Your VP of Store Experience came in asking for budget. They walked into a competitor’s store, or they read something on the flight home, or one of their district managers keeps complaining that the audio in two stores makes the staff want to quit. Now there’s a line item on the table that didn’t exist last year.
You’re the one who has to decide whether it’s real.
Which P&L lines does retail audio actually move? #
Your operations team already tracks four numbers that shift when the audio in a store changes. All four sit on a dashboard somebody reviews every week.
Dwell time. A shopper who stays twelve minutes instead of nine touches more product, asks more questions, and gets further into the purchase decision before anyone has to intervene. Ron Milliman demonstrated the tempo effect in supermarkets in 1982. Forty years of replication have done nothing to weaken the finding. Slower music, longer visits.
Basket size. Most retail analytics teams track dwell and basket as a pair, because shoppers who linger almost always put more in the bag. A shopper who lingers doesn’t just look at one more thing. They reconsider what they came in for.
Willingness to pay. Areni and Kim published the original wine-shop study in 1993. A shopper surrounded by classical music selected more expensive bottles at a meaningful rate above the same shopper surrounded by top-forty. North and colleagues replicated the effect in 1999 with French and German music in a grocery wine aisle. Researchers have reproduced the effect across settings for three decades. When a shopper reads the audio as a match, they revise their sense of the product upward. The price on the tag never changes. The shopper’s read on what it’s worth does.
Return visits. The hardest of the four for an analytics team to isolate cleanly, and the one most operators have an instinct about anyway. Shoppers who feel the store was built for them come back more often than shoppers who don’t.
Why this line item hasn't existed before #
Retail audio has been a facilities line for thirty years. The controller pays a catalog provider something between fifteen and eighty dollars per location per month, the retail team complains twice a year about the selection, and nobody in finance ever looks at the spend next to store-level revenue.
Two things have changed since then.
The first is that retail analytics teams finally have the instrumentation to measure store-level spend against store-level revenue in something close to real time. Operators who never thought of audio as a variable have started asking the obvious question and finding no answer in their current vendor’s reporting. Mood Media, Soundtrack Your Brand, and their competitors sell licensed tracks organized by genre and mood. No salesperson at those companies has ever walked into a CFO’s office and promised a number on dwell or basket, because their product was never built to produce one.
The second is that generative AI music reached commercial grade in 2024. Retailers can now commission their own audio instead of licensing somebody else’s. That removes the per-location performance fees, removes the PRO compliance exposure, and removes the vendor’s argument that the music is whatever it is because that’s what the catalog has. The audio in a shop can be composed for the specific shopper the retailer already profiles on every other variable. That alone changes what a finance team is modeling.
The ROI math, done honestly #
Here is the version of this calculation a finance team can actually run.
Pick a single store in your portfolio. Take the trailing twelve months of average transaction value, transactions per month, and dwell time if your sensor data supports it. Apply a conservative lift from the published literature to one of those numbers. Multiply. That’s the incremental revenue per location per month, under one assumption about which number moves.
Then do the same exercise with a zero-lift assumption. The difference between those two numbers is what you’d be testing for in a pilot. It is almost always large enough to matter, and almost always larger than the cost of any legitimate alternative line item your VP of Store Experience could bring you instead.
No finance team should commit real money on a study from 1993 alone. The published research supports a hypothesis. It does not substitute for your data, in your stores, on your shoppers. That’s what a pilot is for. The point of running one is that at the end, you know.
What a clean test actually looks like #
A pilot a CFO can defend has three parts.
Matched stores. Pick treatment and control locations that look alike on the variables you already use to compare stores: traffic, demographics, format, seasonality, merchandising. You do not need a statistician in the room to do this well, though most operations teams already have one they lean on.
A long enough window. Retail is noisy. A week of data is not data. A month is thin. A quarter gets you somewhere defensible. The exact length depends on your store volume and the variance in your P&L, and that conversation belongs with your analytics lead.
A single-variable change. Nothing else should move in the treatment stores during the test. No merchandising reset. No staffing change. No promotional calendar weirdness. That’s the hardest part of pilot design, and it’s the one most vendors skip, which is why most vendors produce case studies a finance team cannot read.
At the end, you have your numbers. Dwell, basket, conversion, transaction value. Your data, from your stores, measured the way you measure everything else. If the lift is there, you have a real case. If it isn’t, you walk away, and you’ve lost nothing except a quarter of background audio you were going to play anyway.
The question under the question #
Your current music has a cost. Nobody on your team has ever measured it.
When a CFO pushes back on a new audio line item, the real question is usually whether this is the kind of spend that will produce something measurable or the kind that quietly disappears into facilities.
That is a fair question. The honest answer is that most existing audio spend is the second kind. It renews every year, nobody on the retail team loves it, and no one in finance can point to a number the spend moved. The fact that it costs fifty dollars a month per store is the reason nobody asks harder questions.
Retailers who have started treating audio as a measurable variable look at the spend the way they look at lighting or store layout. Those line items carry weight in the capital plan because somebody, at some point, ran the comparison. Nobody on your side has run that comparison on audio yet.
For the financial framing that lets a CFO defend this as a line item, see The CFO’s Case for Retail Audio.
