Circana vs NielsenIQ vs Store-Level: Bev-Alc Data Buyer's Guide

Somebody on your team has a data budget and three browser tabs open: Circana, NielsenIQ, and some flavour of store-level or shelf-audit vendor. The question on the whiteboard is which one to buy.

That framing costs beverage-alcohol brands real money every renewal cycle, because these are not substitutes. They measure different events, at different grain, arriving on different clocks. Ask one of them a question it structurally cannot answer and you’ll spend a year arguing about a number instead of acting on it.

We’re gmware, a software and data engineering firm in Austin, TX, with delivery centers in Bangalore and Mohali, India. We build the warehouses brands land all of this into, and we run Shield Suite, our own retail-intelligence platform covering 60,000+ storefronts. So we have a horse in this race, and we’ll say up front where our horse doesn’t belong: if you need national category share in measured channels, buy syndicated measurement. Nothing we sell replaces it.

So here’s the version we’d hand a brand analyst who has to defend a data purchase to a CFO who wants to know why one line item isn’t enough.

”Versus” is the wrong frame

Every layer above sits at a different point in the chain, and each one goes blind exactly where the next one starts.

Syndicated measurement reads the register in the outlets it measures. That’s the only layer that tells you a human being actually bought your product, and it’s the only layer that shows you the whole category around you rather than just your own portfolio. Shipment data reads the distributor’s dock, which is earlier and broader but stops before the sale. Storefront observation reads the shelf, which is the only place where price, facings, voids and display execution physically exist.

A brand that owns all three can answer “why.” A brand that owns only the first two can usually only answer “what,” a few weeks late. That’s the actual decision in front of you, and it isn’t a vendor bake-off.

If you searched “IRI vs Nielsen,” here’s what happened to both names

A lot of people evaluating this category are comparing two brands that no longer trade under the names they typed. That turns every historical deck into a small archaeology problem, so it’s worth twenty seconds of lineage before anything else.

IRI was Information Resources, Inc., founded in 1979 to tie consumer behaviour to supermarket purchases. It merged with The NPD Group, whose own roots run back to National Purchase Diary Panel Inc. in 1966 (company history). Grocery Dive’s coverage of the rebrand describes the merger as finalized in August 2022, with the Circana name announced in a press release dated 7 March 2023. So IRI is not a rival to Circana. IRI is the larger half of Circana’s parentage, and a legacy IRI contract is a Circana contract.

The Nielsen half is the one that actually causes arguments in meetings, because two different companies get called Nielsen in conversation. The Nielsen at nielsen.com, the one most people picture, now describes itself as “a global leader in audience measurement, data and analytics, shaping the future of media”, and its solutions list runs TV and streaming, audio, digital and cross-media measurement. Nothing on that page is retail scan data or CPG sales tracking. The retail measurement business went the other direction: NielsenIQ announced its independence on 8 March 2021 with the closing of its acquisition by Advent International, and it now brands itself NIQ.

So in a beverage-alcohol context, “IRI vs Nielsen” means Circana versus NIQ. If someone on your team says “let’s pull Nielsen,” ask which Nielsen. One of them cannot answer a shelf question at all, and the other can’t tell you who watched the ad.

One caution for whoever maintains your internal vendor list. NIQ’s own about page today names no owner and points at a separate investor-relations site, so don’t hard-code a 2021 ownership fact into a document you’ll still be reading in 2028. Corporate structure in this category moves faster than the data does.

Who each of these vendors actually is

Marketing pages won’t tell you the structural thing you need to know, which is where each company’s underlying data comes from. That’s what determines its blind spots.

Circana is the combination of Information Resources, Inc. and The NPD Group. The two completed their merger in August 2022 and announced the Circana name in March 2023, with reach at the time described as nearly 7,000 brands and retailers across 20 countries and 500,000 stores. Its core US measurement geography is MULO+, which Circana describes as covering grocery, drug, mass market, military commissaries, club and dollar stores plus e-commerce. The January 2024 MULO+ expansion added eleven retailers and, in the trade write-up of the same release, was pitched as delivering 90% of CPG dollar sales from census point-of-sale data. Read that channel list twice if you sell spirits, because the channel you care most about isn’t named in it. We’re not calling that a gap. We’re saying it writes your first vendor question for you.

