Since its launch, Outset Media Index (OMI) has been introduced as a reference point for analyzing media outlets. Pull it up, and the platform immediately shows scores, rankings, filters, and a tidy list of outlets ready for analysis.
At that moment, you’re seeing the finished product. What you don’t see is the work behind it: the digging, the systems built from scratch, the testing, and the manual checks that keep it running.
For a sense of that groundwork, this piece reflects on a recent interview with Maximilian Fondé, senior media analyst at OMI, who’s been part of the team since day one.
The Numbers Worth Having Were the Hardest to Get
To be what it is today, OMI grew out of years of scattered notes and records the team had built while working with media outlets, from traffic and prices to turnaround, regional reach, and who was likely to repost whom. Bringing it together in one system was the easy half. The harder one was handling data those records never really possessed.
As anyone would, the team reportedly started with some well-known and basically unavoidable tools. Similarweb gave the clearest and most thorough read on how much traffic an outlet drew, where it came from, and how readers behaved once they landed on the site. Moz added analytical depth so Domain Authority went in too, but according to Fondé, mostly because the industry still relies on it as a standalone actionable signal out of reflex.
Each tool gave a piece of the picture. However, what the team wanted most was missing. Those tools could size up an outlet but not answer whether a story published there had the potential to find readers, and exactly that is the most important factor to figure out.
That’s why one thing that came up repeatedly throughout the conversation was how often OMI found itself questioning assumptions that are usually taken for granted in media analysis. A good example is visibility itself.
Maximilian argued that publication and discovery are often treated as the same thing when they are not.
“The mere existence of an article on some media outlet doesn’t mean users will organically discover it,” he said.
That realization eventually shaped some of OMI’s internal infrastructure, including a custom republishing parser that tracks how content actually moves after it goes live.
For example, you can put a story on a respected media outlet and watch it reach almost nobody. Here, a traffic figure tells you the platform is busy, not that your piece surfaced inside it, and plenty of well-placed articles sink without a trace. That gap, between a story existing and a story being found, was one Maximilian kept running into.
A single article rarely stops at the website that runs it first. In crypto, it can show up across a dozen smaller sites in a day, and each pickup brings readers the original publisher never counted on. Treat those second-hand runs as background, and you risk missing most of the people who saw the story. OMI reflects those pickups through its Reprints and Aggregators metrics.
Humans Still Go Through Every Line
Why not let the software carry the whole load? It is the natural question to ask about any data product in 2026, but the OMI team refuses to take the easier road.
Maximilian keeps reading the spreadsheets line by line, watching for the domain that breaks a rule everyone signed off on last month. When he hits a problem, he immediately goes to find the answer himself.
Automation gets a turn afterward, and mostly to catch what a person might have walked past.
“Judgment and conclusions in our process are made exclusively by people,” is how he put it, and the whole system is built around that order: people deciding, software assisting.
Scale forces the same kind of call. Stack a major outlet next to a small independent, the big one wins on raw numbers every time, whether or not it deserves to. Keeping that comparison fair means a person interprets in which conditions what counts before the ranking goes off.
Crypto turns every one of these issues up a notch. Plenty of the websites worth tracking keep quiet about their numbers and share too little to judge them on, so they get dismissed even when they deliver.
Where figures do exist, they remain estimates, and an estimate will not tell you whether the audience behind an outlet is real or padded to look the part. The big analytics providers are little help as well, since none of them treat crypto as its own beat. It all turns into a blurry category, and anyone comparing crypto titles works from a frame nobody drew up for them.
Which is why crypto media couldn’t just borrow the analysis built for finance, it needed something drawn around its own shape as well.
What Keeps It All Together
According to Maximilian, the more all-around a final product looks, the less of it was ever finished by a machine, at least in this niche. The data that mattered most had to be built by hand because no tool sold or provided it, and the judgments that stitched the rankings together still get made by a person because no formula makes them well.
What reaches your screen is a clean score and a tidy list. What stays behind it is the slow part, the searching for stories that were published but were never found, and the deciding that keeps a small publication from being flattened by a giant. That buried work is the whole reason the visible half has worth and credibility.
