Eight years ago, Eytan Buchman, then chief marketing officer at the global freight marketplace Freightos, began emailing a weekly update on ocean freight container prices to reporters. There was no paywall, pitch, or ask — just data.
The email circulated to a small audience, earning Freightos a reputation as a reliable source on freight pricing among a niche of trade reporters.
Then the pandemic hit.
When COVID emptied the shelves and supply chains became front-page news, Buchman’s data was cited at the White House press podium, referenced by The Wall Street Journal and The New York Times, and, eventually, traded as a financial instrument on the Singapore Exchange and Chicago Mercantile Exchange. The Freightos Baltic Index became a global benchmark — and a seven-figure standalone business for the company.
Buchman, who has since left Freightos, calls credible data backed by sharp analysis “a powerful Trojan horse to get into somebody’s head.” Proprietary data opens doors that traditional marketing can’t, he believes.
He isn’t alone. A growing number of brands are pairing proprietary data with news hooks and analysis — some producing little more than a recurring data point; others, work that is virtually indistinguishable from data journalism.
And it’s getting picked up by both mainstream media and LLMs. A widely cited 2024 paper out of Princeton and the Indian Institute of Technology found that adding citations, statistics, and quotations is the most effective way to boost visibility in LLM responses.
The press, meanwhile, is treating these companies the way it once treated think tanks. Ramp’s Economics Lab, a Substack run by an in-house economist, tracks economic signals through the company’s spend data and has been cited by the Federal Reserve and Financial Times. Glassdoor and Indeed run dedicated economic research arms that national media outlets cover. Apartment List publishes a monthly rent report that papers across the country cover.
Cassandra Naji, CEO of the content marketing agency Campfire Labs, has been tracking the rise of data-driven reporting across B2B SaaS for the past year. “The days of easy traffic from search are over. This kind of content can build a moat,” she said. The companies doing it well, however, are a small group of outliers, not the leading edge of a wider trend, Naji believes.
The strategy is clearly paying dividends for Ramp and its peers. It begs the question: why aren’t more brand publishers following their lead?
The political problem
The usual explanation is technical. Most companies, the conventional argument goes, don’t have the data pipelines, analyst headcount, or infrastructure to publish data-driven content.
But in Naji’s experience, the reason is actually simpler: “Content and marketing teams haven’t been granted permission or been able to advocate strongly enough to get it. That’s the biggest challenge,” she explained.
In other words: the data exists, but content can’t touch it. At many companies, there’s still a stubborn notion that marketing-adjacent teams shouldn’t have the same access as engineering or analytics functions. Naji ran a small, informal survey of B2B marketers on LinkedIn and found that 40% cited data access as their primary blocker to creating data-led content.
The exemplar programs, she noted, aren’t headed by marketers. Ramp’s Ara Kharazian is an economist who previously ran economic research at Square. Carta’s Peter Walker, the company’s head of insights, is a research lead. Hiring a credentialed analyst to front the work may function as an internal credibility transfer, making it easier to unlock data and budgets.
Another barrier is editorial. Even when the data is in hand, knowing what and how to publish is its own discipline. The most pressing question today, in Naji’s framing, is not whether a piece of content will rank in search, but if a reporter will quote it. Most marketing teams haven’t been trained to think that way.
For former journalists, on the other hand, that instinct is intrinsic. They’ve spent their careers deciding what’s worth a reader’s attention. Some companies are now hiring them to work alongside economists and other experts to translate analysis into copy that the press will pay attention to.
Speed and distribution
Government data sources are slowing down. Newsrooms are contracting. The institutions that have historically supplied the figures underpinning data journalism are being defunded just as demand for fresh, citable numbers is rising due to AI.
And government statistics agencies tend to publish with multi-month lags. A well-resourced corporate data team, meanwhile, can push out a new chart in a week. The companies getting picked up regularly have moved to quarterly, monthly, and even weekly cadences. The dataset becomes a rolling franchise, rather than a single release.
