#34 - Ara Khazarian
Ara Kharazian is the lead economist at Ramp, where he uses the company’s spend data to produce original economic research. Because Ramp sees how thousands of companies are spending in real time, Ara is able to spot and explain patterns in where business is moving. He’s been especially focused on AI spending inside companies: how much businesses are actually spending on LLM tools, with which companies, and how those patterns are changing over time. Ara tells us all about how Ramp built its insights content motion, why first-party data can be such a powerful way to earn attention, and how he thinks about finding stories inside a company’s data set. We also get into how he built credibility with reporters and researchers, why Ramp makes parts of its data interactive and public, and what other companies should think about before trying to turn their own data into content.
.jpg)
Hey all,
This week I sat down with Ara Kharazian [insert link], Lead Economist at Ramp [insert link], a global corporate spend management platform used by over 50,000 companies.
Ara has been leading Ramp’s effort to turn its proprietary spend data into a content engine. His content byline now accounts for 30% of content readership on Ramp’s website.
This week’s newsletter is a case study in what that looks like.
Ramp has a uniquely rich view into how growth-stage companies spend. Because it sits at the center of their transactions, it can see which vendors companies use, how much they’re spending, and how those patterns shift over time.
This gives them a proprietary dataset with real economic signals…a content goldmine.
Ramp hired Ara in January 2025 to mine that gold. With the title of “Economist” , his job was to figure out what the data was actually suited to reveal and how to turn those insights into research, stories, and an audience.
The plan worked: over the past year, Ara’s readership has grown to more than 1 million users on LinkedIn alone, with millions more across X and Substack.
How Ramp Turned Proprietary Data into a Content Engine
Step 1: Identify what your data actually tells you
One of Ara’s clearest points is that good data-driven content starts with identifying what your data is, leaning into that bias, and building credibility in that space.
Ramp’s data does not represent the average American business, but rather skews toward tech-forward, high-growth companies. These businesses grow roughly 3x faster than the average US business.
Ara didn’t shy away from this bias, he leaned into it.
“If you want to know how fast growing businesses spend… you might want to look towards Ramp data."
The first chart he shipped showed growth in AI spend. He leaned into consistently demonstrating how a large, fast-growing spend category is tied to a major shift in the market.
Step 2: Construct Layered Distribution
Ara built Ramp’s distribution system in layers. At the foundation is the research data itself—products like the AI Index and broader spend index. On top of that sits distribution: a substack where Ara publishes his findings in long-form analysis, and social posts that surface bite-sized insights. His team also created interactive dashboards on Ramp’s website where people can explore the data more fully.
“Interactivity goes a long way…a lot of the research [that] companies put out needs to say ‘hello’.”
That structure made Ramp’s research easier to encounter, understand, and return to. It also meant that when attention converged on a topic, Ramp had the system to meet the demand.
Step 3: Chase Regularity, Not Virality
Ara’s breakout moment was the “996” story. Someone asked whether Ramp’s data could show the rise of extreme work culture in San Francisco tech. He found a correlation between rising late-night and weekend food delivery expenses among SF-based tech workers, a pattern he said did not appear in other US cities.
The post generated 3.6k engagements on LinkedIn and became his biggest social hit. It was covered by The New York Times, and pulled more people into Ramp’s orbit.
“It was like hitting 20 in black jack.”
But Ara’s point was that moments like this are unpredictable. You cannot build the strategy around going viral. You build the system, post consistently, and let breakout moments emerge from the body of work. Content is a long game.
Since Ara started, he’s racked up 3M+ impressions on LinkedIn alone, with millions across X and his newsletter. His content now accounts for 30% of all readership on Ramp’s website. That’s not due to one single viral moment, but building trust through consistent posting and differentiated commentary that only Ramp could provide.
The takeaway
Proprietary data can be powerful source material, if you know what it’s actually suited to reveal.
Ramp’s approach offers a useful model: define the value of the dataset, build layered distribution around the insights, and commit to publishing regularly over time.
Done well, that kind of system can grow into a recognized media motion that creates inbound opportunities by giving away genuinely valuable insight.
.jpg)
.jpg)
.jpg)
%20(1).gif)
.gif)
.gif)
%20(1).gif)
