The Retail World Runs on Guesses. MarketDial Sells Proof.

Every day, Amazon changes a button, a price, a recommendation - and by lunchtime it knows, against millions of shoppers, whether the change helped or hurt. A supermarket chain does something far riskier. It rearranges an aisle, adds a self-checkout, or drops a price across 400 stores, and then waits a quarter to see if sales moved, never quite sure whether the credit belongs to the idea or to the weather. That gap - between how the internet learns and how physical retail guesses - is the whole reason MarketDial exists.

Founded in Salt Lake City in 2016, MarketDial makes what the industry calls test-and-learn software. The plain-English version: it lets a retailer try a change in a small, carefully chosen set of stores, compare those stores against a matched control group that did nothing, and measure the actual lift before rolling the change out to the entire chain. It is A/B testing, the technique that quietly built the modern web, pointed at the physical economy that still makes up the majority of retail spending.

01 / The originTwo people who made a living on opinions

There is a small irony baked into MarketDial's founding. CEO Morgan Davis came out of venture capital, having invested at Peterson Ventures before starting the company. Co-founder Johnny Stoddard came from McKinsey & Company. Venture capital and management consulting are two of the most opinion-driven jobs in the economy - and the pair built a company whose entire purpose is to replace opinions with tested results.

The frustration was concrete. Working with retailers, Stoddard kept seeing the same pattern: chains either rolled out expensive initiatives without testing them at all, or they tested with flawed methodology - no control group, no correction for store differences - and drew confident conclusions from numbers that could not support them. The tools that did it properly were expensive, technical, and, as MarketDial would later argue in court, effectively controlled by a single incumbent. So Davis, Stoddard, and co-founder Joseph Turner set out to make rigorous in-store testing something a merchandising team could actually use.

02 / The productSoftware that ships with a data scientist

MarketDial's platform is built around four pieces, and the second and fourth are the ones customers tend to remember. There is onboarding and data processing, where MarketDial's own team untangles a client's messy point-of-sale and location data. There is test design, which automates the statistical setup - representativeness, comparability, confidence - that a company would otherwise attempt in a fragile spreadsheet. There is output and analysis, which reports the lift, the basket impact, and the optimal rollout. And there is client support, which is where the company does something unusual for enterprise software: it assigns a dedicated data scientist to the account, with no hourly billing.

That last part matters because the hard part of retail testing is not the software - it is trusting the answer. A merchandiser who has run stores for twenty years is not going to bet a rollout on a dashboard they do not understand. Pairing the tool with a human who can defend the math is how MarketDial gets skeptics to act on results.

Figures reported by MarketDial. The company markets its return guarantee and satisfaction score as proof points; treat them as company-stated rather than independently audited.

03 / The problemWhy "did sales go up?" is the wrong question

The trap in retail analytics is that almost everything correlates with almost everything. Launch a promotion in November and sales rise - but sales were going to rise in November anyway. Remodel your best stores and they perform well - but they were your best stores to begin with. The only honest question is not "did sales go up?" but "did sales go up compared to stores that were just like these but didn't get the change?" That comparison - the control group - is the thing MarketDial productized.

Under the hood, the work is heavier than it looks. Matching stores means accounting for size, demographics, seasonality, and format so that the control group genuinely resembles the test group. MarketDial's stack reflects that: alongside a modern web app built in TypeScript and Vue, its engineers run Python analytics and lean on IBM's ILOG CPLEX optimization engine to solve the store-matching and rollout math. The point of all of it is to hand a non-statistician a number they can defend in a boardroom.

04 / The customersGrocery, gas, apparel, and burgers

MarketDial's client list reads like a tour of the American strip mall and the aisles beyond it. Convenience and fuel chains like Maverik, Kum & Go, and Casey's General Stores; big-box and specialty retail like Dick's Sporting Goods, Five Below, The Container Store, and La-Z-Boy; grocery through Meijer; apparel through Calvin Klein; food service through Potbelly; and, internationally, the Australian giant Woolworths and its Big W chain. Consumer-goods makers such as Keurig Dr Pepper use it too, to test how products actually move on real shelves.

05 / The rivalThe incumbent, and the lawsuit

MarketDial did not invent offline test-and-learn. That distinction largely belongs to Applied Predictive Technologies, or APT, whose Test & Learn software has served large enterprises for years and which Mastercard acquired in 2015. MarketDial's pitch has always been that it is the independent, more service-heavy, more approachable alternative to a giant now owned by a payments network.

That framing turned into litigation. MarketDial filed an antitrust complaint alleging that APT and Mastercard hampered its ability to compete by bundling APT's testing software with Mastercard's card services - a claim that, whatever its eventual outcome, tells you the market for deciding-by-experiment is big enough for a Salt Lake City startup to go to federal court over. Most startups say they are disrupting an incumbent. This one filed suit.

06 / The moneyBacked to build, not to burn

MarketDial raised early money from Utah's Kickstart Seed Fund and Peterson Ventures, then closed a $7.5 million Series A in September 2018 led by Crosslink Capital, with Silicon Valley Bank joining. In November 2021 it announced a growth round led by Finistere Ventures, with Crosslink and the seed backers returning. The company has kept round sizes and terms relatively quiet - reported totals range from roughly $10 million to $16 million depending on the source - which fits a business that sells to a limited number of large, high-value enterprise accounts rather than chasing consumer scale.

07 / What's nextFrom shelves to real estate

The newest expansion, SiteHQ, points MarketDial's core idea at a different billion-dollar question: not "what should we change in the store?" but "where should we build the next one?" It is an AI-driven site-selection product that applies the same discipline - model it, don't guess - to real estate decisions. The through-line is consistent. Whether the decision is a price, a planogram, a menu item, or a plot of land, MarketDial's bet is that the physical economy will eventually treat testing the way the internet already does: as the default, not the differentiator.

08 / The takeawayWhat anyone can borrow

The transferable lesson in MarketDial's story is not really about grocery stores. It is a discipline: before your organization makes a big, irreversible, expensive bet, find the smallest valid way to test it first, and insist on comparing the result against what would have happened anyway. MarketDial packaged that habit into software and put a return guarantee on it. It works best for operators with many comparable locations and clean sales data; it works less well where sites are too few or too different to build a fair control group, or where the effect you are chasing is too small to measure against the noise. Within those limits, the pitch is simple, and a little old-fashioned: test before you act.