NielsenIQ, now branded NIQ, became a standalone business on 8 March 2021 when Advent International’s acquisition of Nielsen’s Global Connect unit closed. It then completed a strategic combination with GfK, announced 11 July 2023. For bev-alc buyers the more interesting recent move is on the liquor channel specifically: NIQ says it launched Total US Liquor Open State in July 2024 and that it expanded off-premise coverage by $21.5B, taking off-premise sales reporting past $111B across retail, liquor, convenience and other outlets. Two things follow from that. Liquor-store coverage in syndicated measurement is being actively built out, and any coverage claim you heard three years ago is stale.

NABCA is the trade association for control jurisdictions, and its data exists because in those markets the state itself is the wholesaler, so the state has the records. Worth knowing before you argue with anyone about the counts: NABCA’s own control-state directory lists eighteen jurisdictions with entries while its intro text says seventeen, and individual state pages disagree with each other. Collectively those jurisdictions are described there as roughly 24.7% of US population. We go deeper on that in our control-state data guide.

SipSource is the Wine & Spirits Wholesalers of America’s aggregation of distributor depletions. WSWA describes it as built from single-product transactions between wholesalers and retailers, representing over 70% of wholesale products by volume across all 50 states, contributed by member distribution companies, with VIP doing the harmonization. WSWA’s own page quotes different SKU and outlet counts in different sections, which tells you something useful about how these aggregations get assembled.

VIP is Vermont Information Processing, the software company a large share of US beverage distributors run on. Its iDIG product is the reporting and analytics layer, and it describes combining execution results, retail scan data and shipment trends with market, territory, chain, account and SKU-level views. If you already get depletion reports, they probably arrive through this ecosystem. We wrote the long version in our depletion data field guide.

Who else operates in this space, and what each one reads

Search for competitors or alternatives here and you get a directory listing, which is worse than useless because it implies the names are interchangeable. They aren’t. Several of them are reading a different instrument, and a couple of them aren’t reading an instrument at all. Here’s the honest map.

CGA by NIQ reads bars and restaurants. NIQ’s beverage-alcohol page now carries the line “CGA is now NIQ” above a login for existing CGA clients, and describes Full View™ as integrating “on-premise (bar/restaurant), off-premise (in-store), online, and direct-to-consumer sales data”. If the question keeping you up is on-premise, that’s your instrument. It isn’t an alternative to an off-premise shelf read. It’s the other half of the room.

IWSR reads the market rather than the store. It calls itself “the global leader in data, analytics and insights for the beverage alcohol industry” and names Global Database, Global Forecast Suite, Bevtrac, US Navigator, On-trade value and Radius Innovation Tracker. Nothing on that page states a store- or outlet-level grain, and the only channel it names outright is on-trade. Unpublished isn’t the same as unavailable, so ask. But the altitude reads as country and market, which makes it a planning and forecasting source, not an execution one.

SPINS organises itself around channel alignment. It describes itself as “the only place to align natural, regional, conventional, and specialty retail data into a singular view” and credits that to “exclusive retail partnerships.” Its data platform is listed as Point-of-Sale Data, Product Intelligence and Consumer Panel Data, alongside SPINS Ignite, Liftoff, Omni and the TriLens Panel. One thing to notice before you shortlist it for beverage alcohol: its own front page names no alcohol, beer, wine or spirits category, and the categories it does surface are beauty, body care and snacks. Treat that as a question to put in writing rather than a verdict. It’s the same question you should be asking Circana and NIQ anyway.

Fintech reads invoices. Its named products are PaymentSource® and Scan-Based Trading, it describes itself as an alcohol invoice data and payments platform, and it sells suppliers “exclusive alcohol scan data insights to analyze competitor activity, spot emerging trends, and prioritize opportunities”. Its published component figures are 317,000 retail and hospitality businesses served and 9,200 connected distributors, vendors and suppliers. The same page also shows a rounded 326,000 total that doesn’t reconcile with those two, so quote the components. The boundary worth understanding: an invoice records what the retailer paid the distributor, and scan records realized price after promotion. The retailer’s own margin decision sits between those two numbers, and neither of them is the tag on the shelf.

