Why customer acquisition cost means something different for a $3 yogurt

A software company knows the exact minute a customer arrives: an email address, a card, a timestamp. Chobani knows none of that. Someone lifts a cup out of a chiller, pays a retailer that never passes on the name, and the brand hears about it weeks later as a rounded estimate in a panel report. The buyer signed nothing and can leave without cancelling anything. Every acquisition figure in this module sits on top of that gap, so the first job is to fix what "a customer" even means when the item costs $3 and gets bought twenty times a year.

The unit is a household, not a user

In FMCG (fast-moving consumer goods) the measurable buying unit is the household. One shopper buys for two, four or five people, and the yogurt gets eaten by people who never chose the brand and never will. Panels and loyalty files record the household, so that is what you pay to acquire and what you count.

The working metric is cost per incremental household: acquisition spend in a period divided by the households buying the brand in that period that were not buying it before. Two words carry the weight. "Household" fixes the denominator. "Incremental" strips out the households already in the franchise and the ones who would have bought anyway. A price promotion that moves 40,000 extra cups can add exactly zero households: same buyers, bigger basket, earlier trip.

What belongs in the numerator

• Paid mediaPaid mediaVisitors arriving via paid ads or sponsored placements, where you pay a platform to display your message rather than earning visits organically.View full definition →, including retail media at whatever return the shopper marketing lesson prices out.
• Trade promotion: the discounts, feature ads and display allowances a brand funds to change what the shopper meets at the fixture.
• Listing and slotting fees. Retailers charge to make room for a new SKU, quoted per item per chain, and a multi-chain US launch is routinely a six-figure entry cost before a single ad runs.
• Sampling: in-store demos, event trial, pack-with-purchase.

Distribution belongs here too, which is the part software-trained marketers miss. Chobani's early growth owed much to a decision about shelf rather than media: the product went into the mainstream dairy aisle next to Danone's Activia rather than into the natural-foods section, and physical presence in refrigerated distribution put it in front of new households at a scale no campaign could match. Store count is an acquisition line item.

Cost per incremental household =
    (Paid media + acquisition-weighted trade spend + listing fees + sampling)
    ÷ Net new buying households in the period

Worked example, using plausible category figures:

• Paid media: $400,000 per quarter
• Trade spend weighted to acquisition (not all trade spend acquires; much of it is loyalty pricing to existing buyers): $250,000
• Sampling and demos: $150,000
• Total: $800,000, against 200,000 net new buying households
• Cost per incremental household = $4

Against $1.00 to $1.20 of gross profit per cup, that $4 buys back in roughly four purchases. Whether four purchases is cheap or ruinous depends on the repeat-purchase-adjusted household value the lifetime valuelifetime valueLifetime Value: the total revenue (or profit) a customer generates throughout their entire relationship with your business.View full definition → lesson models, not on the margin of the first cup.

The incrementalityincrementalityThe share of results (sales, conversions, revenue) that only happened because of a marketing action, not what would have occurred anyway.View full definition → trap

The hard part is the denominator. Volume sold during a promotion is not volume acquired. Some of it is pantry loading by existing buyers, some is a trip pulled forward from next week, some is cannibalised from your own full-price sales. Without a baseline forecast, matched control stores or a held-out region, a brand counts everything that moves in promotion weeks as new and reports a flattering cost per household.

Deal-recruited households also behave differently from full-price triers. Chilled dairy is promoted hard, and when Danone and Chobani are discounting against each other in the same door, a share of the measured "new buyers" is the same price-driven households rotating between the two. They register as acquisition four times a year and belong to nobody.

The denominator you cannot count precisely

Household panels project from a sample. Nielsen, which sells the panel data brands use for exactly this count, reads tens of thousands of households and grosses them up to the population. For a brand bought by 0.5% of households, that is a few hundred panel homes, and the quarter-to-quarter wobble in your new-buyer count can be sampling noise rather than marketing. Small brands should read penetration on rolling twelve-week or annual windows and treat a single quarter's cost per household as directional. Retailer loyalty data is far larger and cleaner, but it sees one banner only: it cannot tell you the household that bought you somewhere else, or the one you lost to a competitor across the road.

Where the real-world numbers land

Reference points, all estimates that vary by sub-category, retailer and geography:

• US gross margins for packaged food brands: commonly cited at 25 to 45%, with dairy at the lower end and confectionery higher.
• Sampling cost per trial in the US: often put in the $2 to $8 range depending on venue and product, per industry sources such as the Path to Purchase Institute.
• Trade promotion spend as a share of gross revenue for US CPG brands: frequently estimated at 15 to 25% of gross sales, per Nielsen/NIQ category reporting.

Read those together and one thing falls out: trade spend usually dwarfs paid media in an FMCG budget, yet most acquisition costacquisition costCustomer Acquisition Cost (CAC) is the total sales and marketing spend divided by the number of new customers gained in a period. It measures how efficiently you grow.View full definition → calculations use the media line alone. That habit understates true cost by a multiple and makes every efficiency comparison meaningless.

🎬 [VIDEO: "How CPG Brands Calculate Customer Lifetime Value" - youtube.com/results?search_query=CPG+customer+lifetime+value+calculation - search for recent explainer content from marketing analytics channels covering repeat purchase modeling in consumer packaged goods]

Aldi, and the households you cannot buy at any price

Around 90% of what Aldi sells is own label, across an assortment of a couple of thousand items rather than the tens of thousands a conventional supermarket carries. For a branded yogurt with no listing there, those households are not expensive to acquire; they are unavailable. No media budget reaches a shopper standing in front of a shelf your product is not on.

The second-order effect is the one that catches finance by surprise. As discounters take category volume, the addressable base for branded acquisition shrinks, so the same media plan chases fewer reachable households and cost per incremental household rises with no change in creative, targeting or price. Teams read that as media fatigue and rebrief the agency. The honest reading is that a distribution decision, taken two years earlier or never taken at all, set the ceiling the media is working under.

The practical trap: cutting sampling because it looks expensive

Sampling almost always posts the highest cost per incremental household of any channel. It also removes the risk of an unfamiliar $3 purchase, which is the actual barrier in chilled dairy, where a bad cup gets remembered. A team ranking channels on acquisition cost alone will defund sampling and overfund paid social every single planning cycle.

The fix is to hold cost per incremental household next to the repeat rate of each channel's cohort, using the benchmark ranges the retention lesson supplies. A $6 sampling cohort repeating at 55% beats a $2 social cohort repeating at 18% well inside a year, and the gap widens after that.

Key Takeaways

• Acquisition in FMCG is priced per incremental household, not per signed-up user: the household is the buying unit, and "incremental" excludes existing buyers and anyone who would have bought without the spend.
• The numerator must include trade promotion, listing and slotting fees and sampling alongside media. Media-only maths understates real cost by a multiple.
• Promoted volume is not acquired volume. Without a baseline, control stores or a holdout, pantry loading and pulled-forward trips get counted as new households.
• Panel estimates carry sampling error. A low-penetration brand sits in a few hundred panel homes, so quarterly swings in new-buyer counts are often noise; loyalty data is bigger but blind outside its own banner.
• Distribution sets the ceiling. Where discounters like Aldi run mostly own label, the addressable base shrinks and cost per incremental household climbs even when the marketing has not changed.