Distribution, too, has changed shape. Ramp’s data feeds the Bloomberg Terminal and is queryable in Claude, ChatGPT, and Perplexity via the Model Context Protocol, which lets LLMs pull real-time data directly from a provider’s servers. The same dataset can now operate as a PR asset, a financial markets product, and a citation source for AI, all from a single publishing pipeline.
“They’re increasing that surface area of touchpoints in a way that a year ago I wouldn’t have thought about as a distribution plan,” Naji said. For a small number of companies, getting data in front of an audience has stopped being a marketing campaign and started becoming a technical buildout.
Corporate data, of course, carries inherent biases. Buchman argues that at least they’re transparent: “If Ramp is reporting an increase in credit card spend, I know what they want to show, and I can control for that. I think that’s refreshing compared to trying to distill the angle from traditional media.”
An index based on card spend might be accurate, but that doesn’t make it a complete representation of the market.
But the bias can still be hard to pin down. When Ramp’s AI Index reported in May 2026 that Anthropic had overtaken OpenAI in US business adoption, the finding was picked up by mainstream media within hours. OpenAI pushed back, explaining its biggest customers don’t pay by credit card — a reference to the large, directly invoiced enterprise deals that a card-based index such as Ramp’s tends to miss. The headline number was accurate according to Ramp’s data, but that doesn’t make it an independent measure of the US AI market.
None of this means corporate data can’t be trusted; only that it should be read closely, with an eye to what the methodology captures and what it leaves out.
Starting small
Proprietary data is a powerful asset, and you don’t need the scale of a company like Ramp to own a corner of it.
Freightos’s data storytelling operation began with a weekly email. Naji’s minimum viable version is similar: a single recurring chart, built on whatever proprietary data a company is already generating, and published on a monthly or weekly cadence. A free tool such as Datawrapper is enough to get the job done. “Sit down and think: what’s the one thing we can package up and keep updated on a rolling basis that no one else has?” said Naji.
Executive buy-in for time-intensive data reporting projects tends to be the harder challenge. Buchman’s approach is incremental: take a small risk, prove a return, then take a larger one. “It’s hard to swing for the fence on day one, but it’s much easier to run to first base,” he said.
Before the Freightos Baltic Index, Buchman’s team ran a string of smaller data experiments to see what would resonate. They created a fake company to solicit price quotes from the top 20 global freight firms, then turned the results into competitor benchmarks they could pitch back to each one. Another time, they scraped Amazon’s career database to map which qualifications the company wanted when hiring for supply chain roles. “Each time, we saw how much credibility you earn when you use data to tell the right story,” Buchman said.
Buchman has three rules for keeping data credible: publish the methodology. Run findings even if they don’t flatter the company. And find an outside partner to validate the data, as Freightos did with the Baltic Exchange. Above all, he said, the dataset should be treated like a product, not a campaign.
The chart is just a starting point. Buchman’s weekly email eventually became a newsletter, a self-service platform, a webinar series, and finally, a seven-figure data product.
A widening lead
Most brand content, Naji said, is moving in the opposite direction of the companies mentioned in this story. “They’re producing more content just because they can [with AI], rather than doing less and putting out really high-quality, data-driven stuff,” she said.
When asked where the field would be in 18 months, Naji hesitated. The brands with serious resources — she names Ramp, Carta, Ahrefs, SparkToro, ActiveCampaign, and Clay — will likely pull further ahead, she said. The citation patterns will reinforce themselves, and a small group of brand publishers will likely end up owning a disproportionate share of airtime and authority: “We might see brand oligarchs emerging,” Naji cautioned.
“Institutions that have historically done this kind of content are contracting: government agencies and newsrooms. There’s a small number of companies with proprietary data and a lot of money to invest in editorial rigor. They’re rushing to fill the gap, and they’re doing it really well.” —Jennifer Guay
