You have data nobody else has. Every business in this room does. Usage data. Pricing data. Performance data. The stuff that sits in your CRM, your ops dashboard, your billing system — numbers your competitors would kill to see.
Hello, I’m Jeff Payne, and you are listening to The Jeff Payne Show, Episode #21: The Benchmark Nobody Built.
Here’s the uncomfortable part. Most of you are sitting on it. Doing nothing.
And even the ones who do publish it — write it up, put it on the website, call it a “report” — most of them still get zero credit for it. AI never cites the page. Google never ranks it. The number exists. Nobody sees it.
Today, we’re talking about why that happens and what the handful of companies that got it right did differently.
Growth Memo — Kevin Indig and Amanda Johnson, two of the sharper minds in SEO research right now — pulled 301 pages that AI systems actually cited across seven industries. Over a thousand citations total.
They went looking for one thing: how many of those 301 pages were genuine first-party research. Meaning the data and the method were on the page — not a summary of somebody else’s report with a nicer headline.
The answer: 8. Eight pages, out of 301. That’s under 3%.
Now here’s the number that should make you lean forward. Those eight pages pulled 90 of the 1,100 citations in the whole set. Call it 8% of all the citation volume — from 3% of the pages.
Original research is rare. And when it shows up, it punches almost three times above its weight.
Here’s where it gets specific, and where most people will get the lesson wrong.
The instinct is: “Great, I’ll publish original data.” Stop there, you’ve missed it.
Strip those 8 pages apart, and the picture changes fast. Of the 90 citations, 75 came from one content type, clustered around one kind of question: Which of these named options is fastest, cheapest, or best, measured head to head.
Not “We ran a survey.” Not “Here’s our State of the Industry report.” A benchmark. Named competitors, ranked on a specific yardstick, numbers on the page.
Owning proprietary data isn’t the asset. A benchmark that answers a comparison question is the asset. That’s a much narrower target — and it explains why so much “original research” content quietly fails. A survey of 500 marketers with a nice infographic doesn’t answer anything anyone’s actively trying to decide. A benchmark does.
AI systems — and increasingly Google itself — are hunting for one specific thing: information that’s hard to get anywhere else, that directly resolves a comparison someone’s about to act on.
One piece of content accounted for almost half of the work in this entire study. A data warehouse company called Fivetran published a 2022 benchmark comparing four major cloud warehouses — named them all — on speed and cost. Real customer usage data, not a synthetic test. That single page pulled forty-four citations on its own. Add its companion piece and it’s 58 of the 90 — nearly two-thirds of every first-party-research citation in the entire dataset, coming from one company’s benchmark.A page from 2022. Still earning citations four years later.
Why that page and not the hundred other “research reports” published in the same window? Six things, and every one of them is a choice, not luck.
And the URL never moved. Four years, same address.
That’s not a research team’s job. That’s an editorial discipline. Most companies that fail here didn’t fail at collecting the data — they failed at every other step.
If you’re sitting there thinking, “we don’t have data like that,” you almost certainly do. Every service business has usage patterns, timelines, cost breakdowns, and win rates, something a buyer in your category would kill to see compared honestly.
The question isn’t “do we have proprietary data.” It’s “Is there a buyer decision in our category that nobody’s built an honest, named comparison for yet?” That gap is usually wide open, because building it right is genuinely more work than it looks — clean methodology, a stable page, the willingness to name competitors and show your seams.
Most companies that lose here didn’t lose the data. They buried it in a narrative nobody could lift, gated it behind a form, or moved the URL in the next site redesign — and the citation just evaporated.
Find the comparison question in your category that nobody’s answered with real numbers yet. That’s the whole brief.
So here’s the question worth sitting with after this episode ends.
Somewhere in your business right now, there’s a number only you can produce. The question is whether it’s sitting in a spreadsheet nobody sees — or built into the one page that answers the question your next customer is already asking an AI system to settle for them.
Why almost nobody gets credit for the original data they publish — and the one format AI rewards almost to the exclusion of everything else.
You have data nobody else has.
Every business does. Usage data. Pricing data. Performance data — the kind of numbers sitting in a CRM or a billing system that a competitor would genuinely love to see. Most companies sit on it and do nothing. And the ones who do publish it — write it up, put it on the site, call it a “report” — mostly still get zero credit. AI never cites the page. Google never ranks it. The number exists. Nobody sees it.
A recent analysis from Growth Memo’s Kevin Indig and Amanda Johnson explains why, and the answer is narrower and more useful than “publish original data.”
Most companies publish their best data and still get zero credit for it.
8 Pages Out of 301
Growth Memo pulled 301 pages that AI systems actually cited across seven industries — over a thousand citations total — and checked how many were genuine first-party research: data and methodology actually on the page, not a writeup of someone else’s numbers.
The answer: 8. Under 3% of the set.
Those eight pages pulled 90 of the 1,100 citations in the entire dataset — roughly 8% of citation volume from three percent of the pages. Original research is rare. When it shows up, it punches nearly three times above its weight.
Original research is rare — and when it shows up, it doesn’t just perform better. It performs almost three times better.
The Lesson most people will get wrong
The instinct is to stop there: publish original data, done. That’s not quite it.
Of those 90 citations, 75 came from one content type, clustered around one kind of question: which of these named options is fastest, cheapest, or best, measured head to head. Not a survey. Not a “State of the Industry” report. A benchmark — named competitors, ranked on a specific yardstick, numbers on the page.
Owning proprietary data isn’t the asset. A benchmark that answers a comparison question is the asset. AI systems — and increasingly Google — are hunting for one thing: information that’s hard to get anywhere else, that directly resolves a comparison someone’s about to act on.
Owning the data was never the hard part. Building it into something worth citing is.
The Page That Did Half The Work
One piece of content accounted for almost half of this entire study’s primary research citations. A data warehouse company, Fivetran, published a 2022 benchmark comparing four major cloud warehouses — named them all — on speed and cost, using real customer usage data rather than a synthetic test. That single page pulled forty-four citations on its own. Its companion piece brings the total to 58 of the 90 — nearly two-thirds of every first-party research citation in the set, from one company’s benchmark.
A page from 2022, still earning citations four years later.
What made it work wasn’t luck. It named the players instead of hedging. It showed its method — what was tested, how, with what data — instead of just the headline number. It was structured around the exact questions someone would ask, so a passage could be lifted directly. It linked to raw data instead of gating it. It posted dated corrections instead of quietly editing mistakes away — which, counterintuitively, made the rest of the numbers more trusted, not less. And the URL never moved.
That’s not a research team’s job. That’s editorial discipline.
The Move
If the instinct is “we don’t have data like that” — most businesses do. Usage patterns, timelines, cost breakdowns, win rates: something a buyer in your category would genuinely want to see compared honestly.
The real question isn’t whether you have proprietary data. It’s whether there’s a buyer decision in your category for which nobody has built an honest, named comparison yet. That gap is usually wide open, because doing it right is more work than it looks: clean methodology, a stable page, the willingness to name competitors, and showing your own limitations.
Most companies that lose here didn’t lose the data. They buried it in a narrative nobody could lift, gated it behind a form, or moved the URL in the next redesign — and the citation just evaporated.
The gap isn’t a lack of data. It’s a lack of anyone willing to name the competition and show their work.
COMPLETE THE FORM TO
BOOK A STRATEGY CALL
"*" indicates required fields
CHAPTERS:
COMPLETE THE FORM TO
BOOK A STRATEGY CALL
"*" indicates required fields
Subscribe and Share – WE APPRECIATE YOUR SUPPORT
