For at least the past decade, B2B marketers grew accustomed to creating content that I’ve politely called “SEO book reports.”
Stitch together the commonalities among the top ten-twenty existing articles against a target keyword, slap on H2s, H3s, bullet points, maybe add a little paragraph or cute anecdote in the introduction, and build some links. Hope and pray that the additional paragraph is enough to usurp the incumbents. That arbitrage is mostly dead. It has been decaying in effectiveness for a while, even absent any changes with AI-driven answers in AI Overviews, AI Mode, or ChatGPT.
Of course, AI-powered answers render this type of content even less effective, as the friction for consensus answers falls to nearly zero, meaning there’s no real reason to click onto the 3rd or 4th rendition of “what is marketing automation” written by a non-subject-matter-expert. ChatGPT will just tell you…right away.
Effectively, it’s easier than ever to create this type of content, too, as a $20 ChatGPT subscription and cursory knowledge of prompting will get you a median-quality draft. And when median-quality content is nearly free to create, it becomes abundant, eroding the value and visibility of anything below median-quality.
One antidote is information gain, and I mean that in a broad and colloquial sense (as in, unique value created and distributed), though Google’s patented version of the term is also useful (i.e. extra value a document adds beyond what a user has already consumed). LLMs appear, at least according to some early research, to prioritize trustworthy and unique data, especially in relation to top-of-the-funnel queries.
As an aside (for another newsletter), this type of data is also readily distributable via authoritative publications, which are also cited with more frequency than small company blogs. So lots of bang for your buck in creating new data points.
Winning visibility, therefore, is less about how elegantly you remix consensus and more about how bravely you create something the web (and more particularly, your target audience) hasn’t seen. Enter proprietary data.
(Side note: every time I link to a company’s original research in this essay, I get a mini-dopamine kick as I get to reinforce the idea that this stuff is effective and people like to cite it).
Why Proprietary Data Beats Platitudes
- It’s Hard to Fake. Anyone can echo mantras; only Ahrefs (and perhaps a handful of their competitors) can analyze one million SERPs overnight. As of yet, AI can’t make up new data from thin air (though it will try and try, repeatedly, and to our collective detriment).
- It compounds authority. Every citation, backlink, and quote snowballs brand equity, raising your click-through rates in classic search and your presence in zero-click answers.
- It powers endless repurposing. One dataset seeds blog posts, LinkedIn carousels, podcast talking points, webinars, social content, and sales decks.
- It often aligns with product value or brand point of view. The closer the dataset lives to your core platform or offering, the clearer the bridge from insight to signup.
The Five Data Assets You Can Build Next Quarter
Below is the field guide we teach at Omniscient. Pick whichever suits your resources and risk tolerance.
Obviously this is not exhaustive, nor is it a scientifically ordained set of research categories. But I’ve found it incredibly useful in teaching teams and getting them excited about what sorts of data they can bring to the world.
| Asset Type | Source | Speed | PR Potential | Product Tie-in |
|---|---|---|---|---|
| 1. First-Party Data | Product logs | Medium | ★★★★☆ | Inherent |
| 2. Market Survey | Panel / users | Medium | ★★★★☆ | Narrative |
| 3. Micro / Case Study | Your own experiment | Fast | ★★☆☆☆ | Anecdotal |
| 4. Data Aggregation | Public datasets | Fast | ★★★☆☆ | Thematic |
| 5. Product-Led Experience | UGC / Freemium | Slow | ★★☆☆☆ | Direct |
1. First-Party Data
Ahrefs mined its backlink index to show links still correlate with rankings. The post itself earned—wait for it—hundreds of backlinks. Meta.
Starter question: What behavior do users perform inside our platform every day, or what do we have unique first party access to, that would answer business questions of our prospects or be interesting to the market?
2. Market Survey
Jasper polled 500+ marketing leaders on AI content adoption. They’ve done this annually starting in 2023, owning the narrative, but also driving leads, traffic, and links.
Starter question: Which hot-button topic in our industry keeps our buyers awake, yet no credible numbers exist?