Crisp doesn’t measure anything, which is exactly why it belongs on this list. Crisp Inc. calls itself “the leading vertical AI company for retail data”. It says it “connects with all of your retailer data sources” including retailer portals, syndicated feeds, ERPs and EDI, and that it “creates a universal schema for products and stores across retailer and syndicated data sets.” That is plumbing, and good plumbing earns its money. But it standardizes feeds you already hold rights to. It cannot originate an observation nobody sent it.

Dimensional Insight occupies the same slot from the BI side. Dimensional Insight, Inc. sells Diver Platform® with ProDiver, DivePort, DiveTab and Measure Factory. It describes that as “a governed analytics platform that unifies data, business rules, and analytics into a single, trusted foundation,” and runs a beverage-alcohol vertical for tracking “production, inventory, depletions, pricing, and sales performance across brands, distributors, and territories.” Here’s the concrete reason a control-state buyer should care. NABCA’s dataset overview says SAM access “is available through ProDiver, DivePort and Standard Data Sets,” and ProDiver and DivePort are Dimensional Insight products. If you subscribe to control-state data you may already be reading a Diver screen without knowing whose software it is.

Two things fall out of that map. The word “competitor” is carrying weight it hasn’t earned, because most of those names read a different instrument and two of them read none. And your shortlist isn’t decided by the vendor list at all. It’s decided by which instrument answers the question your board asked you last month.

Seven questions that decide the contract

This is the part most evaluations skip, and it’s where the regret comes from. Every one of these should be answered in writing by the vendor before signature, and every one of them applies equally to us.

Grain deserves an example, because it’s the one people nod along to and then get wrong. Say a market read shows your lead SKU down six points of share, and assume that market holds 900 stores carrying you. Six points could be 54 stores each softening slightly, or it could be twenty stores that stopped carrying you entirely while the rest held. Those are opposite problems with opposite fixes, and a market-level number cannot tell them apart. Neither can a chain-level rollup. Only store-grain data can, which is why void and coverage detail matters more than another decimal place of accuracy on the aggregate.

The re-use question is the sleeper. Brands routinely buy syndicated data intending to blend it with distributor depletions in their own warehouse, then discover mid-implementation that the contract restricts derived outputs or onward sharing. Ask about derived works, retention after termination, agency access and retailer-facing use, all four, separately.

Where syndicated measurement is simply the right buy

Plainly: if the question is national or regional category share, competitive brand performance across a measured channel, or anything a retailer’s category manager will accept as a shared version of truth, buy syndicated. Circana and NIQ exist because that job is genuinely hard, requires retailer participation you cannot assemble yourself, and requires consistent category definitions maintained over years.

A storefront layer is not a substitute for that and we won’t pretend otherwise. Observation from outside the register can tell you what is on a shelf, what it costs and whether the display is up. It cannot tell you units sold, and it cannot give you a category total that a buyer at a national chain will accept in a line review. If someone sells you shelf observation as a replacement for measurement, push back.

The same discipline runs the other way for on-premise. Bars and restaurants behave differently enough from stores that blending them without thinking produces confident nonsense, which we get into in our on-premise versus off-premise breakdown.

Where all three layers go quiet at once

Three gaps survive no matter how much of the above you own.

The first is the unmeasured storefront. A meaningful share of off-premise beverage-alcohol volume runs through independents, and coverage there is uneven across every source in this post. Depletions know product shipped to those accounts and stop. Syndicated measurement is expanding into liquor retail, as NIQ’s Total US Liquor Open State work shows, but the honest posture is to ask for the outlet list rather than assume.

The second is the actual price a consumer sees. Your list price, your suggested price and the number on the shelf tag in a store 1,400 miles away are three different numbers, and only one of them affects whether the bottle sells. Deriving price from measured dollars divided by units gives you an average across a promotional period, not the tag a shopper looked at on Saturday. That’s the whole reason we built Price to Consumer around observed shelf price rather than modelled price.

Then there’s execution, which is the one that costs the most and gets measured the least. You funded an endcap in 340 stores. Depletions confirm cases shipped. Syndicated data may show a lift or may not. Neither one tells you the display was ever built, and both will happily attribute a flat result to the creative rather than to a program that was never executed. Marketing Monitoring exists for that specific argument, with on-the-floor evidence store by store.

