Matthew
Elsley.

Matthew Elsley runs a company whose product you have already heard, in a store you have already visited, without noticing. QSIC is the software behind the audio at 7-Eleven, McDonald's, Coles Express, Speedway and Stripes. He is 6,000 miles from where he built it, and 13 years past the moment anyone would have called it a good idea.

01What He Is Doing Right Now

In January 2025 QSIC closed a $25 million Series B led by Hedosophia. The pitch, more or less, was that the ambient noise inside a convenience store is inventory. Elsley has been making some version of this pitch since 2012, when the category of "retail media" did not yet exist as a line item, when in-store audio meant a Muzak subscription, and when the software behind the ceiling speakers was mostly considered a facilities cost. Hedosophia wrote the cheque; Elsley moved the money into product.

The proximate use of the money is a rollout called Gulp Radio, which will play across 12,000 7-Eleven, Speedway and Stripes locations by the end of 2025. Gulp Radio is not, strictly speaking, radio. It is a store-specific mix of licensed music and dynamically generated audio ads voiced by a generative AI model that QSIC calls Lucy. Lucy knows the weather, the local price of a Slurpee, and what is on the shelf. She reads the ad copy accordingly. It is the sort of thing that would have sounded like science fiction in a 2013 pitch deck, and it is now the pitch deck.

02The Long Bootstrap

Elsley met Nick Larkins, his co-founder, in school in Melbourne. They started QSIC in 2012 as, essentially, a commercial music streaming service. The unglamorous version of that business is licensing music for retail environments in a way that does not require the store manager to break copyright law with a Spotify Family plan. QSIC did that, but it did it with a lot more instrumentation than the market thought was necessary: an eye on which songs correlated with which sales patterns, which times of day changed behavior, which stores had different acoustic profiles. The instrumentation is now called retail media analytics. In 2013 nobody was buying instrumentation.

The bootstrap years lasted a while. Elsley has said, in more than one interview, that the company was funded early on by borrowing from friends and family and maxing out credit cards, before angel money arrived from some of QSIC's own early customers. This is a familiar founder anecdote with one specific detail: he had a young family at the time, and had already shut down an earlier business to focus on this one. He has founded three companies by the age of 28. Two of them are not QSIC. He does not talk about the first two much.

The Melbourne office in those years had a nickname among the early team - the haunted office - which was not really a comment on ghosts and more a comment on rent. QSIC was one of those Australian software companies that raised the smallest possible amount of money for the longest possible time. Then, in 2021, in the middle of COVID, Carthona Capital led a $4 million Series A. Elsley has been consistent about crediting the pandemic with giving the team something scarce for a startup, which is time to think.

03The Category Catches Up

Between 2021 and 2024 something happened to the language of retail. The industry decided that every physical store was also a media network - shelves as ad inventory, register screens as ad inventory, receipt tape as ad inventory - and a category called retail media went from footnote to line item at $150 billion. QSIC had been building for this without knowing what to call it. The company reported over 220% year-over-year growth through Q2 2024 and rolled out into a widening set of North American accounts. By late 2024 the software was reaching more than 100 million shoppers each month. In-store audio, the industry's most ignored channel, had turned into its most measurable one.

Elsley moved to Dallas. This choice is worth pausing on. Australian founders who expand to the U.S. usually do so via San Francisco (if their customer is another software company) or New York (if their customer is an advertising buyer). QSIC's customer is neither. Its customer is a convenience-store operations lead in Irving, Texas. Dallas is where the buyers are. Dallas is also where the flights out to a Speedway store in Ohio and a Stripes location in San Antonio are cheap. Proximity as a moat is not a very fashionable strategy, but it is the one he chose.

04Lucy, Talking

Lucy is QSIC's generative model for on-demand audio ad creation. The reason Lucy exists is that the old way of running an in-store audio ad required somebody in a studio, a booking process, weeks of lead time, and a very narrow ability to say anything specific to a particular location. Lucy does the opposite of all of that. She writes and voices the ad, in real time, using the retailer's own first-party data - price, inventory, local promotion, sometimes even the weather. If a store in Phoenix has too many Gatorades in the cooler and it is 104 degrees outside, Lucy will notice.

The 14% number QSIC uses is average sales lift for retailers running their tech. It is the sort of number that any B2B software company will happily quote and any analyst will happily discount, and it is the number Hedosophia effectively bet on when it led the Series B.

05Voice, In Person

On the state of the pitch meeting

Elsley's speaking style is unshowy. On stage at the P2PI Retail Media Summit and at Shoptalk he tends to underplay QSIC's numbers, which - given that founders as a class do the opposite - is what people tend to remember. In podcast interviews about the QSIC founding story he tends to spend more time on process than on the company itself: how he thinks about hiring a CTO first, how he thinks about picking one investor over another when both cheques are the same size, how he thinks about listening to the room before pitching it.

06Timeline

07Small Things Worth Knowing

Elsley did his business degree at Swinburne in International Business and his VCE at Xavier College. Before QSIC he had two other companies, one of which was Easy Control, a smart-home integration business, which - if you squint - is not entirely unrelated to a startup whose current job is patching AI voice models into a 7-Eleven ceiling. His generative model is named Lucy. His 7-Eleven audio product is called Gulp Radio. He is on the record as saying the best thing about the pandemic, from QSIC's perspective, was that it forced a quieter period of reassessment. The company he built has raised roughly $30.4 million in total, and the biggest cheque of that was written thirteen years after the idea.

The category he ended up in is one he did not name and did not invent. It arrived at his address around 2023 and knocked. He answered.

08Where To Find Him