How Does a Promotion Show up in Syndicated Data?
Syndicated data sees a promotion as three retailer-executed tactics: a temporary shelf price reduction, a retailer feature, and a secondary display. The provider flags each store-week with the condition it observed and reports every sales fact under that condition. Anything the manufacturer does outside the store is invisible.
That framing decides what promo analysis can and cannot measure, so it is worth getting exact before any lift is computed.
The three tactics, as the providers define them
NielsenIQ describes its retail measurement as looking at promotions in three buckets: discount (temporary price reduction), feature (the store circular), and display (sometimes called secondary display) [1]. SPINS lists the same three plus special packs, and names TPR as the most common [2]. The CPG Data Tip Sheet, the practitioner reference most analysts learned from, explains why the providers stop there: these are the activities the retailer controls, they are short in duration, and they are visible to any shopper who walks in, which makes them measurable by the automated processing that builds a tracking database [3].
Each has a definition tighter than the everyday word.
A temporary price reduction counts only when the promoted price is more than 5% below the regular price. If the lower price persists, the provider treats it as the new regular price and stops counting it as a promotion. NielsenIQ puts that reset at 7 weeks [1]; the Tip Sheet puts Circana's (formerly IRI's) at 6 [3]. So a "promoted" price of 2 for $5 that runs week in, week out is reported as a $2.50 regular price, which is also how the shopper experiences it.
A feature is retailer print: the circular, the newspaper insert, the mailed coupon book. Manufacturer coupons such as FSIs are not features [3]. NielsenIQ describes it as inclusion in the retailer's in-store circular or a newspaper insert [1].
A display must sit in a secondary selling location and hold product the shopper can pick up and buy. Nielsen adds that it must be temporary; Circana also counts permanent displays [3]. Field auditors visiting a sample of stores each week record both the location and the selling price of displayed items [3], which is why display coverage in the data can look thinner than what your broker photographed.
Three tactics become six conditions and hundreds of measures
Because the tactics combine, most databases carry at least six merchandising conditions: any promo, no promo, feature and display, feature only, display only, and TPR only. A "quality merch" cut groups the three that involve a feature or display and excludes price-only weeks [3]. The overview page of this guide tabulates all six with the misread each one invites.
Then the multiplication happens. Every fact is reported for every condition, so "dollar sales" becomes dollar sales any promo, dollar sales no promo, dollar sales feature and display, and so on down the list. The Tip Sheet notes that a database can easily carry hundreds of promotion measures this way [3]. The volume is the reason promo work intimidates people; the structure is simple once you see that it is a handful of facts crossed with a handful of conditions.
Two mechanics to carry into every read. First, the reported percent discount is (base price minus promoted price) divided by base price, computed on the stores in that condition, and it is an average per unit: a buy-one-get-one appears as a 50% discount [3]. Second, "merchandised" and "promoted" are synonyms in Nielsen, Circana, and SPINS data; both refer to those three retailer tactics and nothing else [3].
The promoted price is a ladder
Promoted price splits by condition, and the split is consistent enough to be a rule of thumb. The Tip Sheet's worked case shows a promoted price of $2.58 that was really $2.62 on TPR-only weeks and $2.22 on quality-support weeks [4]. The quality price is almost always lower than the TPR-only price, because a lower price is what buys the feature or the display. Feature and display together is always lowest. Display only tends to run higher than feature only, because some retailers display at full price the week after a feature to sell through what they ordered [4].
Two further practices follow from that source. Analyze promotion at the promoted product group level, the set of items the retailer prices and features together, rather than at the individual UPC, because that is what the shopper saw on the shelf and how manufacturers plan events [4]. And read promoted price at the retailer level: national promoted pricing is useful for strategy and of limited value for understanding what happened, and some retailers, Walmart being the standard example, run promotions that either persist long enough to become base price or are features at full price, so their promoted price barely differs from their average price [4].
The taxonomies differ by provider
The six conditions above are the common core, but the providers do not name or nest them identically. Sous's Nielsen rulebook, built from Nielsen's own facts guide, recognizes ten promo conditions, including "feature or display," "price reduction," and a "subsidized" cut. Its Circana rulebook treats causal (merchandising) conditions as a separate dimension that any measure can be sliced by, rather than as part of the measure's name, because that is how Circana's measures guide lays them out. Those are Sous's encodings of each provider's conventions, kept separate on purpose. The practical consequence for a brand running two providers is that a promo query written for one does not translate word for word to the other, and a tool that assumes it does will return a number without a warning.
There is one arithmetic rule the taxonomies share. The "only" conditions are exclusive of each other at the store-week, and "any promo" is their union. Summing feature only, display only, and feature and display, then adding any promo on top, counts the same store-weeks twice. Sous handles this by asking you to pick one promo condition before a promo query runs, and recommends any promo as the default when the question does not name a tactic. The aggregation trap chapter of the Syndicated Data guide treats this as one of the four ways syndicated math breaks.
What the data does not see
| Visible in syndicated promo data | Not visible |
|---|---|
| Shelf price reduction of more than 5% below regular | Manufacturer coupons (FSIs, Catalina, digital) [3] |
| Retailer circular, insert, or mailer | Manufacturer bonus volume or bonus packs [3] |
| Secondary shoppable display recorded by an auditor | Retail media and other advertising [2] |
| Promoted price by condition, at PPG and retailer level | Loyalty or card-holder pricing, unless it changes the shelf price |
| Base and incremental, modeled by the provider | The trade dollars the brand actually paid (that lives in your own files; see Chapter 3) |
The right-hand column is why the return layer cannot be computed from the syndicated file alone. Chapter 3 covers where that cost lives and how the deal type changes who bears the risk. The what syndicated data misses chapter covers the broader blind spots.
Common questions
What counts as a promotion in SPINS, Circana, or Nielsen data? A shelf price more than 5% below regular, a retailer feature, or a secondary display, each executed by the retailer. Combinations of the three produce the merchandising conditions.
When does a TPR stop being a TPR? When the reduced price has persisted long enough for the provider to call it the new regular price: 7 weeks per NielsenIQ, 6 weeks per the Tip Sheet's account of Circana.
Is "merchandised" the same as "promoted"? In syndicated data, yes. Both mean one or more of the three retailer tactics.
Why does my display coverage look low? Display is recorded by field auditors visiting a sample of stores, and Nielsen counts only temporary displays. A permanent fixture your broker photographed may not be a display in the data.
Why is my promoted price different under different tactics? Because it is a ladder. Feature and display carries the lowest price, TPR-only the highest, and the blended "promoted price" sits somewhere between depending on mix.
Can I add the feature-only, display-only, and any-promo numbers together? No. Any promo already contains the others. Pick one condition per question.