Price Elasticity and Promotion Planning by Vertical

Price elasticity of demand is the percentage change in units sold divided by the percentage change in price that caused it, and for a merchandise planner it is the number that says how much volume a markdown, a promotion or a price rise will actually move. It sits underneath every markdown depth, every promotional calendar and every decision to hold a ticket, whether or not anyone has written it down. The formula is the same in every category; the traps in estimating it are not. This guide sets out the formula and how to read it, how to estimate elasticity honestly from your own markdown and promotion history, the margin-neutral lift that tells you whether a discount pays, how brand and price tier change the answer, and then follows ten verticals through the places they genuinely differ.

It belongs to the by-vertical series beside markup and pricing by vertical, which sets the first ticket and the margin it carries, and markdown and exit strategy by vertical, which spends the reductions planned against it. Merchandise planning KPIs by vertical sets out the sell-through and margin measures that elasticity analysis is read against, markdown optimization in retail covers when, how deep and where to take markdowns, and dynamic pricing for fashion retail covers when moving price continuously helps an apparel brand and when it backfires.

What price elasticity is and how to read it

The definition is a ratio of two percentage changes:

Price elasticity = % change in units ÷ % change in price

A markdown that cuts price 20% and lifts units 30% has an elasticity of 30% ÷ −20% = −1.5. A price rise of 10% that loses 5% of units has an elasticity of −5% ÷ 10% = −0.5. This guide keeps the sign, so "more elastic" means further from zero.

The arc, or midpoint, method

The simple version divides each change by its starting value, so the answer depends on which price you start from. Between a $60 ticket and a $42 markdown, the drop is 30% of $60, but the rise back is 42.9% of $42. The arc method — also called the midpoint method — divides each change by the average of its two values instead:

% change in units = (Q2 − Q1) ÷ ((Q1 + Q2) ÷ 2)

% change in price = (P2 − P1) ÷ ((P1 + P2) ÷ 2)

Arc elasticity = % change in units ÷ % change in price

On a markdown or a promotion, where the price moves a long way in one step, the midpoint method is the one to use, because it gives the same figure going down as coming back up. The worked example below shows the gap: the simple method reads the same two points as −1.67 or −0.78 depending on direction, while the midpoint method reads −1.13 both ways.

Elastic, inelastic and what each one means for revenue and margin

Elastic is a revenue statement, not a margin statement. A discount gives up margin on every unit, including the ones that would have sold at full price, so a markdown on an elastic product can raise revenue and still lose margin dollars. Whether it pays depends on the gross margin at the full ticket — the subject of the margin-neutral lift section below — and, at the end of a season, on what the unsold units would realize at exit.

Why a planner cares: four decisions elasticity sits under

Four decisions assume a value for elasticity whether or not anyone has estimated one.

• →Markdown depth. A first markdown is set to move a forecast quantity of residual stock in a forecast number of weeks. Too shallow and the stock reaches the next step anyway; too deep and the margin given up on units that would have sold at the shallower step is never recovered.
• →Promotional lift. A promotion is planned as a lift over baseline. A plan that assumes a lift the category cannot produce loses margin dollars, usually visible in the weeks after the event rather than during it.
• →Full-price sell-through. Every unit sold on promotion is sold below full price, and a promotional calendar that takes a growing share of a class's units off full price changes its sell-through rate profile and its markdown rate at once.
• →The margin trade-off on the ticket itself. A price rise to protect initial markup when cost moves is an elasticity bet: it holds only if units fall by less than the margin gained. Planning margin on a moving cost base covers the cost side of that decision; elasticity is the demand side.

Estimating elasticity honestly from your own history

There is no published coefficient for your category, customer and price tier. The estimate has to come from your own price changes, and your own history lies in three predictable ways: it confuses the season with the price, it records sales rather than demand, and it misses demand that moved rather than grew.

Where the price changes come from

Three sources produce usable price variation, in rising order of reliability.

1. →Markdown events. Every planned markdown is a price change with a date, but markdowns are confounded with a weak product and the season's decline, because markdowns are taken late and on styles behind plan.
2. →Promotions. A promotion is a temporary price change with a start and an end, which gives a before, during and after. A temporary price behaves differently from a permanent one, so promotional elasticity and markdown elasticity are different numbers and should be stored separately.
3. →Price tests. A deliberate test holds a price in one set of stores or one time window and changes it in a matched set. In stores, the matched set comes from store clustering: test doors and control doors drawn from the same cluster, so climate, demographics and capacity are similar — the method is in store clustering and localization by vertical.

