Summary. A store’s value is not contained in its own sales line. Opening one measurably lifts online sales in the surrounding area, and closing one measurably damages them. That spillover is created by what happens inside the store, which in practice means the person standing in it. This paper covers what the effect is worth, how to bring it into a store’s own numbers, and what it asks of an associate that a conventional sales role does not.

The store’s effect is bigger than the store

ICSC’s Halo Effect III study analyzed 69 retailers and 2,103 individual stores across 50 states and Washington DC between 2019 and 2022, drawing on roughly $848 billion in credit card transaction data and measuring the thirteen weeks following each opening or closure.¹ The findings run in both directions.

Opening a store	Online sales in the surrounding trade area rise 6.9 percent on average in the following weeks
Opening a store, emerging DTC brand	13.9 percent
Closing a store	Sales fall 11.5 percent
Closing an apparel store	19.4 percent decline
Closing a home store	32.2 percent decline

Read the closure rows as the more useful half. They say a store is contributing demand that does not appear in its own numbers, which means a location judged only on its four-wall sales is being judged on a fraction of what it does.

Note also that the losses on closure run larger than the gains on opening. A store builds its trade area gradually and takes it away at once, which is an argument for treating an underperforming location as a capability question before it becomes a closure decision.

What the halo is made of

A store lifts surrounding online sales because it does things a website cannot: it makes the brand physical, it lets someone handle the product, and it puts a person in front of the customer.

The first two are property and merchandising decisions. The third is the only one that varies day to day, store to store, and shift to shift. It is also the only channel a brand owns that talks back. Everything else the customer encounters, from the campaign to the product page, is a broadcast. The associate is a conversation, and conversations are where a brand is either confirmed or quietly contradicted.

That is the argument for treating the associate as a brand role rather than a transaction role. Not because it sounds better in a deck, but because the spillover the store generates is produced in that interaction.

“The halo is a real estate effect, not a staffing one”

The reasonable objection is that the study measured openings and closures, which are property decisions, and that attributing the effect to associates is a leap. Signage, footfall and physical presence would generate awareness whether or not anyone inside was well briefed.

Three responses, and one concession.

Presence explains the average, not the variance. Every store in the sample had a sign and a location. If presence alone produced the halo, comparable stores in comparable catchments would produce comparable spillover, and any retailer who has looked knows they do not.

The mechanism has to run through an experience. A customer who walks in, is ignored, and leaves has been made aware of the brand and has also formed a view of it. Awareness with a poor impression attached is not a neutral outcome for the online sales that follow.

The controllable part is the only part worth discussing. Property and merchandising decisions are made annually by a small group. The interaction happens several hundred times a day in every location and is the one variable a business can change this quarter.

The concession, and it should be stated plainly: the ICSC research does not measure staff quality and cannot separate its contribution from the rest. The claim here is not that the study proves associates cause the halo. It is that the halo is produced inside the store, that the interaction is the largest thing inside the store that anyone controls, and that a business persuaded by the first table should be interested in the second point.

Valuing a store on what it actually contributes

The practical consequence of the halo research is that most store-level financials understate the location. Correcting that is a calculation, and it changes both closure decisions and capability budgets.

1. Take online revenue in the store’s trade area. Define the area consistently, by drive time or postal geography, and use the same definition everywhere.
2. Apply a halo share. The study’s averages give you a defensible starting point by category, and if you have opened or closed anything in the last three years you can estimate your own from that natural experiment.
3. Add it to four-wall sales to get a total contribution figure for the location.
4. Re-rank the estate on that basis. Stores move, sometimes sharply, and the ones that move most are usually smaller locations in dense areas that look marginal on their own line.

Two uses follow. The first is defensive: a closure list built on four-wall performance is systematically wrong, and the home and apparel figures above suggest expensively so. The second is the budget argument. If a store contributes materially more than its own sales, the return on making the people inside it better is larger than the store’s P&L implies, and capability spend should be sized against total contribution rather than against four-wall revenue.

What a brand ambassador needs that a salesperson does not

The gap between the two is specific, and it is mostly a content problem rather than a hiring problem.

• The story behind the product, not just its specification. Why the brand makes it this way, what it replaced, what it is for. A customer who already read the specification is asking for the part that is not written down.
• Language the brand actually uses. A fleet where every associate describes the same product differently does not have a brand problem in marketing, it has a distribution problem in training.
• Enough context to recommend against something. The most persuasive thing an associate can do is steer a customer away from a product that will not suit them. That takes comparative knowledge, which is exactly what generic product training omits.
• Permission to talk about the brand, not just the promotion. Heritage, materials, provenance and craft are what justify a price. They are also the first things cut from a training calendar dominated by this week’s offer.
• An answer for the customer who will buy later, online. A large share of the halo is exactly that person. An associate measured only on today’s transaction has no reason to serve them well, which is a targets problem rather than a training one.

Consistency at scale is a distribution problem

Any single store can be trained to do this. The difficulty is doing it in every location, in every language, with a workforce that turns over, at the speed products now change.

That reframes the requirement away from content quality and toward reach and timeliness. A brand story that is excellent and reaches half the fleet three weeks after launch is worse, commercially, than a good one that reaches everyone on day one. Consistency is what makes a brand recognizable across a market, and consistency is a delivery property.

It also has to survive local variation. A flagship and a suburban location share a brand and not a customer, and the practical answer is a common global message with room for regional layering, rather than either extreme of one rigid script or fully devolved content. That split has its own paper.

The trap in the halo data

The temptation with ICSC’s numbers is to treat the store as automatically valuable, which is not what the study says.

The measured effect is what happened when retailers opened stores. It is a statement about the average store opening, not a promise about any specific location, and it certainly does not imply that a store contributes the same halo regardless of how well it is run. A location where nobody can answer a question is still casting a halo, and the halo it casts is the brand impression it created.

The closure figures cut the same way. A brand that keeps a location open but under-invests in the people inside it has kept the cost and kept only part of the benefit.

What to measure

1. Trade-area online revenue by store, tracked monthly. Without this the halo stays a study rather than a management number.
2. The spread in total contribution across comparable stores. Same format, same catchment type, different results, is the clearest evidence you have that the variable is inside the building.
3. Brand-language consistency. Ask five associates in five stores to describe the same hero product. Record the answers. This is uncomfortable and it is the fastest diagnostic in the whole list.
4. Share of training minutes spent on brand and story rather than on promotion mechanics. Most fleets discover the ratio is worse than they assumed.
5. Trade-area online performance after a capability change, against matched stores held back. This is the only way to attribute any of it, and it takes a quarter.

Where to start

Pick the product where your associates are least confident, usually the newest or the most technical, and treat it as a brand story rather than a specification sheet. Deliver it to the whole fleet in the same week. Then compare units per transaction on that product against the previous launch.

One accessories brand doing exactly this recorded a 21 percent lift in handbag sales at one store in a quarter, alongside a $20 increase in average dollars per transaction at another.² The write-up sets out both figures and their limits.

Model what a fleet-wide capability change is worth on your own numbers.

References

1. ICSC, The Halo Effect III: Where the Halo Shines, December 12, 2023. 69 retailers and 2,103 individual stores across 50 states and Washington DC, 2019 to 2022; approximately $848 billion in credit card transaction data; thirteen-week window following each opening or closure. icsc.com
2. Multimedia Plus client case study; figures stated publicly by the brand’s retail marketing leadership at NRF’s Big Show, January 2017. Individual store results, reported by the client. Retailer not identified.