Stop Pitching Reporters. Start Being the Thing They Cite.

There is a category of owned media asset that outperforms everything else B2B companies publish, and most marketing teams do not build it because it does not look like marketing while it is being built.

It is a recurring number that nobody else has.

Ramp runs an Economics Lab staffed with an actual economist, Ara Kharazian, who previously led economic research at Square. It publishes a Business Spend Index constructed from anonymized transaction data across more than 50,000 US businesses, and an AI Index drawing on more than 70,000. Kharazian's analysis has been picked up by the New York Times, Bloomberg, NPR's Planet Money, and Axios. We have covered the output ourselves, which is the entire point of the exercise.

Carta publishes State of Private Markets quarterly, written by Kevin Dowd, who covered venture capital and private equity at Forbes and PitchBook before joining. The reports are the default reference for startup fundraising terms, and they get cited in venture coverage constantly.

Neither company is doing content marketing in any sense their peers would recognize. They are running research operations that produce primary source material, and the citations follow automatically because journalists and analysts need the numbers.

Why this beats everything else in the owned media stack

Three reasons, and the third one is the reason this piece exists now rather than three years ago.

Nobody can replicate it. Your competitor can hire better writers, out-publish you, and cover the same subjects with more resources. They cannot produce your transaction data, your pricing data, your hiring data, or your platform telemetry. Proprietary data is the only durable moat in owned media, because it is the only input that is genuinely yours.

It inverts the pitch. Media relations is a function built around asking. A recurring index reverses the direction of the relationship entirely. Reporters on the beat build it into their coverage cycle, and after four or five editions the release becomes something they anticipate rather than something you have to sell.

It is the format AI systems treat as citable. As category messaging gets flattened into interchangeable summaries, the material that survives is material that answers a question no other document can answer. An original quarterly figure with a documented methodology is close to the ideal shape for that. Everything else in your content library is competing to be one of many sources on a subject with plenty of sources.

What qualifies

Not all internal data becomes an index. The bar has four parts, and most company data fails at least one.

It has to measure something outside your company. The Ramp index page states directly that it is not a measure of Ramp's own financial performance, and that distinction is what makes it credible rather than promotional. A metric that goes up when you do well is a growth chart. A metric that moves with the economy regardless of how you are doing is an indicator.

It has to be large enough to generalize. Tens of thousands of accounts, or a defensible argument about why your sample represents the population. A survey of 200 customers is not an index, and treating it as one will get you dismissed by the exact people you are trying to reach.

It has to recur on a fixed schedule. The value compounds through repetition. A one-off report is a campaign. A quarterly report that has run for three years is infrastructure, and the third year is when reporters start citing it without being told it exists.

The methodology has to be published in full. This is the requirement companies most often skip and the one that determines whether serious people will touch the output. Ramp publishes the mechanics: same-store comparison across businesses active in consecutive months, Törnqvist weighting, machine learning models applied to strip out platform-lifecycle effects like onboarding ramp-up and churn, size caps so no segment dominates. That level of disclosure is what separates an index from a chart in a press release.

What it costs

More than a blog, less than a brand campaign, and the cost sits somewhere unfamiliar on the org chart.

The staffing model in both examples above is the notable part. Ramp hired an economist. Carta hired a journalist. Neither hired a content marketer, and that choice is doing most of the work, because the output has to survive contact with people who evaluate research for a living. A marketer producing this material will unconsciously optimize for the conclusion that flatters the company, and readers who handle data all day will detect it in the first chart.

Budget for one senior researcher or data journalist, engineering time to build a reliable extraction and anonymization pipeline, and legal review that nobody plans for and everybody needs. Customer transaction data carries contractual constraints and privacy obligations that vary by agreement and jurisdiction. Route the aggregation and disclosure approach through counsel before the first edition rather than after, because the version of this that gets built quickly and cleaned up later is the version that gets killed.

Then expect twelve to eighteen months of publishing before meaningful citation. The first two editions will be ignored. The third gets picked up by a trade outlet. Somewhere around the fifth or sixth, if the methodology has held, a national business desk uses it, and after that the compounding is fast.

Why most companies will not do this

Everything above is available to any B2B company whose product generates data, which is very nearly all of them. Almost none will build it, and budget is not the obstacle.

An index only works if it is allowed to report bad news. Spending falling. Hiring slowing. Deal terms tightening. Adoption of the exact thing you sell flattening out. The first time a company quietly adjusts a number because the quarter looked wrong, the people who use that number for a living will notice, and an asset that took two years to build becomes worthless in roughly one edition.

That is a governance problem rather than a marketing one, and it is the same problem as editorial independence in a brand newsroom. Someone has to decide in advance that the output does not get softened when it is inconvenient, and then write down who is not permitted to soften it. Companies that cannot answer that question should not start, because a compromised index is worse than no index. It teaches your most sophisticated readers that your numbers move for reasons other than the world moving.

We should be direct about our own position here. This publication is owned by Outlever, which builds newsrooms for B2B companies, so we have a commercial interest in more companies concluding that publishing is worth the trouble. We do not currently run an index of our own, and readers are entitled to weigh that against the argument above.

What we would offer in return is that the standard in this piece is checkable. External measurement, a sample large enough to generalize, a fixed schedule, and a published methodology. Hold anyone making this argument to all four, including us, and especially the fourth. Companies that will not show their workings are usually protecting something other than a trade secret, in a market where almost every other signal has already been flattened into the same summarized paragraph.

Your competitors can copy your positioning by Friday. They cannot copy your data, and they certainly cannot copy four years of publishing it on schedule.