Controlling for seasonality

The single largest error in reading a markdown is comparing markdown weeks with the full-price weeks before them, because the markdown weeks sit later in the season and would have sold less at full price anyway. The baseline is what the markdown weeks would have sold at full price, not what the earlier weeks did sell. Build it from a seasonal index: the same weeks last year for a comparable style that stayed at full price, or the rate of a control group of styles or doors that did not move price this year. The worked example below shows how much this one correction changes the reading.

Stock-outs and broken size runs censor demand

Units sold are not demand. A week in which the item, or the sizes most customers want, is out of stock records only what was available to sell: demand above stock becomes lost sales that never enter the history. This censoring biases elasticity both ways:

• →A markdown week with a broken size run shows a weak response because the customer who would have responded could not find a size. It reads as inelastic when the product was elastic and simply ran out of the sizes that sell.
• →A full-price baseline with stock-outs understates the baseline, so the lift measured against it looks larger than it was.

The correction is to screen the history before estimating: drop or flag style-weeks where in-stock on the core sizes fell below a threshold the team chooses, and record why. In footwear and apparel the screen runs on the size run, not the style total, because a style with plenty of units in the outlying sizes is out of stock to most of its customers.

Cannibalization and halo

A discount moves demand sideways as well as forward. Cannibalization is demand the promoted item takes from its siblings: a markdown on one crew-neck color takes units from the full-price colors beside it, and the category gains less than the promoted item does. Halo is the opposite: a promoted item brings customers who buy something else at full price, the way a promoted hero product can lift the attached items around it. Read a promotion at the class or category as well as the item, or the item's lift is counted as the category's.

Promotion planning: lift, pull-forward and the post-promotion dip

A promotion is planned as a lift over a baseline, and it is judged on what is left after three adjustments. Gross lift is the promotion week's units over the baseline. Pull-forward is the part of that lift that would have sold at full price in the following weeks anyway; the post-promotion dip is where it shows up. Cannibalization is the part taken from sibling styles. Net incremental units are what remains.

The figures below are illustrative, chosen to show the arithmetic rather than any category's pattern. A sweater with a $60 ticket and a $24 landed cost — $36 of margin per unit at full price — sells a baseline of 200 units a week in season. It runs a one-week promotion at 20% off, $48, which leaves $24 of margin per unit.

A sibling style with the same $36 unit margin sells 20 fewer units during the promotion week, so the category gains 50 − 20 = 30 incremental units, not 100.

Now the margin dollars. In the promotion week the sweater earns 300 × $24 = $7,200, exactly its baseline of 200 × $36 = $7,200: on the week's own numbers the event broke even. Across the window it earns $7,200 + (170 + 180) × $36 = $7,200 + $12,600 = $19,800 against a baseline of $7,200 + 400 × $36 = $21,600, and the sibling's lost units cost another 20 × $36 = $720. The event that broke even in its own week cost $2,520 of margin dollars across the window.

The margin-neutral lift formula

Whether a discount can pay at all is a margin question, and it has a closed-form answer. Hold gross margin dollars constant, with P the full ticket, C the landed cost, d the discount as a share of the ticket, Q0 the baseline units and Q1 the units needed on promotion:

Q0 × (P − C) = Q1 × (P × (1 − d) − C)

Divide both sides by P and write g for the gross margin percentage at the full ticket, (P − C) ÷ P:

Q1 ÷ Q0 = g ÷ (g − d)

So the unit lift that holds gross margin dollars is:

Required lift = d ÷ (g − d)

Three things follow. The required lift rises steeply as the discount approaches the margin: on a 60% gross margin, 20% off needs a 50% lift, 30% off needs 100%, and 40% off needs 200%. At a discount equal to or deeper than the gross margin, each promotional unit earns nothing or loses money and no lift holds margin dollars. And a thin-margin product needs more lift for the same discount than a rich-margin one, which is why the same promotional depth asks more lift of a thin-margin opening price point than of a richer-margin best tier. The margin in the formula is gross margin on full landed cost, the same basis as the gross margin percent and initial markup formulas; where a vendor funds part of the promotion through markdown money or trade funds, the funded share comes off the discount before it goes in.

