01Overview
The frequency illusion — also called the Baader-Meinhof phenomenon — is the experience of suddenly seeing something everywhere after noticing it for the first time. Coined by Arnold Zwicky in 2006, it names a familiar sensation that turns out to be genuinely misleading. Two mechanisms are at play: selective attention (noticing something once lowers the threshold for noticing it again) and confirmation bias (each subsequent sighting gets remembered as evidence of prevalence, while non-sightings quietly disappear).
For design teams, the phenomenon is most dangerous in research synthesis. The first strong theme a researcher encounters doesn't just feel significant — it primes them to find it everywhere that follows.
02Detailed explanation
The everyday version is harmless: you learn a new word and hear it twice in the next day. You buy a car and suddenly see your model everywhere. The word and the car were always there. Your attention wasn't primed for them until now. In design work, the same mechanism has consequences:
- After three interviews where users mention a pain point, it starts appearing in every subsequent transcript — because the researcher is now primed for it. The sample hasn't changed. The filter has.
- After one usability session where someone struggles with the nav, every subsequent session seems to confirm the nav is broken. The threshold for noticing nav-adjacent behaviour has dropped to near zero.
- After reading a competitor's case study about a specific interaction pattern, that pattern starts appearing in every product audit — including products that were audited before the case study, if the team goes back.
The pattern isn't becoming more common. The attention is becoming more selective. The distinction is invisible from the inside.
03Why it exists
Priming is a core memory mechanism — recently activated concepts are more accessible and therefore more likely to be noticed when encountered again. This is useful. It means recently learned information can be applied quickly without deliberate search.
Confirmation bias reinforces the pattern once primed: matches get noticed and remembered, non-matches don't register. Together they create a compelling, false signal of frequency that is almost indistinguishable from a genuine trend. Almost.
Three users mentioning something is not evidence that it's common. It might just mean you asked three users in a row.
04Effects on users
After a negative experience with a feature, users notice every friction point in similar features — including ones they didn't consciously notice before. One bad interaction primes the attentional filter for the whole category.
First-time users notice design inconsistencies that long-time users are completely blind to — because their mental model is still being formed, so every element is novel and attended to. Long-time users, conversely, stop noticing patterns they've normalised. Both are frequency illusions of a kind: one overestimates problems, one underestimates them.
After a friend mentions a product flaw, users start experiencing it — even when the flaw existed identically before the conversation.
05Effects on designers & teams
Research synthesis is most vulnerable. The first strong theme surfaced in a session primes the researcher to see it everywhere. If synthesis begins before all sessions are complete, early themes actively reshape how late sessions are read.
Competitive audits are a second hotspot. An audit begun after noticing one specific pattern in a competitor product will find that pattern in every subsequent review — sometimes in products where the pattern is marginal or absent. The audit finds what it was primed to find.
Design reviews have the same structure: when one comment is raised first, it shapes the direction of every subsequent observation. The room converges on the initial observation, not because it's the most important issue, but because it came first.
06Practical takeaways
- Count before you conclude. Once you've identified a theme, go back through all transcripts and count how many actually contain it before declaring it a pattern. The number will be smaller than it feels.
- Code data before naming themes. Raw tagging resists post-hoc narrative formation. If you name themes first, subsequent tags become confirmation. The sequence of operations matters.
- Rotate who leads research synthesis. The lead researcher's primed observations will dominate if they're the only analyst. Multiple independent coders — especially one who came to the research late — catch what the lead has normalised.
- Use blind competitive audits. Approach competitor products without a specific pattern hypothesis. Document what you observe first, then look for patterns. Hypothesis-first audits find the hypothesis.
- Notice when "I keep seeing this" is actually "I kept looking for this." They are not the same. The difference is a single sentence in your synthesis notes — and it changes what the finding means.
07Design examples
The theme that took over
The first strong observation in a research sprint primes every subsequent session to feel like it confirms the same finding. By the debrief, the theme feels unanimous. Go back and count — it's usually present in a minority of transcripts, just loudly in one.
The pattern that's suddenly everywhere
One notable interaction pattern in one competitor's product starts appearing in every audit that follows. The pattern was always there. The primed attention is new. The resulting audit overweights one pattern relative to every other finding.
The first comment shapes the rest
The first observation in a design review primes subsequent reviewers to notice the same class of issue. By the end, the team has extensively audited one dimension and left others untouched. Round-robin formats and silent individual note-taking before discussion break the cascade.
The metric that won't let go
Noticing one data anomaly creates confirmation-seeking behaviour in every subsequent dashboard review. The anomaly stays prominent in memory; the rest of the data is filtered through it. What looks like diligent monitoring is pattern-matching to a primed signal.
08Ethical risks
When frequency illusion operates in advocacy — "our users keep asking for X" — it can drive product decisions based on primed observation rather than measured need. The users who didn't ask for X don't appear in that story. They rarely do.
In organisations where a senior voice first identifies a pattern, frequency illusion can make that pattern seem ubiquitous across all subsequent research. The finding gets repeated, cited, and built into roadmaps — while the original observation was a single data point that became a lens.
Self-test: When did you last look for evidence that a theme you've identified doesn't apply — and find it?
10Suggested reading
Suggested reading is temporarily unavailable. Please check back later.