You open a dashboard expecting a list of names, a spreadsheet of emails, a map of device IDs. Instead, you are presented with a population that does not exist. It has demographics, purchasing habits, and daily commutes. It votes, browses, and buys. It is a statistically perfect mirror, built from over 10,000 attributes, but it contains no actual people. This is the first screen of Arima’s Synthetic Society, and it asks you to trust a reflection more than the real thing [Arima company page].
The Statistician's Mirror
Arima is the seven-year-old project of Winston Li, a Toronto-based statistician who spent years inside the marketing measurement engines of PwC Canada and Omnicom Media Group [Advertising Research Foundation, October 2026]. His frustration, common in the industry, was a growing chasm: the need for granular consumer insight was exploding just as privacy regulations and platform changes were walling off the raw, identifiable data that fueled it. Li’s answer was not to find a better way to collect personal data, but to stop collecting it altogether. The Synthetic Society is a cloud-based platform that simulates populations, allowing marketers to run media mix models, test campaign scenarios, and derive insights from synthetic individuals who behave like real ones, but who leave no privacy trail [Arima company page]. The company, which now employs 25 people, has raised a seed round from a consortium of Canadian investors including Graphite Ventures and the MaRS Investment Accelerator Fund to scale this vision [Amii AI Startup Pitchbook, April 2025] [CB Insights].
Why a Simulacrum Sells
The bet hinges on a shift in what constitutes actionable data. For decades, marketing intelligence meant aggregating and analyzing traces left by individuals. The new model, which Li evangelizes on industry podcasts, argues that you don’t need to know who bought the sneakers; you need to know the precise constellation of attributes,age, location, income, media consumption, mobility patterns,that predicts a sneaker purchase [The Analytics Power Hour, June 2025]. By fusing census data, market research, purchase records, and mobility signals, Arima builds a high-fidelity model of a market. A brand can ask it how a price increase might play in Montreal, or which digital channel is most efficient for reaching young parents, and get an answer derived from simulated behavior. The product’s appeal is its clean, compliance-friendly architecture: it is, by design, a system of inferences, not a database of persons.
The Traction and the Team
Arima’s early traction is measured in headcount and institutional recognition. Growing to 25 employees by April 2025 signals a move beyond a founder-led consultancy into a product company with engineering and sales needs [Amii AI Startup Pitchbook, April 2025]. Li’s background is the company’s cornerstone. His credibility as a former director of data science at major firms provides a bridge to the very enterprise marketing departments that feel the privacy pain most acutely. This is not a product built by AI researchers in a vacuum; it is built by someone who has sat in the client’s chair, wrestling with the limitations of traditional data. The company’s presence on panels like the Advertising Research Foundation’s Marketing Effectiveness Accelerator further cements its position as a serious tool for a serious problem, not a speculative tech demo [Advertising Research Foundation, October 2026].
Where the Reflection Could Distort
The core risk for Arima is not technological feasibility, but market conviction. Synthetic data is a rapidly evolving field, but convincing a marketing VP to stake a million-dollar media budget on the behavior of simulated people requires a profound leap of faith [The Analytics Power Hour, June 2025]. The company must prove its mirror is not just statistically sound, but commercially predictive. Its success depends on answering three critical questions:
- Model fidelity. Does a synthetic population react to a new ad campaign or a competitor’s promotion the same way a real one would? The proof will be in consistently accurate forecasts, not just impressive attribute counts.
- Sales motion. Marketing mix modeling is a sophisticated, high-ticket service often sold through consultancies. Arima must build a repeatable SaaS sales process for a product that displaces a service.
- Competitive landscape. While no direct competitors are named in available sources, the space for privacy-safe analytics is crowded. Arima’s wedge is the depth of its synthetic population,10,000 attributes,but it must defend that technical moat as larger platforms build similar, broader capabilities.
The company’s seed funding from Canadian institutional investors provides a runway to tackle these challenges, but the round size and valuation remain undisclosed, leaving the scale of its resources a private matter [CB Insights].
The Next Twelve Months
For Arima, the coming year is about translation: turning a compelling statistical concept into undeniable commercial evidence. The key metrics to watch will be named enterprise customer logos and published case studies that move the conversation from “this is how it works” to “this is how it drove ROI.” The team will likely expand further, particularly in sales and customer success, to manage the complex onboarding this product requires. Another round of funding may emerge as the company seeks to accelerate growth and potentially expand its synthetic models into new verticals beyond marketing.
Ultimately, Arima is answering a cultural question that extends far beyond media budgets. In a world increasingly wary of surveillance, what does it mean to understand a population? The Synthetic Society proposes that you can know a people intimately without knowing a single person at all. It is a bet on the power of the pattern over the individual, on the map being more valuable than the territory. The success of that bet will determine whether marketing’s future is built on watching us, or on simulating us.
Sources
- [Arima company page] About Us - Arima's data | https://arimadata.com/company/about/
- [Amii, April 2025] Amii AI Startup Pitchbook | https://static1.squarespace.com/static/67a3aea41c5d89522a0b0522/t/685c1b5a117f48467a21ef14/1750866779082/Amii_AIStartup_Pitchbook_04-2025.pdf
- [CB Insights] Arima - Crunchbase Company Profile & Funding | https://www.cbinsights.com/company/arima-2
- [Advertising Research Foundation, October 2026] MARKETING EFFECTIVENESS ACCELERATOR (MEA) | https://thearf.org/arf-events/marketing-effectiveness-accelerator-mea-advisory-committee-2026/
- [The Analytics Power Hour, June 2025] #274: Real Talk About Synthetic Data with Winston Li | https://analyticshour.io/2025/06/24/274-real-talk-about-synthetic-data-with-winston-li/
- [Arima company page] Synthetic Society - Arima's data | https://arimadata.com/products/synthetic-society/