The test is made against net lift, not gross. In the illustration above, the gross lift was 100 ÷ 200 = 50%, exactly the 50% that 20% off needs on a 60% margin; the net lift after pull-forward and cannibalization was 30 ÷ 200 = 15%. That is the whole difference between an event that looks margin-neutral and one that loses $2,520.

A promotion can still be bought deliberately — to acquire customers, clear a size run before it breaks or hold a slot in a retail partner's calendar — but that is a cost to plan and name, and the margin-neutral lift tells you how large it is.

Worked example: arc elasticity from a markdown, then the margin-neutral test

The figures below are illustrative, chosen because they divide cleanly. They are not benchmarks and not drawn from any brand.

A brand takes a 30% markdown on a merino crew-neck sweater in weeks 5 to 8 of its selling window: $60 to $42. Landed cost is $24, so the full-ticket gross margin is ($60 − $24) ÷ $60 = 60%. Every core size was in stock in all eight weeks, so no week is screened out for a broken run.

Step 1 — the observed history.

Step 2 — the seasonal baseline. Last year, a comparable style that stayed at full price sold 10% less a week in weeks 5 to 8 than in weeks 1 to 4. The baseline for the markdown weeks is therefore 100 × 0.90 = 90 units a week at $60, not 100.

Step 3 — arc elasticity.

• →% change in units = (135 − 90) ÷ ((90 + 135) ÷ 2) = 45 ÷ 112.5 = 40.0%
• →% change in price = (42 − 60) ÷ ((60 + 42) ÷ 2) = −18 ÷ 51 = −35.3%
• →Arc elasticity = 40.0% ÷ −35.3% = −1.13

The style is elastic. Two wrong readings of the same history show why the method matters:

• →Without the seasonal baseline, comparing 135 with 100: % change in units = 35 ÷ 117.5 = 29.8%, and elasticity = 29.8% ÷ −35.3% = −0.84. The style reads as inelastic, and a planner would conclude that markdowns do not work on it.
• →With the simple method, dividing by the starting values: units rose 45 ÷ 90 = 50% on a 30% price cut, an elasticity of −1.67; measured back up from the markdown, units fall 45 ÷ 135 = 33.3% on an 18 ÷ 42 = 42.9% price rise, an elasticity of −0.78. Same two points, two answers.

Step 4 — revenue and margin at the markdown.

Revenue rose by $270 a week, as an elastic reading predicts. Margin dollars fell by $810 a week. The markdown was elastic and margin-negative at the same time. On the margin-neutral formula, 30% off a 60% margin needed a lift of 0.30 ÷ (0.60 − 0.30) = 100%, or 180 units a week; the style delivered 135 ÷ 90 − 1 = 50%. Whether the markdown was still right depends on the exit: if the units not cleared would have gone to off-price or outlet below $42, clearing them earlier at $42 can be the better trade, and the comparison that decides it is against the exit value, not against full price.

Step 5 — testing a proposed 20% event. For the same style at the same point next season, the merchant proposes 20% off, $48, instead of 30%.

• →Margin-neutral lift = 0.20 ÷ (0.60 − 0.20) = 50%: 135 units a week at $48 earn 135 × $24 = $3,240, the full-price baseline.
• →Lift implied by the −1.13 estimate: % change in price = (48 − 60) ÷ 54 = −22.2%, so the midpoint change in units is −1.133 × −22.22% = 25.2%, and (units − 90) ÷ ((units + 90) ÷ 2) = 25.2% solves to about 116 units a week, a lift of about 29%, earning about $2,780 a week.

On this estimate, 20% off does not hold margin dollars either; it clears about 26 more units a week than full price at a lower margin cost than the 30% markdown. The decision between 20% and 30% is then about how many units must clear by the exit date, which is a sell-through target, not an elasticity question. The estimate came from a late markdown, so applying it to an earlier event is a judgement the season hindsight has to check.

Brand, line and price-tier effects

What "brand elasticity" means

Brand elasticity is a working name for how much a brand's equity insulates its demand from price moves. A brand with a distinct product, a customer who seeks it out and a consistent full-price history usually loses fewer units when its price rises and gains fewer when it discounts than an undifferentiated product in the same category, because its customer is buying the brand as well as the item and has fewer close substitutes. A brand without that equity competes on price with everything near it, and its elasticity is set by the gap to the nearest alternative.

