Table of Contents
Ever tried writing an unreliable narrator and felt like you were juggling knives? I have. The hard part isn’t the twist itself—it’s getting readers to buy in early, then slowly realizing, “Wait… that doesn’t add up.”
Most drafts I see (and I’ve done this too) rush the reveal. They drop the “gotcha” too soon, and the story ends up feeling like a trick instead of an earned moment. What I’m after is different: I want the unreliability to feel inevitable in hindsight.
In this post, I’ll walk you through a practical way to build unreliable narration step-by-step. I’m also going to show you exactly what to write on the page early on, what to change later, and what the reveal should prove—so you’re not stuck with vague “add contradictions” advice.
Quick overview of the method: establish trust with specific, believable observations; plant contradictions on purpose (not randomly); make the reason for unreliability connect to character; then pay it off with a reveal that feels fair.
Key Takeaways
- Build early trust with concrete details (time, place, sensory specifics). Then hide one “oddity” that doesn’t matter… yet.
- Use contradictions in three flavors: memory errors (unintentional), intentional lies, and selective omission. Mix them so the narrator feels real.
- Give the unreliability a job. It should explain why the narrator misses something, misreads something, or avoids the truth.
- Show bias through internal reactions—what they notice, what they dismiss, what they refuse to describe. Don’t just tell us they’re biased.
- If you use multiple narrators, keep the voices distinct and let each one contradict the others in a different way (not all of them “lie” the same way).
- Plant unreliability clues immediately (Chapter 1), then escalate them every few scenes. By Chapter 7, readers should have enough breadcrumbs to suspect the twist.
- Design the reveal so it answers a specific question you raised earlier. Fair twists feel like “ohhh, that’s what that meant.”
- Use the narrator’s unreliability to deepen character and plot. The reveal should change what the protagonist (and readers) do next.

1. Start with a Trustworthy Voice for Your Narrator
The first step in writing unreliable narration is making your narrator feel competent and present early on. Not “mysterious.” Not “vague.” Present.
In my own drafts, the narrator usually wins trust when I give them details that only someone there would notice—things like the exact time, the smell in a room, the way a character’s hands move when they lie.
Here’s what I mean, step-by-step, using a simple example story:
Page 1 (Chapter 1): Write trust + one harmless oddity.
Example narration (what you’d write): “The knock came at 7:12. The neighbor’s porch light flickered twice before it steadied. I remember the blue mug because I set it under the faucet to catch the drip.”
That’s believable, right? Specific. Visual. Sensory.
Now add one oddity that doesn’t scream “plot.” Maybe: “When I turned the knob, it didn’t squeak. That surprised me.” Why mention the squeak? Because later, the reveal will show why it mattered.
What I noticed in workshop feedback: readers didn’t question the narrator because of the “oddity.” They questioned them later because the oddity reappeared and contradicted something the narrator said.
Rule of thumb: trust first, confusion later. If the narrator sounds unreliable on page 1, readers stop reading. They don’t “enjoy the mystery”—they just brace for disappointment.
So, establish credibility with accurate observations and a consistent voice. Then use the trust you earned to make the later contradictions feel sharper, not random.
2. Hide Clues That Reveal Unreliability Gradually
Gradual unreliability is the sweet spot. If you dump contradictions all at once, it feels like the narrator is performing. If you never contradict them, the reveal feels unearned.
Let the issues grow. Small at first. Then bigger. Then undeniable.
There are three contradiction types I like to rotate:
- Memory error: they’re wrong without realizing it. (They remember the blue mug, but it was actually red.)
- Intentional lie: they know better and choose a false detail. (They say they didn’t open the door.)
- Selective omission: they tell the truth… but skip the part that changes everything. (They mention the knock, but not who answered.)
Page 20 (early scenes): introduce memory error in a way that seems minor. Keep it under one sentence when possible. “I’m pretty sure the mug was blue.” That little uncertainty is gold.
Page 50 (midpoint): introduce a second contradiction that forces the reader to compare. Maybe the neighbor later says, “Your porch light never flickers.” Now the narrator’s earlier detail doesn’t just feel wrong—it feels impossible.
Page 90 (late escalation): switch to selective omission. The narrator can’t keep explaining everything away. Instead, they avoid a question: “I didn’t think about the door again. Not until later.”
Also: let other characters react. If someone says, “That’s not what happened,” but the narrator keeps moving like nothing happened, readers start tracking the gap.
In my experience, the “unreliable” feeling comes less from the contradiction itself and more from how the narrator responds to pressure—do they get defensive? Do they change the subject? Do they get weirdly specific about irrelevant details?
3. Make Unreliability Purposeful and Clear
Unreliability shouldn’t be random. If it’s just there to surprise readers, it’ll feel cheap.
Instead, ask one question while you outline: Why does this narrator need the reader to believe them?
Sometimes the answer is ugly. Sometimes it’s emotional. Sometimes it’s practical.
Here are a few “jobs” unreliability can do:
- Protect someone: they lie to keep a loved one out of trouble.
- Protect themselves: they rewrite events to avoid guilt.
- Survive a memory: they confabulate because the real memory is unbearable.
- Maintain control: they shape the story to keep people following them.
In a draft I revised recently, my narrator kept “forgetting” one detail. The reveal landed flat—until I rewrote the reason: they weren’t forgetting; they were blocking the moment because it connected to a specific trauma. Suddenly the omissions weren’t gimmicks. They were character.
Mini-checklist (use this while drafting):
- Can I point to one scene where the narrator’s unreliability protects something?
- Does the narrator’s tone match the reason? (Guilty, calm, angry, terrified?)
- Does the reveal explain the unreliability—not just the plot?
Do that, and the twist stops feeling like a magic trick. It becomes part of the character’s journey.
