RSi and the expensive art of the empty shelf

The empty shelf is a remarkably efficient critic. It ignores the packaging, declines the advertising and dismisses the forecast. A shopper reaches for a product, finds a gap, and makes another choice. Somewhere else, a spreadsheet may still insist that everything is perfectly in stock.

That disagreement is the territory of RSi, the retail analytics business now associated with Circana. Its work begins with sales and inventory records, but the interesting question is physical: can someone actually buy the thing? A carton in the back room, a damaged package or a misplaced item may satisfy an inventory count while disappointing a customer.

01 / A sale lost between the back room and the basket

Inventory and availability sound like twins. They have rather different manners. Inventory records how much product should be present. Availability concerns whether that product is accessible and purchasable at the shelf. The distinction becomes expensive when a brand spends money creating demand for goods that shoppers cannot find.

RSi’s response was to examine retail performance at the level of an item, in a store, on a particular day. Chain-wide totals can conceal a local problem: healthy sales elsewhere need not mean that a particular store is serving its customers. More granular reporting gives a supplier a place to investigate and a retailer a more precise question to answer.

The resulting business sits between the retailer that holds the operational records and the manufacturer that needs to understand what happens after delivery. It collects and organizes downstream data, then uses analytics to support replenishment, promotions and shelf execution. The software’s audience includes supply chain planners, sales teams and people visiting stores. They need different views of the same underlying problem.

02 / Before the shelf, there was the tag

RSi began in 2003 as T3Ci, with a focus on radio-frequency identification data. RFID offered a way to observe product movement through electronic tags. The early business therefore approached retail visibility through a particular technology, rather than through the broad collection of sales and inventory feeds that would later define it.

In 2007, T3Ci acquired VeriSign’s Retail Data Services business. The Retail Solutions name was reported in January 2008. This is the useful turn in the company’s history: the business widened its view from tag events to retailer data sharing. It could help interpret demand using records that retail operations already produced.

The lesson a reader can borrow is a modest one. A useful problem may survive a change in the technology used to solve it. Here, the enduring question was what happened to products close to the customer. RFID supplied one kind of evidence. Point-of-sale and inventory data supplied another, with a wider set of applications.

03 / The quiet business of agreeing on the numbers

Retail data sharing has a social history as well as a technical one. In a 2013 account, RSi co-founder Jonathan Golovin described how suppliers had begun pressing retailers to share point-of-sale information, sometimes proposing pilots to demonstrate the value. Manufacturers wanted earlier evidence of what was selling instead of discovering the answer after a promotion had already moved on.

Shared data does not automatically produce shared understanding. A supplier needs a consistent view across retail partners, while the retailer needs analysis relevant to its own stores. RSi’s expertise includes the unglamorous work of collecting, cleansing and harmonizing records so that operational reporting can become usable. The polished chart is the end of that work, rather than its beginning.

Its commercial model reflects that role: enterprise software delivered as a service, supported by analytics and customer expertise. Historical customers named in contemporary coverage include Nestlé, Procter & Gamble, Unilever and Clorox. These are companies with products spread across many stores and reasons to care about execution at individual locations.

“Suppliers want real-time data. Retailers want strategic insights.”Jonathan Golovin, RSi founder, 2013

Software development was central to the company’s identity. Co-founder Shantha Mohan ran its global product-development team. In her 2016 retrospective, she described founders who challenged each other and engineers working closely with other functions. It is a revealing detail: a retail analytics product must survive contact with operations, customer service and commercial needs, as well as with a database.

04 / Advertising discovers the stockroom

In June 2015, RSi launched Ansa. It applied daily store-level sales and inventory data to digital campaign planning and measurement. Brand marketers and agencies could connect online activity to what happened at physical checkouts, and assess performance at a finer geographic level.

The connection is appealingly practical. An advertisement can attract attention to a store without ensuring that the advertised product is available there. Bringing inventory into campaign analysis gives a marketer another explanation for weak results and another consideration when deciding where to spend. A creative idea cannot persuade an absent box to appear.

Ansa arrived through advertising and media partners, including MediaMath, MaxPoint and Valassis. That distribution put retail intelligence inside workflows used by media buyers. It also gave RSi another use for its existing data foundation: the same store records could inform both supply chain decisions and campaign analysis.

05 / The dashboard must eventually meet a person

RSi’s January 2020 OSA End-to-End launch brought together predictive and corrective alerts, inventory management, promotional forecasting and field execution. The stated aim was to help teams concentrate on higher-value store actions. In October, IRI acquired the company, planning to connect RSi’s supply chain capabilities with its Liquid Data platform.

The scale at that moment matters, provided it keeps its date attached. The acquisition announcement reported more than 500 global CPG customers and retailer data spanning over 300,000 stores. These describe RSi in 2020. Circana’s broader footprint today belongs to the combined parent business.

IRI and NPD merged in August 2022 and adopted the Circana name in March 2023. Today, Circana markets On-Shelf Availability and Liquid Supply Chain offerings. The former describes daily store-item alerts, mobile-capable prioritization and scorecards. The latter emphasizes harmonized visibility across retailers. The current product names belong to Circana; RSi’s older portfolio explains the business that entered it.

06 / The economics of doing something about it

RSi’s growth required capital. Contemporary reporting puts its September 2010 Series C at $15.2 million, led by Investor Growth Capital, with SAP Ventures, Venrock, Bessemer Venture Partners and Red Rock Ventures participating. The stated uses included international expansion and acquisition opportunities.

For a buyer, the more relevant cost is the whole operating commitment. Evaluate the software fee alongside data access, integration, training and the people who will respond to alerts. Then compare those costs with the value of recoverable sales and less wasted effort. This is a purchasing framework, not a quoted RSi price or a promise of returns.

Competition makes the buying questions sharper. NielsenIQ offers an OSA Barometer that also analyzes store-level sales to identify availability issues. Retailer portals and internal business-intelligence tools provide other ways to inspect data. RSi’s distinctive proposition was the combination of multi-retailer data preparation, retail-specific analysis and execution workflows. Whether that combination is valuable depends on the retailer coverage and operational support a particular buyer needs.

07 / Tomorrow morning has its advantages

Circana’s December 2025 discussion of shelf visibility makes a useful concession: daily updates available the next morning may be more practical than minute-by-minute information. Staffing constraints and product availability limit how quickly a store can respond. A faster feed cannot create a delivery or an employee.

That observation gives the reader a way to copy the approach. Begin with one retailer and a defined group of products. Agree on the data, assign responsibility for responding, and review whether the interventions improve availability. Keep the comparison honest: promotions, seasonal demand and changing distribution can all complicate a sales result.

The conditions matter. Poor records weaken the signal. A retailer unwilling to share data limits visibility. A team with no authority or time to correct a problem leaves the analysis stranded. RSi’s story rewards an ordinary question that software buyers sometimes postpone: who will do the work once the system tells us what is wrong?

The answer ends where the shopper begins. There is a product, a shelf and a moment to buy. All the data in the world must eventually help those three keep their appointment.