How to choose a retail analytics platform

"Retail and consumer analytics" covers genuinely different jobs under one label — market-share measurement, in-store physical behavior, a retailer's own loyalty data, on-shelf execution, and commerce-platform reporting. The starting question isn't which vendor is best, it's whose data you're analyzing: the whole market's, your own store's, or a specific retailer's shelf.

Syndicated market-share data: three providers, three data sources

Circana, NielsenIQ and Numerator all answer the same basic question — how is a brand or category performing across the market — but they build that answer from different raw data. Circana, formed from the 2022 merger of IRI and The NPD Group, and NielsenIQ both aggregate retailer point-of-sale and e-commerce feeds across many retailers into market-share, pricing and distribution measurement; NielsenIQ's coverage skews toward traditional CPG at global scale, while Circana's combined heritage spans both CPG and durable-goods categories, such as electronics, apparel and toys, that NielsenIQ covers less directly. Numerator takes a different approach entirely: instead of aggregating retailer-supplied POS feeds, it tracks a large panel of opted-in shoppers' actual purchase receipts, which adds demographic and behavioral segmentation — who bought it, and did an ad reach them first — that POS-based measurement can't see on its own.

If the question is "what's our market share this quarter," Circana or NielsenIQ is the more direct fit. If the question is "who's buying our product and why they switched," Numerator's panel data answers something POS aggregation structurally can't.

Your own data vs. everyone's data

dunnhumby is a different kind of vendor from the syndicated three above: it works with an individual retailer's own loyalty-card and transaction data — famously built Tesco's Clubcard analytics — to guide that retailer's pricing, promotion and personalization decisions, effectively acting on donor-level customer value rather than category averages. This is first-party, retailer-specific data science, not market-level benchmarking; a retailer engaging dunnhumby isn't trying to see the whole category, it's trying to act on its own customers' behavior. CPG brands that sell through a retailer using dunnhumby may also get access to that retailer's media and insight products, a separate relationship from buying Circana or NielsenIQ data directly.

Execution on the shelf vs. strategy in the market

RSi is more operationally scoped than the syndicated data providers: it collects point-of-sale and inventory data directly from a manufacturer's retail partners to flag day-to-day execution problems — out-of-stocks, on-shelf availability, promotion performance at a specific retailer — rather than reporting category-wide trends. A CPG category manager deciding on strategy wants Circana or NielsenIQ; a CPG sales team managing distribution and stock-outs across retail partners wants RSi.

Physical stores and commerce platforms

Two tools address retail operations rather than market or shelf data. RetailNext measures the physical store the way web analytics measures a site: sensors and cameras track foot traffic, dwell time, conversion and queue length, essentially a store-level funnel, plus asset-protection features using the same camera infrastructure. Aptos is different again — a full commerce platform covering point of sale, merchandising and order management, with sales and inventory reporting built in, so the analytics come bundled with the systems that actually run store and e-commerce transactions, rather than being layered on top of them.

SymphonyAI Retail rounds out the category from the planning side: machine learning applied to demand forecasting, price and promotion optimization, and supply chain or replenishment planning, aimed at automating decisions previously handled with simpler statistical methods, and typically deployed integrated with a retailer's existing ERP and POS systems rather than standing alone.

How pricing works

Every tool in this category is quote-only, sold as data subscriptions, analytics engagements or multi-year software contracts rather than self-serve signup — there's no published rate card anywhere here. Scope varies: the syndicated providers price by category and geographic coverage; RSi scopes by number of retail partners and categories covered; RetailNext prices per store and sensor deployed, since it requires physical hardware; Aptos and SymphonyAI Retail are sold as enterprise multi-year contracts. Budget for a sales conversation at every stage of this category.

A shortlist by situation

• CPG brand or retailer needing market share and category trends: Circana or NielsenIQ — check which one's category coverage, CPG or durable goods, fits your product.
• Need to understand who's buying and why, beyond POS totals: Numerator's opted-in shopper panel.
• Retailer wanting to act on your own loyalty data: dunnhumby.
• CPG manufacturer managing stock-outs and on-shelf execution: RSi.
• Understanding physical store traffic and conversion: RetailNext.
• Need a commerce platform with built-in analytics, not a standalone tool: Aptos.
• Automating demand forecasting, pricing and replenishment decisions: SymphonyAI Retail.

Questions to ask vendors

• Is the underlying data retailer-supplied POS, an opted-in consumer panel, or our own first-party data, and how does that affect what questions it can actually answer?
• What category coverage does this include, and does it match what we sell?
• For in-store sensor analytics: what's the hardware installation and maintenance commitment per store?
• Does pricing scale by category, retailer partner, store count, or a flat enterprise contract?
• Can we get store-level or SKU-level detail, or only aggregated category reporting?

Common mistakes

Buying syndicated market-share data to solve an execution problem, such as out-of-stocks at specific retailers, is a common mismatch; that's what RSi's retailer-specific feeds are built for, and Circana or NielsenIQ's category-level view won't show it. The reverse mistake is engaging a first-party specialist like dunnhumby before a retailer has enough of its own loyalty data to act on — that engagement depends on data maturity the retailer needs to have already. And treating panel-based data and POS-aggregated data as interchangeable is a real methodological difference, not just a vendor preference — check which one a specific claim in a report is actually built on before citing it.

See Circana vs NielsenIQ and dunnhumby vs Numerator for two common head-to-head decisions, and every tool in this category.