Promotion planning for category managers

Why promotion is a category decision, not an item one

How do you get your SKU a July feature at Sprouts? The honest answer is that you mostly do not, and the reason is what promotion planning actually is. When I asked for a July slot on Verde Fresca's chunky fresh salsa, the buyer pulled up her promo grid: my ad slot was already Casa Marisol's, the endcap behind it belonged to the store brand, and four salsa promo weeks were left in the quarter against eleven brands who wanted them. That grid is why promotion gets planned at the category level, not one SKU at a time. The ad features, the endcaps, the promo weeks are a fixed, shared resource, and the buyer allocates them to grow the entire $18.4M refrigerated salsa set, not to subsidize whichever supplier asked first.

I sat on the supplier side, so I felt this from the outside. My job was to make my item the best use of a slot the buyer was going to fill regardless. Her job, running category management, was to spend the calendar so total category dollars and blended margin came out ahead. Those are different jobs, and a promotion pitch that ignores hers loses. This chapter is Step 6d of the eight-step process: how the calendar gets built, how lift is measured against a baseline, and how a category manager keeps a promo from just moving share between two brands while the section as a whole goes nowhere.

Promo mechanics: the five instruments and what they cost

A category manager works with a small set of promotion instruments, and each one buys a different amount of lift for a different margin cost. The refrigerated salsa set at Sprouts uses all five over a year. Here is how they compare, with the everyday $4.99 mainstream fresh price as the reference point.

The lift ranges look generous until you separate the two kinds of volume hiding inside them. Some of a promo's extra units are incremental, genuinely new sales the category would not have made. The rest is pulled forward from the weeks on either side, the same shopper buying the same jar a week early because it was on deal, or traded over from the brand's own other SKU. A BOGO that reads +120% on the promoted week and then leaves a crater the week after mostly borrowed from the future. Separating incremental from pulled-forward is the entire measurement problem, and I wrote the method up in full in choosing a baseline for post-promo lift. The short version: the baseline you pick can swing the reported lift by 30 points on the same promo, so the number is meaningless until you name the baseline next to it.

Promotion planning: timing, cadence, and forward-buy risk

Once you know what each instrument costs, planning the calendar is mostly about when and how often, not just what. Four constraints shape every salsa calendar I have built.

Timing against a crowded week. Refrigerated salsa peaks in summer, and so does everyone's promo pressure. The weeks around Memorial Day, the Fourth of July, and Labor Day are wall-to-wall salsa features across every brand in the set. Running your deepest discount into that wall is the worst timing there is: the category is already moving on occasion demand, the ad slots are expensive and oversubscribed, and your feature competes with five others for the same eye. A discount that would earn real incremental lift in a quiet April week gets swallowed in the July crush, where it mostly funds volume the category was going to sell anyway.

Competitive weeks. If Casa Marisol has the endcap in week 27, that is a week to sit out, not to match. Two brands both cutting price in the same week on the same shelf trade units with each other and hand the category manager a margin bill with no incremental volume to show for it. The buyer sees that pattern immediately when she runs the recap.

Forward-buy and pantry-loading risk. Refrigerated salsa has a short shelf life, so pantry-loading is milder than it is for a shelf-stable broth at Kroger, but it is not zero. A deep, pre-announced TPR trains the weekly shopper to stock two jars and skip next week. That shows up as a trough right after the promo, and if you measure on the promoted week alone you never see the borrowed sales come back out.

Cadence. This is the one that most often wrecks a category. Promote the same SKU every third week and you teach the shopper to wait for the deal. Everyday velocity erodes, the reference price in the shopper's head becomes the promo price, and you have converted a healthy full-margin item into a permanent-discount item that only sells on ad. A routine category like salsa can support a promoted SKU maybe six to eight weeks a year before the everyday base starts to soften. Past that, you are not lifting the category, you are repricing it downward one promo at a time. That is why promotion planning has to coordinate with the everyday price ladder, covered in category pricing and price architecture: the promo price only means something relative to the everyday price it discounts from.

A worked example: Verde Fresca's summer TPR at Sprouts

Take a concrete plan. Verde Fresca's chunky fresh SKU sits at an everyday $4.99 and runs a TPR to a promo $3.99, a 20% cut, for one week in early July at all 345 Sprouts stores. Here is the read on that single SKU.

The +56% headline is the number a junior analyst puts in the deck. The +29% is the number the buyer actually cares about, because a little over half of the apparent lift is not new category volume at all. Six hundred units were the same shoppers buying a week early, and they come back out of the following week as a trough. Five hundred units were stolen from Verde Fresca's own second fresh SKU sitting two facings over, a within-brand cannibalization that nets the category nothing and costs Verde Fresca margin on both jars. Strip both out and the category gained about 1,200 real incremental units off a 20% price cut it paid on all 6,400 promoted units.