The insulation runs both ways: a well-insulated brand has more room to hold a ticket when cost moves and less to gain from a discount. And frequent promotion erodes the insulation it relies on. A customer who has seen the brand at 30% off every few weeks learns a lower reference price and waits for the next event, so the full-price weeks weaken and the promotion weeks carry more of the volume; the history then shows rising lift from discounting that is partly a trained customer. There is no published coefficient for any brand's insulation; the only honest reading comes from its own history, split by tier and by channel.

Why a good-better-best ladder behaves differently by tier

A single elasticity for a whole category averages tiers that respond differently. In most categories the opening price point draws the customer most aware of price and most willing to substitute, so it tends to respond most to discounts and to lose most to price rises. The better tier carries the core volume. The best tier frames the ladder and carries a customer buying a specification, a material or the brand at its fullest; it tends to respond least to price, and a deep discount on it can draw customers up from the better rung rather than bring new ones in.

A discount on the better rung that brings its price close to the opening price point takes units from the opening tier — cannibalization inside the ladder. One coefficient for the whole ladder hides the tier differences and the substitution between rungs, which is why estimates are stored by tier, and why the markup and pricing guide treats the ladder as a set of rungs rather than a single price.

The ten verticals compared

The table sets out where elasticity and promotion planning differ. The formula does not change between rows; what drives demand and what fools the estimate do, and the sections after the table take each vertical in its own terms.

Apparel: seasonal fashion against never-out-of-stock basics

Apparel runs two elasticity problems under one roof. Seasonal fashion is bought once, sells against a season clock and exits through a markdown cadence: a first markdown timed to a sell-through threshold, deeper steps after it, and an exit to outlet or off-price. Its elasticity is read from markdown events, and every one of them is confounded with the season's decline and with the fact that the styles marked down are the ones selling behind plan. Never-out-of-stock basics — the core tee, the five-pocket jean, the white shirt — are replenished continuously, so they have long full-price histories and a promotional pattern instead of a markdown one: event promotions, multi-buys, and a sitewide offer that catches them alongside the fashion layer. The never-out-of-stock entry and core and NOOS replenishment programs cover how the core is run.

The two need different estimates. A seasonal style's elasticity is a property of that style late in its life and does not transfer cleanly to next season's style, which is a different product; the useful reading is at the class and the markdown step rather than a coefficient per style. A basic's elasticity is a property of a product the customer knows, measured repeatedly, and it is where promotional lift, pull-forward and reference-price erosion show most clearly, because the same item is promoted again and again.

The apparel-specific trap is the size run. A fashion style marked down in week 10 is usually already missing its best-selling sizes, so the markdown reaches a customer who often cannot find a size, and the response looks weak. Read without a core-size in-stock screen, that history says the class is inelastic and that the next markdown should be deeper, when the honest reading is that the class was elastic and ran out of the sizes that sell. Screen style-weeks on core-size in-stock before estimating, and read the result by class and markdown step. Markdown optimization strategies covers the cadence itself.

Footwear: the franchise price point, model-year closeouts and the size-and-width run

Footwear prices at the model, and a franchise price point is the number least free to move. The customer knows what the running franchise or the boot line costs, a new colorway on a carryover model joins at that price, and the price usually moves at a model-year changeover, the date it can change without repricing pairs already in the channel. So in-season price variation is small, and most of a footwear brand's usable price history comes from one event: the model-year closeout, when the outgoing model steps down while its successor launches at the new price. That event mixes three things at once — the old model's price falling, the new model arriving beside it, and the customer choosing between them — so the closeout's response is partly substitution between two models, not new demand.

Where dealers sell under a minimum advertised price policy, the advertised floor removes deep promotion as a lever on current models, so the brand's own history of discounting on a current model may be thin or confined to its own channel.

The vertical-specific trap is size and width availability. A model's demand is spread across a size run and, for many franchises, more than one width, and closeout stock is by definition the pairs that did not sell: the outlying sizes and the wider or narrower fits. A closeout at 40% off that sells slowly may be selling slowly because the remaining pairs fit few customers, not because the customer does not respond to price. Screen closeout history by size and width position, and estimate response on the positions that were in stock, or the model-year step-down will be set too deep next year. Planning a model-year changeover and assortment planning for footwear brands cover the run and the changeover date.

Accessories & bags: gifting tiers and the evergreen core

Accessories carry two price behaviors under one brand. The evergreen core — the black leather tote, the house wallet, the belt in the signature hardware — is planned to sell at full price season after season, and a brand that discounts it trains its customer to wait. The seasonal colorway on the same body is bought once and takes the fashion markdown. The two have the same ticket and very different histories, so an elasticity read across them is an average of a product that is rarely discounted and one that always is.

