B $99.00, C $8.40 — and recompute.

Check the blended margins: $1,638,000 ÷ $4,160,000 = 39.38%, and $1,672,800 ÷ $4,299,400 = 38.91%. Two of the three lines improved their own margin percentage: B rose from 45.0% to 46.8% on a $6 price increase, and C rose from 30.0% to 33.3% on a 60-cent one. Only A fell, from 36.0% to 34.0%, because its price dropped $1.50 against a flat cost. Yet the blend fell half a point despite gross profit dollars rising $34,800.

If your unit costs also moved between the periods, the margin story needs a fourth effect for cost, and the cleanest way to keep it honest is to compute each line's margin from the same cost basis in both periods — the gross margin calculator does the per-line arithmetic, and the contribution margin calculator does it after variable selling costs, which is the version that tells you which line to push.

Practical rules for building this in a real system

• Choose the grain deliberately. Price volume mix at SKU level on 40,000 SKUs produces a mix effect that is real but unusable. Run it at the level someone owns: product family, brand, channel or customer segment. You can always drill.
• Use realised price, not list price. Average selling price should be net revenue ÷ units after discounts, rebates and credit notes. If your list price is in the calculation, every rebate shows up as mix and the analysis misleads. Model stacked discounts separately with the discount cascade calculator.
• Handle new and lost products explicitly. A line with zero prior units has no prior price, so price and mix are undefined for it. Give new and discontinued products their own columns: a new line's whole current revenue (current price × current units) is incremental, and a lost line's whole prior revenue is a deduction. Forcing either into mix makes mix meaningless.
• Separate currency before you separate mix. In a multi-currency business, translate both periods at the same rate to get a constant-currency price volume mix, then report FX as a fourth effect. Otherwise every devaluation appears as a price cut.
• Units must be comparable within a line. Cases, eaches and pallets in the same unit column produce a mix effect that measures packaging, not commercial behaviour. Normalise to a base unit of measure first.
• Always print the reconciliation row. If price + volume + mix does not equal the revenue change to the dollar, do not publish. A closing check is the cheapest quality gate in reporting.

What each effect tells you to do

Getting price and volume by line out of your ERP

The maths is a few multiplications. The effort is assembling matched pairs: net revenue and units by line for two periods, on the same product hierarchy, after credit notes, excluding intercompany, with returns netted against the right period. Product hierarchies change between periods, which is usually where the reconciliation breaks — and it is why this analysis often runs quarterly instead of monthly. Building the ranked revenue base for it is the same exercise as a top customers by revenue report, with the grouping swapped from customer to product.

This is exactly the shape of question ERPray was built for. Ask "net revenue and units by product family for this quarter and the same quarter last year, excluding intercompany" and get both periods computed live from your own account, with the query shown underneath so you can confirm how returns and credit notes were treated before you decompose the change.