You’re sitting in a conference room or maybe just scrolling LinkedIn while half-watching your rank tracker refresh.
Someone’s on stage, or on your feed, showing a gorgeous line chart with a tidy, dramatic spike. “AI search is transforming discovery,” they beam, pointer hovering over the spike like it’s divine revelation, forgetting that without context, a spike is just a jumpy line on a chart.
You lean in. No date range. No axis labels. No mention of the SERP layout change that happened that week.
In the corner? Their company logo. And yes, they sell AI search tools.
If you work in SEO, you’ve seen this movie before. It’s not research. It’s marketing with a graph of authority taped to the front.
Why SEOs Are Prime Targets for PowerPoint Science
We are data people. We live in dashboards. We can recite CTR by position from memory.
But we’re also in an industry that moves at “Google just rolled out another unconfirmed update” speed. That means we sometimes grab onto anything that feels like clarity.
That’s why AI search research wrapped in the right language: “recent findings,” “user study,” “performance lift” can slip past our filters. And in that moment, you risk locking a strategy into place based on data that’s about as solid as a rank report without the location filter set.
The Marketing-First Research Playbook
Here’s how the sausage usually gets made:
- Publish a one-pager “study” → just enough to intrigue, never enough to audit.
- Gate the “full” version behind an email wall.
- Repackage for drip campaigns → the same graphs, sent on schedule.
- Slide into your inbox with an offer for their AI-powered ranking widget.
It’s a quick win for them, a lead list and a bump in brand recall.
But for you? No strategic depth. No durable insight. Another “industry study” that collapses when you ask it the same questions you’d ask a client about their analytics data.
The SEOs Framework for Calling BS on “Research”
You already have the instincts, the below just gives them sharper edges.
1. The Core Question Test
If they can’t sum up the research question in two clear sentences, it’s a thought piece with charts.
2. Data Lineage
Where’s the dataset from? How was it collected? What time frame? We’d never trust “organic traffic is up 200%” without knowing if it’s all branded queries.
3. Methodology Transparency
Could you follow their steps and reproduce it? Or is it more like a Google patent analysis where they “can’t share the client data” but still draw sweeping conclusions?
4. Incentive Audit
If the publisher’s business depends on you agreeing with their findings, that’s not a crime, but it is a reason to demand more proof.
5. Timestamp Rigor
In AI search, “Q1 2025” is vague to the point of useless. Exact dates matter. The difference between before and after an SGE UX change can be 48 hours.
Old Logic vs. New Logic (SEO Edition)
Old Logic:
- If it’s got charts, it must be legit.
- If it’s “recent,” it’s relevant.
- If other SEOs are sharing it, it must be good.
New Logic:
- If the methodology’s invisible, the claim’s suspect.
- Dates, scope, and context are non-negotiable.
- A chart without open data is decoration, not documentation.
Borrow Rigor from Beyond the SERP
E-commerce analysts don’t publish a conversion lift without sample size, split-test details and device segmentation.
UX researchers don’t claim “users prefer design A” without disclosing the task flow and how many failed it.
Apply that mindset to AI search “research,” and a lot of the fluff dies instantly. And what’s left? The stuff that actually makes you a smarter strategist.
A Better Game Plan for SEOs
If you’re publishing research:
- Be timestamped, transparent, reproducible.
- Share your thinking, not just your findings.
- Separate insight from sales → your best lead-gen will be the respect you earn.
If you’re reading research:
- Ask if you could replicate the result with the same dataset.
- Ask who benefits most if you believe it.
- Look for what’s not there → the null results, the negative cases, the SERP features that tanked the metric.
Final Word
Opinion pieces are fine, just label them.
The second you start calling opinion “research,” you’re playing a different game. And SEOs can smell spin from a mile away. We’ve been reverse-engineering Google’s talking points since before “not provided” was a thing.
In AI search, the people who’ll matter in 3 years won’t be the ones with the prettiest spike charts. They’ll be the ones who chose rigor over reach, clarity over clicks and truth over tactics.
Leave a Reply