The demand driver that fools accessories estimates is gifting. Small leather goods — card cases, key fobs, compact wallets — sit at gift-tier price points and concentrate their volume into the gifting peaks: the year-end holidays and, for many brands, occasion dates through the year. Promotions scheduled into a peak take credit for the peak. A gift-tier event read against a non-peak baseline overstates elasticity badly. Build the baseline from the same peak in a year without the event, or from comparable full-price items in the same weeks, before reading the promotion.

Gift sets priced below the sum of their parts give away margin the way a discount does, and they belong in the margin-neutral test with the set price as the discounted price. Planning accessories lines and planning the gifting calendar cover the core-and-seasonal split and the peaks.

Home & furniture: long consideration, event weekends and financing

Furniture is bought slowly. A sofa or a dining set is researched for weeks, compared across retailers and often bought on a date the customer chose in advance, which makes the promotional calendar part of the purchase decision. In the US, event weekends on public holidays — Presidents Day, Memorial Day, Labor Day — and the year-end period are when much of the category's event pricing runs, and a customer who is considering a sofa in the weeks before one of them has every reason to wait for it.

That produces the vertical's elasticity trap: pull-forward and push-back between events. The weeks before an event are depressed by customers waiting for it, the event weekend is inflated by them, and the lift measured against the depressed weeks is far larger than the demand the event created. The honest baseline is the run rate across the full cycle between events, not the weeks immediately before one, and the honest measure of an event is the cycle's total against a cycle without it, which is rarely available, so the reading carries a caveat.

Two levers are specific to the category. Financing offers — a promotional term on a credit plan rather than a cut in the ticket — move demand on high-ticket pieces without changing the price, and where the retailer pays the finance provider for the promotional term, that cost belongs in the event's margin test. Floor-model clearance is a reduction line that follows a floor set, sold as-is with no price history of its own. The supply side shapes all of it: SKUs with long lead times and container minimums mean an event cannot be chased with more stock, so an event that outsells its inventory records its stock, not its demand, and the response reads lower than it was. OTB planning for home goods and merchandise planning for home and furniture brands cover the container and the calendar.

Outdoor: weather-driven demand, prebook pricing and MAP

Outdoor's demand is set by the weather as much as by the price. An insulated jacket, a snow-sport shell or a sleeping bag rated for the cold sells when the cold arrives, and a fall that runs warm delays the category by weeks. A markdown taken in a warm week reads as inelastic, because the customer was not yet shopping for insulation at any price; the same markdown two weeks later, after the first cold spell, can read as strongly elastic. Estimates for weather-driven classes are only usable with weather in the baseline — the same weeks in a year with a similar temperature pattern, or a control group of doors in a region where the weather did not change.

The price is set once a year. The dealer price list goes out with the line, dealers prebook against it, and moving the price mid-year reprices goods already on dealer floors, so in-season price variation on current product is small. Where dealer-protected products sit under a MAP policy, deep promotion is removed as a lever while the model is current, and the price history that exists concentrates in the model-year closeout, the at-once residual after the season and pro-purchase programs where the brand runs one. MAP moves the price question from markdown depth to closeout timing: when the closeout opens matters more to the units it moves than how deep it goes, and a closeout opened before the season's demand has played out sells current-season units at closeout prices.

Merchandise planning for outdoor brands covers the prebook and the weather risk, and demand forecasting by vertical covers building weather into a baseline.

Sporting goods: by-sport seasonality, launch pricing and prebook

Sporting goods run a separate season for each sport: bats and gloves into spring, football into a late-summer window, basketball through the winter, ski and snowboard into a winter that may start late. Demand rises into a sport's season and falls away after it at any price, so a discount in the last weeks of a sport's season reaches few customers, and one in the pre-season weeks reaches customers who were about to buy anyway. Both read wrongly as elasticity: the end-of-season event as low response, the pre-season event as high. The baseline has to come from the same weeks of the same sport's calendar, and an elasticity estimated in one sport's season does not transfer to another sport's.

Within a sport, the ladder is a flagship at a launch price with the previous flagship and lower-spec models stepping down beneath it. The launch resets the ladder: the outgoing flagship takes the next rung down or goes to closeout, and its response at the lower price is partly customers choosing it over the new flagship — substitution inside the brand, not new demand.