4. Show the Narrator’s Perspective and Bias Clearly
Here’s the thing: unreliability isn’t always “lying.” A lot of the time, it’s interpretation.
To make that clear, write the narrator’s bias in the way they notice things. What do they zoom in on? What do they ignore? What do they assume?
I like to use first-person for this because it lets you show the narrator’s internal logic. But even in first person, you can’t just say “I was paranoid.” You need to show the paranoia working.
Example: same event, different mental lens.
Scene: a shadow moves behind a curtain.
Reliable-style: “A shadow moved behind the curtain. I checked the window and found it latched.”
Biased-style: “The curtain twitched like it was breathing. People don’t move like that unless they’re hiding something.”
Notice what changed? The second version makes a leap. That leap is the bias. Later, when the reveal proves the narrator’s assumption was wrong, readers feel the distortion.
Also, show internal reactions that don’t match what’s happening externally. If the narrator is calm while danger escalates, that mismatch is creepy in a good way.
And please—avoid the “tell” trap. Instead of “I lied because I was traumatized,” show the narrator avoiding a specific topic, changing the subject, or focusing on irrelevant sensory details to keep control of the scene.

5. Incorporate Multiple Narrators for Depth and Complexity
Multiple narrators can be amazing for unreliable narration because it gives readers multiple “truths” to compare. But there’s a catch: if everyone is unreliable in the same way, it turns into noise.
I recommend giving each narrator a different flaw:
- Narrator A: confident but wrong (overestimates what they saw).
- Narrator B: honest but incomplete (they omit what they don’t know).
- Narrator C: strategic (they lie to protect their agenda).
Start by establishing each voice clearly—different sentence rhythm, different priorities, different vocabulary if it fits the character.
Then weave them so their versions conflict on key points. Not everything. Just the points that matter to the plot.
Example (mystery scene): one character says, “I heard the scream at 7:12.” Another says, “There was no scream—only the porch light.” A third insists, “The door was already open.”
Notice how each version points in a different direction. That’s how you encourage readers to assemble the real story themselves.
And yes, occasionally show reactions between characters. If Narrator A hears Narrator B’s version and goes quiet, readers will remember that silence later.
Want more help with multi-viewpoint structure? Check out tips on writing multiple narrators.
6. Introduce Unreliability Early and Keep the Tension Building
Don’t wait for the “big reveal” to start planting doubts. By the time you reveal anything, readers should already have a reason to question the narrator.
Here’s the approach I use when I outline:
- Chapter 1: establish trust + one small oddity
- Chapter 2–3: add a second contradiction that’s still easy to explain away
- Chapter 4–5: introduce a pattern (the narrator keeps getting the same kind of detail wrong)
- Chapter 6–7: escalate with an omission or a direct conflict with another character’s account
For example, early on the narrator might forget a minor detail (“I think the mug was blue”). Later, they misremember something that should be impossible to miss (“The porch light never flickered”). Finally, they refuse to answer a question that would clarify everything.
Foreshadowing works best when it doesn’t feel like foreshadowing. A tone shift, a repeated phrase, a weirdly specific memory—that’s not a billboard. It’s a breadcrumb.
If you want to build tension alongside this, see how to build tension in writing.
7. Set Up Your Reveal by Planting Clues and Red Herrings
A fair reveal isn’t just “surprising.” It’s provable in hindsight.
That means you need two things working together:
- Clues that point to the truth (even if readers don’t connect them yet)
- Red herrings that make the wrong conclusion feel tempting
Design your clues so they look ordinary at first. Then make them matter later.
Here’s a concrete example of how this can land:
Early: the narrator mentions the porch light flickering twice.
Middle: another character mentions the porch light never flickers.
Late: the reveal shows the narrator wasn’t lying about the light—it was flickering because of a specific event they were too distracted (or too traumatized) to notice properly.
In other words, the reveal answers a question you raised: “Why did the narrator focus on that detail?”
Red herrings should also be grounded. Don’t toss in random suspects just to confuse people. Give them a reason to be suspicious, and then let the story show why they’re not the answer.
If you’re aiming for a twist that doesn’t feel unearned, you’ll probably like tips for setting up effective storytelling surprises.
8. Use Unreliable Narration to Deepen Character and Plot Development
The best unreliable narration does more than shock readers. It reveals how the character thinks—and it changes the plot in meaningful ways.
So instead of asking, “How do I make the narrator wrong?” I ask, “What does being wrong cost them?”
If the narrator distorts events because of trauma, then the distortion should affect their choices: who they trust, what they investigate, what they avoid.
If they lie intentionally, then their lie creates consequences. They’ll have to keep track of what they said. That’s where plot tension comes from.
Example: a narrator omits that they opened the door. Later, when evidence shows the door was opened from inside, they panic—not because the plot changed, but because their lie collapses.
That’s character development. That’s cause and effect.
If you want to connect narrative voice to mental states, check out how mental health shapes storytelling.
When you use unreliable narration this way, it stops being a gimmick. It becomes a lens—one that readers can’t stop looking through.
FAQs
Give them credible, specific observations right away—time, sensory detail, and straightforward cause-and-effect. Readers trust competence. Then you can introduce one small oddity that only becomes suspicious later.
Use a pattern: start with minor memory uncertainty, then escalate to a contradiction that conflicts with another character’s account. Later, add omission—what they refuse to say—so the unreliability becomes harder to “explain away.”
Connect the unreliability to a clear motive or mental state. Then make the clues match that motive: if it’s trauma, the narrator avoids specific moments; if it’s guilt, they distort details that would implicate them. Clarity comes from consistency.
Write their bias into the narration: what they notice, what they dismiss, and how they interpret neutral facts. Show internal reactions and leaps in logic, not just labels like “I was paranoid.”