What this data costs, honestly

Nobody publishes a rate card. We read the public sites of NIQ, SPINS, IWSR, Fintech and Crisp while writing this, and not one of them lists a price anywhere. We don’t publish ours either. Any article that hands you a dollar range for syndicated beverage-alcohol measurement is guessing, and a guess is worse than nothing because it anchors your expectations before you ever get on a call.

There is exactly one place in this category with published commercial terms, and its value is the shape rather than the number. NABCA’s control-state dataset overview states that “One user ID is provided per subscription, and additional ID’s are priced at $100 per month per ID.” It adds that “All SAM subscriptions are quarter-to-quarter, billed quarterly in advance, with no long-term commitment required,” that “There is no charge for user training and hotline support,” and that if you “Subscribe to 10 markets and receive the remaining 8 at no charge.” Per market, per seat, short term, training bundled. That is the architecture of almost every quote you’ll receive here, including from vendors who won’t show you the arithmetic.

So instead of a number, here’s what moves the number.

Two of those rows are where the money actually leaks. Grain is the first. Store-level detail costs more than aggregates, and it’s the only version that tells you which stores to call, so buying aggregate to save budget buys you a report instead of an action. Delivery is the second, and it’s a commercial question wearing a technical costume. Flat-file or API access into your own warehouse is worth materially more to you than a dashboard login, and vendors price the two very differently.

Put these in the RFP in writing, because none of them show up in a first proposal unprompted.

1. The outlet universe for your states, by channel, with a count.
2. Whether store-level detail is included or an upcharge, and in which markets.
3. Release latency in days from event to availability, not the word “weekly.”
4. History depth included at signature, versus backfill priced separately.
5. Seat count, and the rate to add one mid-term.
6. Re-use rights: derived works, warehousing, retention after termination, agency and retailer-facing use.
7. Who maps their item codes to your SKUs, and at whose cost.
8. A defined expansion path, so adding markets in year two is a pre-agreed step rather than a renegotiation from zero.

Number seven is not a footnote. In the integrations we run, the connector is days of work and the reconciliation of product masters, account names and reporting calendars across a national footprint is months. Whoever you buy from, budget for the mapping.

One last thing, since you can’t compare published prices and a reference call is the closest thing to a review you’ll get. Ask for two references in your category and two in your states, then ask all four the same two questions. How long did the first ninety days really take, and who ended up doing the product-master mapping. Those answers separate vendors more reliably than any feature grid, and they’re free.

What to buy first, by scale

Sequence matters more than selection at every stage below.

If you’re emerging, buy nothing. Your distributors already give you depletion views, and a spreadsheet plus a person who calls accounts will outperform any platform at that size. Spending on syndicated measurement before you have distribution worth measuring is the most expensive mistake in this category.

If you’re mid-scale, meaning distribution has outgrown the phone call and you’re being asked to explain regional performance, add the shelf next. Not measurement. You already know roughly what shipped; what you don’t know is which stores are out of stock, where the price drifted, and which programs got built. That’s store-grain and it’s actionable this week.

If you’re national and sitting in line reviews with chain buyers, you need syndicated measurement, and the argument for it is partly political. You need a number the retailer also uses. Buy it, and layer storefront observation underneath so that when the measured number moves you can find the stores responsible.

What we’d recommend

Stop shortlisting and write down the decision first. If the decision is “which SKU do we cut in this category,” that’s syndicated. If it’s “why did this market fall,” that’s store-grain. If it’s “did the money we spent turn into a display,” that’s observation with evidence attached. The vendor falls out of the question almost automatically once the question is honest.

Between Circana and NIQ specifically, we’d stop treating it as a capability comparison. Run both through the seven-question table above with your own states and your own channels filled in, insist on the outlet universe in writing, and pilot two markets including one you already suspect is misreporting. Whichever answers that with less hedging is your vendor. That test is more informative than any feature grid, including this one.

Then look at what’s left over, because for most beverage-alcohol brands we talk to, the leftover is the entire storefront layer. They can describe their share to two decimal places and cannot tell you today’s shelf price in Dallas or whether last month’s display program was ever built.

Tell us which sources you already pay for and what question keeps going unanswered, and we’ll give you a straight answer on whether we’re the right layer to add or whether you should spend the money elsewhere. We do this work with beverage-alcohol brands and distributors every week, and sometimes the honest answer is that your existing contracts already cover it.