Team business is a different market inside the same brand: uniforms and team equipment sold through team dealers at quoted team prices in order windows tied to the sport's season, with little price-driven variation because the order is set by the roster. Consumables — balls, grips, strings — sit at the other end, with stable prices, continuous replenishment and the most conventional promotional behavior in the category, including stock-up before a season that shows as a post-promotion dip after it. Merchandise planning for sporting goods brands covers the by-sport calendar.

Health & beauty: retail partner calendars, gift sets and the shade range

Beauty's promotional calendar is often not the brand's. In mass, the retailer sets its own price points and runs its own promotional events, and the brand participates through trade promotion: temporary price reductions, feature and display, off-invoice allowances and scan-based funding that sit between gross and net sales. The lift a brand sees in a retailer's event is a mix of its own price cut, the retailer's feature and display, and the traffic the event itself brought, and separating the price effect from the display effect needs events that varied one without the other. In prestige, a brand that holds one price across every door and avoids public discounting promotes through gift-with-purchase and value sets instead, and its price history is thin by design.

Holiday gift sets are the category's largest promotional instrument: a set of full-size or travel-size products at a price below the sum of its parts, made once for the season. Its margin test uses the set price as the discounted price, and its sell-through describes the gifting customer, not the everyday franchise.

Two traps are specific to beauty. Pantry loading on consumables — cleanser, shampoo, a daily moisturizer — means a promotion pulls forward months of a loyal customer's usage, so the post-promotion dip can run far longer than in apparel, and a lift read only in the event weeks overstates the event badly. And one price across the shade range means a franchise's response depends on which shades were in stock: a customer whose shade is missing does not substitute to the next one, so a shade stock-out during an event is lost demand that reads as weak response. Stock near its sell-by limit under retailers' dating rules is a clearance problem rather than a price one — see dating rules and weeks of supply; period-after-opening (PAO) governs use once the pack is opened and plays no part in a promotion's timing. Merchandise planning for health and beauty brands covers the shade ladder and the retailer calendar.

Toys & games: the gifting peak and the licensed window

Toys concentrate the year into the fourth-quarter gifting peak, and the retailer's promotional calendar runs on top of it. The peak is not a price response: units rise through November and December at full price as well as on promotion, so an event inside the peak measured against a pre-peak baseline credits the price with the season. The baseline for a peak event is the same weeks in a year without it, or comparable items at full price in the same weeks. Post-peak clearance reads as low elasticity because the gift-giver has already bought.

The retailer usually sets the price. Where a retailer plans its toy aisle to a set of retail price points, the brand designs to one, and promotional depth is negotiated with the account and often funded by markdown allowances in its terms.

The second demand driver is the licensed window. A product tied to a film, a series or a game sells against the property's release and the attention around it, and the window has a start, a peak and an end. A price cut in the weeks after the property's attention has faded recovers little, not because the customer is insensitive to price but because the reason to buy has gone; a cut during the peak of attention moves units that would mostly have sold anyway. Elasticity inside a licensed window is a function of where in the window the price moved, and a coefficient read from one license does not transfer to the next. The sell-off date in the license contract is the hard stop that turns the end of the window into a clearance deadline. Planning a licensed product window, planning the gifting calendar and merchandise planning for toy and game brands cover the windows.

Baby & juvenile: the due date and the registry

Juvenile demand is timed by the baby, not by the price. A car seat, a stroller or a crib is needed by a due date, and the purchase happens in a window before it whether or not a promotion is running. Price moves the purchase within that window and between products, but rarely moves it out of the window, so a promotion's lift is largely timing and substitution: a parent who was going to buy in six weeks buys now, or buys the promoted model instead of the one beside it.

The registry changes who decides. A registrant chooses items and the givers buy them, often at the price shown on the registry at the moment they buy. That gives the category a price behavior no other vertical has: the registrant's choice responds to price and features, while many givers buy what is listed. Registry retailers commonly offer a completion discount to the registrant on items still unbought as the arrival date approaches, which is a planned reduction rather than a promotion, and its units come from the registry, not from new customers.

Juvenile hard goods are priced on certified platforms with feature tiers on top, sold separately and in travel-system bundles below the sum of their parts; the bundle discount belongs in the margin-neutral test like a gift set. Model-year and pattern changeovers are the category's main source of price variation, and where dealers sell under a MAP policy the closeout is bounded as it is in outdoor. Planning with registry demand and merchandise planning for baby and juvenile brands cover the registry and the platforms.

