Promotion effectiveness

Which promotions add margin, which give it away, and how to clear the season

24 products, 104 weeks. 6 promotions add margin, 15 lose it and 3 cannot yet be called. P18’s promotion sells 139 extra units a week and loses £1,626 of margin doing it.

Prepared for a constructed retailer. Every figure below was generated by the model from a constructed population when this report was produced — not any firm’s data or result.

How these figures were produced — the model

Your questions, answered

What a promotional calendar needs to know, in a line each

Do our promotions make money?: Mostly not. 6 of 24 add margin, 15 lose it and 3 cannot yet be called.
Which promotion costs us most?: P18: £1,626 of margin lost every promotion week, at 41% off.
How can a promotion sell more and still lose money?: P18 sells 139 extra units a week on promotion and loses £1,626: the discount on units that were selling anyway, the dip the week after and the sister product’s lost sales outweigh the extra units.
Are stock-outs hiding what promotions do?: 22 of 24 products sold out in some week. Read as demand, those weeks understate the lift.
How should we clear the season?: With a markdown plan per product, which beats the standing rule by £4,644 across the range.

Background and approach

Not “did it sell more”, but “did it add margin”

Almost every promotion sells more units. What it costs is the discount given on units that would have sold anyway, the dip the week after as customers who stocked up stop buying, and the sales its sister product loses while it is on deal.

24 products in pairs of substitutes, read weekly over 104 weeks.
Each product’s price response, promotional lift, after-promotion dip and effect on its sister product are estimated from its sales — none is assumed.
Weeks where a product sold out are left out of the demand estimate, because they record what the shelf held, not what customers wanted.
Every promotion is read as margin per promotion week, with a 90% interval.

Key findings

Most promotions here sell more and earn less

Promotions read24one per product

Add margin6whole interval above zero

Lose margin15whole interval below zero

Not determined3interval spans zero

Margin lost£6,387a promotion week, across the losers

Markdown plans gain£4,644over the standing rule

Where the margin goes

The three costs a unit count does not show

For the promotions that lose most: the margin made or lost on the product’s own units against a normal week, the dip in the week after, and what the sister product lost.

Own unitsThe week afterSister product

Margin per promotion week.

Show as a table

	Own units	The week after	Sister product
P18	−£1,375	−£183	−£68
P17	−£510	−£279	−£69
P22	−£185	−£347	−£104
P20	−£398	−£170	−£50
P07	−£366	−£71	−£66
P15	−£75	−£105	−£220

P18 loses the most: £1,626 a promotion week (−£1,731 to −£1,525), at an average discount of 41%.

Every promotion

Each promotion’s margin, with its interval

Left of the line loses margin; right adds it. A whisker that crosses the line is not determined — not break-even, and not a reason to keep the promotion.

Show as a table

	Estimate	From	To
P18 · Category 9	−£1,626	−£1,731	−£1,525
P17 · Category 9	−£858	−£965	−£745
P22 · Category 11	−£637	−£732	−£542
P20 · Category 10	−£618	−£683	−£554
P07 · Category 4	−£502	−£542	−£464
P15 · Category 8	−£401	−£487	−£314
P05 · Category 3	−£389	−£510	−£269
P19 · Category 10	−£321	−£383	−£251
P16 · Category 8	−£225	−£385	−£43
P24 · Category 12	−£180	−£214	−£144
P08 · Category 4	−£174	−£219	−£126
P02 · Category 1	−£158	−£238	−£69
P12 · Category 6	−£150	−£205	−£95
P06 · Category 3	−£103	−£168	−£37
P13 · Category 7	−£63	−£123	£2
P09 · Category 5	−£44	−£57	−£30
P23 · Category 12	−£37	−£89	£19
P03 · Category 2	£13	−£20	£47
P04 · Category 2	£47	£7	£92
P10 · Category 5	£118	£78	£163
P01 · Category 1	£156	£54	£254
P21 · Category 11	£204	£78	£335
P11 · Category 6	£207	£91	£326
P14 · Category 7	£297	£181	£427