3. Micro Study
LawnStarter founders documented a month of biphasic sleep (the post has since been deleted). The Guardian picked it up, among other publications and interviews. N=1 doesn’t guarantee prestige, but quirk plus velocity can often pick up substantial interest.
Starter question: What unconventional test can we run this week that either delights or horrifies our industry?
I have to also recall here one of my favorite examples of this from The Hustle, where they experimented with 30 days of LSD micro-dosing as purchased from the Dark Web.
4. Data Aggregation
Kevin Indig synthesized 19 studies on Google’s AI Overviews into a comprehensive review of their impact of performance and click-through-rates. Meta-analyses are incredibly useful, especially in high variance topics (think: nutrition and fitness). They’re also somewhat easier to produce, as the data already exists, it’s just up to you to synthesize it and derive aggregate insights.
Starter question: Can we be the librarian who catalogs scattered data into a definitive reference?
5. Product-Led Experiences
This is a bit different from the other categories, but I lump it under the objective of “creating unique value,” and that value is often data-driven in nature. For example, Zillow’s whole programmatic SEO play. Or Zapier’s integrations pages ranking for “<App A> to <App B>.” Each landing page is both utility (undermet user need) and SEO moat (with native product tie-in).
We’ve worked with clients on many such plays, involving integrations, user-generated content, templates, and freemium versions of their products with lots of search demand.
Starter question: Can we externalize a sliver of our platform—calculator, template, benchmark explorer—as a free mini-tool?
From Dataset to Dominance: The Distribution Playbook
Creating data assets is half the battle. Winning (and really leveraging all the work and costs that go into the work) requires determined amplification and repurposing.
- SEO sprinkling – Embed fresh stats in existing articles and new content to be published. Like himalayan sea salt, it really brings out the flavor in otherwise mundane or consensus topics.
- Digital PR – Pitch journalists with exclusive angles (“Only 12 % of CMOs trust AI copy unedited”). Tools like Stacker help break into legacy media.
- Organic social – Carve 10 graphics from the report; post a thread teasing one chart per day. Derive POVs and thought leadership from core takeaways. Post, post, post.
- IRL Assets – Print pocket-sized executive summaries; hand them out at trade-show booths. Print out books even (I saw LinkedIn do it at a recent meetup in NYC). Physical / print media FTW.
- Sales Enablement – Arm BDRs with one-pager infographics or documents. “According to our 3,000-site audit…..”
- LLM Seeding – Host key findings in crawlable HTML, not just a PDF. AKA, open up those gated reports.
Many more ways to repurpose if you really get creative (not to mention the easy ones like paid marketing / remarketing campaigns on LinkedIn).
How to Launch Your First (or Next) Data Play in 90 Days
- Inventory – Meet product, data, and CS teams. Take inventory of data available, business questions, customer pain points, and general curiosities.
- Choose the Juiciest Question – One whose answer changes behavior (and ideally backs into your product utility or brand point of view).
- Scope Small to Start – I’ve let unwieldy projects through due to unbridled ambition, but I’ve found it’s best to build the muscle with smaller scale projects and then scale up.
- Ship Minimum-Viable Study – Clean basics: methodology, sample size, anonymization, a handful of key charts and insights. Narrative woven through the central asset (typically whitepaper or blog post) to extract and showcase top level findings.
- Plan a Distribution Week – Outreach list, five social assets, email blast, two partner webinars.
- Measure & Iterate – Track backlinks, mentions, signups, but also qualitative buzz. Double down on what resonates.
And then do more of it. Learn, iterate, improve. It’s always a bit of guesswork to calibrate market demand with publisher demand with audience demand with your product with what data you can actually collect and surface, so while it helps to identify some of these audience questions up front, sometimes you strike gold and your report picks up surprising amounts of interest and natural tailwinds.
In an AI-flattened landscape, proprietary data is a way to stand out, create unique value, and leverage that value throughout most if not all of your other GTM efforts. It’s the raw ore that algorithms can’t conjure (with integrity), the narrative oxygen that journalists crave, and the conversion jet fuel your funnel secretly desires.
Go make something worth citing and sharing.
Want more insights like this? Subscribe to Field Notes