Jewelry & watches: the metal cost base and perceived-value pricing

Fine jewelry has a cost base that moves with the market: gold, platinum and silver priced by weight and purity, plus stones, bench labor and findings. So the price question in jewelry runs both ways. When metal rises, the ticket either rises or the margin falls, and a price rise driven by metal is an elasticity bet made without much history, because a brand that reprices rarely has few observations of what a rise does to units. The repricing policy — a trigger band, metal priced separately, or absorption — is covered in markup and pricing by vertical and planning margin on a moving cost base; the elasticity question is how far units fall at each rung when the ticket moves.

The demand side is shaped by perceived value. A piece of fine jewelry is bought as a gift, a commitment or an heirloom, and its price is part of what it says. A deep public discount can lower the value a customer places on the piece, and on pieces customers already own, so the category's response to price cuts does not follow the textbook curve: a discount can reach fewer customers than expected because it reads as a lower-quality piece. Fine jewelry exits through remount, melt and consolidation more than markdown, so a brand's price history sits mostly in fashion and bridge lines. Gifting occasions — the year-end holidays, Valentine's Day, Mother's Day — and bridal concentrate demand, so the occasion effect comes out of any event reading.

Watches add a published price. Where a watch brand publishes its retail price and sells through authorized dealers, the dealer's planning variable is which references to hold and how deep, not what to charge; new references arrive at the brand's release, and a discontinued reference is the category's closeout. Merchandise planning for jewelry and watch brands covers the metal exposure and the dealer model.

How elasticity analysis goes wrong

Each of these failures produces a number that looks reasonable.

Reading a stocked-out week as low elasticity

A markdown or promotion week in which the core sizes, the popular shades or the event inventory ran out records what was available, not what was wanted. The response looks weak because the stock was gone, and the next event is set deeper to compensate. Screen for core in-stock first, and treat an event that sold out as a lower bound on its response.

Comparing markdown weeks with the weeks before them

Compared with the full-price weeks before it, the worked example's markdown read −0.84; compared with what those weeks would have sold at full price, it read −1.13. The season's decline in apparel, the weather in outdoor, the sport's season in sporting goods and the peak in toys all make a late markdown look weaker and a pre-season promotion stronger than it was.

Ignoring pull-forward

A promotion measured in its own week counts the customers who would have bought next week as new. In the illustration, gross lift was 100 units and net incremental was 30. Read promotions across the whole window, including the weeks after, and in categories with pantry loading or event-driven waiting — beauty consumables, furniture — the window is longer than it looks.

One coefficient for every tier

A category elasticity applied to every rung of a good-better-best ladder overstates the response of the best tier and understates the opening price point, and it hides the substitution between rungs. Store estimates by tier and channel, and test a promotion against the tier it runs on.

Treating a borrowed number as your own

A coefficient from another business describes that business's customer. Where a starting range is needed before your own history exists, label it an illustrative planning starting point, not a benchmark, and replace it once a few events have been read.

Where RetailNorthstar fits

RetailNorthstar is a merchandise planning platform — OTB, assortment and line planning, buy planning, allocation, sizing, PO and WIP tracking and analytics on a shared data model. As the market and performance intelligence page puts it, velocity, weeks of supply and sell-through rate are shown at the point of decision, not in a separate BI report.

Estimating elasticity, building the baselines, choosing markdown depths and deciding which promotions to run remain judgements that belong to the merchant and the planning team. OTB planning holds the receipt budget, which the planned reductions from a promotional calendar feed into, and margin planning sets out how OTB margin targets, buy depth and full-price allocation connect in one workflow.

Related resources

• →Markup and Pricing by Vertical — the first ticket and the margin it carries
• →Markdown and Exit Strategy by Vertical — the exit levers each vertical uses
• →Markdown Optimization in Retail and Dynamic Pricing for Fashion Retail — markdown timing and depth, and continuous pricing
• →Merchandise Planning KPIs by Vertical — the sell-through and margin measures elasticity is read against
• →Markdown Percent, Gross Margin Percent, Initial Markup and Maintained Markup formulas — the margin arithmetic
• →Price Elasticity, Markdown Optimization, Markdown Rate, Markdown Cadence and Promotional Planning — the terms