Product	Average discount	Extra units a week	Net margin	Verdict
P18	41%	139	−£1,626	loses margin
P17	28%	195	−£858	loses margin
P22	28%	284	−£637	loses margin
P20	31%	276	−£618	loses margin
P07	28%	222	−£502	loses margin
P15	29%	101	−£401	loses margin
P05	29%	242	−£389	loses margin
P19	29%	83	−£321	loses margin
P16	25%	296	−£225	loses margin
P24	27%	73	−£180	loses margin
P08	28%	163	−£174	loses margin
P02	24%	231	−£158	loses margin
P12	26%	143	−£150	loses margin
P06	29%	115	−£103	loses margin
P13	28%	185	−£63	not determined
P09	24%	122	−£44	loses margin
P23	27%	82	−£37	not determined
P03	24%	265	£13	not determined
P04	28%	147	£47	adds margin
P10	28%	313	£118	adds margin
P01	26%	151	£156	adds margin
P21	26%	287	£204	adds margin
P11	25%	188	£207	adds margin
P14	24%	285	£297	adds margin

More units, less margin

P18 sells more on promotion, and loses margin doing it

The promotion a unit report would call a success.

Extra units139a promotion week

Average discount41%off shelf price

Price response−0.65−1.07 to −0.22

Net margin−£1,626a promotion week

The extra units do not cover the discount given on the units that would have sold anyway, the dip after, and what its sister product loses.

Stock availability

A promotion that sells out looks weaker than it was

Promotion weeks are the weeks shelves run empty. Read as demand, a sold-out week says the promotion lifted sales less than it did — and the next promotion is planned smaller, or cut.

Sold-out weeks read as demandSold-out weeks left out

Show as a table

	Sold-out weeks read as demand	Sold-out weeks left out
P09	−20%	22%
P10	24%	59%
P04	31%	35%
P18	−4%	11%
P23	21%	29%
P20	−45%	36%
P16	−22%	8%
P11	3%	47%

22 of 24 products sold out in at least one week; 146 product-weeks in all were left out of the demand estimate.

For the plan: stock the promotion. A promotion that sells out spends its last days discounting a shelf nobody can buy from, and its results understate the next one.

Clearing the season

Markdown plans that clear the stock for more margin than the standing rule

For each product’s end-of-season stock, every markdown path that only deepens is costed under the fitted demand, salvage included, and the best is set beside the standing rule of 30% off throughout.

Gain over the rule£4,644across 24 products

Clearance6 weeksmarkdowns only deepen

Standing rulePlan

Show as a table

	Standing rule	Plan
P16	£1,773	£2,878
P18	−£1,170	−£213
P14	£1,845	£2,193
P08	£504	£787
P03	£340	£587
P01	£1,252	£1,484
P22	£1,462	£1,692
P17	£813	£1,037

Product	Plan, week by week	Gain
P16	10% → 20% → 20% → 20% → 30% → 30%	£1,105
P18	0% → 0% → 0% → 0% → 0% → 0%	£957
P14	20% → 20% → 30% → 30% → 30% → 30%	£348
P08	20% → 20% → 20% → 20% → 30% → 40%	£283
P03	10% → 20% → 20% → 20% → 20% → 30%	£248
P01	20% → 20% → 20% → 30% → 30% → 40%	£233
P22	20% → 30% → 30% → 30% → 30% → 40%	£230
P17	20% → 30% → 30% → 30% → 30% → 40%	£224

Conclusions and recommendations

What to do with the promotional calendar

Stop or re-price the promotions that lose margin: P02, P05, P06, P07, P08, P09, P12, P15, P16, P17, P18, P19, P20, P22, P24.
Keep the promotions that add margin: P01, P04, P10, P11, P14, P21.
Test before deciding on P03, P13, P23 — their intervals span zero; run each on a set of stores against the rest.
Re-cost P18’s promotion: it sells more units and loses margin.
Read every promotion on margin after the week-after dip and the sister product, not on units.
Stock promotion weeks: sold-out weeks make a promotion look weaker than it was.
Clear the season with the markdown plans, which beat the standing rule by £4,644.

Methodology

How this was produced

Weekly units per product modelled on price, promotion, the week after a promotion, the sister product’s promotions and season; sold-out weeks left out.
Promotion margin costed from those estimates, with 1,500 draws for its interval.
Markdown paths costed under the same demand, with the merchandiser’s out-of-season decay and a salvage value.
The population is constructed. Every figure was generated by the model when this report was produced.

The model: elasticities, lifts and intervals

Promotion and price effectiveness — sample report