Now the category manager's question: is that a good trade for the whole set? At a $4.99 everyday price and a $1.00 promoted discount, Sprouts gave up roughly $6,400 in margin on the promoted week to gain 1,200 incremental units. Whether that clears depends on the category's margin on those incremental units and whether any of the 1,200 were new-to-category shoppers versus regulars who simply switched from a competitor. If the incremental volume mostly came at Casa Marisol's expense, the category netted close to zero and the buyer funded a share fight. If it pulled fresh trips into the salsa fixture, it was worth doing.

Two recommendations come out of this plan, and both are moves I would put in front of the buyer:

Move the deep discount off the July wall. That same 20% TPR run in a shoulder week in late April or September, when no one else is featuring salsa, earns more incremental per margin dollar because it is not competing for attention or funding occasion volume the category already had. Save the July slot for a shallow feature that rides the occasion rather than a deep cut that subsidizes it.

Protect the fermented line from discounting entirely. Verde Fresca's two fermented SKUs are growing +34% year over year inside a fermented segment that is up +28% for the whole category. That line is winning on its own demand at its $6.49 everyday price. Discounting a growing item does one thing: it hands back margin on volume that was already coming. There is no lift to buy when the base is already climbing, so a promo on the fermented SKUs is close to pure margin giveaway. Leave it at full price and spend the calendar on the flat fresh SKUs that actually need the push.

The category manager's lens: does the promo grow the set?

The test the buyer applies to every promo ask is simple and it is not the one suppliers pitch. She is not asking "will this SKU sell more." A discount always sells more of the discounted item. She is asking "does the category end up ahead." A promo that moves 1,200 units from Casa Marisol to Verde Fresca has done nothing for the $18.4M set: same total dollars, lower blended margin because someone got a discount, and a buyer who now has to explain to her boss why she funded a brand swap. She will not do it twice.

What clears her test is incremental category volume or a new category shopper. A feature that pulls a household who does not currently buy refrigerated salsa into the fixture grows the set. A display that lifts the grilling occasion so shoppers buy salsa and chips together on a trip they would not otherwise have made grows the set. A within-set price war does not. This is the same discipline the whole eight-step process runs on: the buyer's incentive is total category dollars and margin, and every tactic gets weighed against that scorecard. Promotion is just the tactic where the temptation to optimize for one brand is strongest, because the lift on the promoted item always looks great in isolation.

Anti-patterns that quietly cost the category

Three failure modes show up in almost every weak promo plan I have reviewed.

Promoting a KVI that only pulls demand forward. Known-value items, the SKUs shoppers price-check, feel like the right thing to promote because they move volume. But the routine shopper buying the chunky fresh salsa every week does not buy more salsa because it is 20% off; she buys the same salsa a week early and skips next week. The promoted week looks huge and the trailing weeks sag. Net incremental on a true KVI in a routine category is thin, and you paid full discount for volume that was already yours.

Measuring on raw promoted volume. The +56% headline in the worked example is the trap. If the recap deck reports promoted-week units against a baseline that already includes pull-forward, or against no stated baseline at all, the promo looks like a hit every single time and the category manager over-invests in the next one. The fix is measuring incremental lift over a clean baseline, dropping the immediate pre-promo week when pull-forward is plausible, which is exactly the method in the post-promo lift baseline chapter.

Over-promoting a growing item. This is the fermented-line mistake in general form. When a segment is already up +28%, a promo on it cannot separate its own effect from the growth that was coming anyway. Run a TPR on a +34% SKU and the promoted week will look great, but a trend-adjusted baseline shows most of that lift was organic. You gave away margin to accelerate something that did not need accelerating. Spend promotion on the flat and the soft, never on the winner.

Doing this in Scout

The half-day version of this is stitching the promo recap together by hand: pull the promoted SKU's weekly units out of the SPINS extract, build a baseline window, eyeball the trough, guess at the cannibalization from the sibling SKU, and hope the buyer agrees with your baseline choice. Scout's scenario explorer computes lift on a SPINS-extracted promo against several baselines at once, trailing N-week, year-over-year, pre-plus-post excluding halo, and a trend-adjusted view, so the incremental-versus-headline gap shows on one screen instead of four spreadsheet tabs. The pull-forward and destocking windows are settings, so the "real incremental" number in the worked example above is a configuration rather than a manual subtraction.

What Scout does not do is build your calendar or make the buyer's call. It assembles the evidence: which weeks are already crowded in the category, what a past promo actually netted after cannibalization, whether a growing SKU's lift was real or organic. The judgment about which slot to spend, which item to protect, and whether an incremental trade clears the margin target stays with the category manager. Scout gets you to the review with a defensible number instead of a raw promoted total.

The short version

• Promotion is planned at the category level because the calendar is a shared, fixed resource, and the buyer allocates ad features, endcaps, and promo weeks to grow the whole set, not to subsidize one brand.
• Every promo instrument, from TPR to BOGO, mixes real incremental volume with pulled-forward and cannibalized volume, so the number that matters is lift over a clean baseline, not raw promoted units.
• A promo that only shifts share between two brands does nothing for the category; move deep discounts off crowded summer weeks to quiet shoulder periods, and never discount a SKU that is already growing on its own.

Related: Category pricing and price architecture · Measuring post-promo lift