Research Nuggets for the Solo Journey.

You've heard two voices.

"Use AI for everything. It's faster."

"Don't touch AI. It kills research quality."

Both miss the point.

The question isn't whether to use AI.

It's where.

A Beyond Research workshop tested AI across six research phases — from problem breakdown to report building.

They found something useful.

Different tasks carry different risks.

Some high. Some low. Some structured. Some ambiguous.

So they built a simple 4-quadrant model.

High Risk + Ambiguous → Don't use AI.

Problem breakdown. Primary data. Scanning sensitive sources.

These need ethical sensitivity, business nuance, sometimes political awareness.

AI can generate hypotheses with confidence. But it doesn't feel organizational tension.

High Risk + Structured → AI as Co-Pilot.

Scanning existing data. Designing methods. Drafting report content.

AI drafts plans. Suggests sampling. Compares methods. Catches gaps.

You hold the wheel. AI navigates.

Low Risk + Ambiguous → Thought Partner.

Brainstorming. Devil's advocate. Reframing problems.

When stakes are low but ambiguity is high, AI expands your thinking.

Like a junior researcher who reads fast — still needs supervision.

Low Risk + Structured → Automate.

Data cleaning. Formatting. Notes into outlines.

If imperfect won't break a decision, let AI run it.

Save your brain for thinking.

The shift after testing?

Participants moved from over-trust and over-fear to mature skepticism.

AI in research isn't a prompting skill.

It's context awareness. Risk assessment. And knowing who bears responsibility for the call.

Adapted from Beyond Research posts: BR158.

← Back to all nuggets

You've finished an interview and felt nothing.

You asked all the "right" questions. Every answer felt flat.

You checked the box.

But you didn't learn anything.

This happens when questions are too broad. Too polite. Too easy to forget.

"What do you think about our app?" isn't a question.

It's a placeholder.

The person you're talking to doesn't remember why they care enough to answer.

No story. No emotion. Just a shrug.

Here's the shift.

Instead of "What do you think about our app?"

Try "When was the last time our app made you frustrated — and what happened right before that?"

Same topic. Different effect.

The first is interrogation. The second is a memory trigger.

Four rules that change how researchers interview:

• Warm them up first. Don't go heavy right away. Start with easier terrain. Build trust before you ask them to dig into their own tensions.
• One question at a time, no branching. Ask. Listen. Wait. Then ask next. Branching questions confuse people and make them forget what they were saying.
• Make it easy to answer. Don't ask people to theorize. Ask them to remember. "What did you do?" beats "What do people usually do?"
• Give examples, so they know what you mean. "Tell me about a time you felt lost in a product" is clearer than "Tell me about confusing moments."

Once you change how you ask, everything changes.

People open up.

They share stories you didn't even know how to ask for.

The session stops feeling like an interview.

It starts feeling like a real conversation.

That's when insights start to breathe.

Adapted from Beyond Research posts: BR135, BR7, BR144, BR133.

← Back to all nuggets

You've seen this.

Solid research. Clear recommendations. Sharp visuals.

And then nobody moves.

The deck sits in a folder.

The stakeholder nods in the meeting — and schedules the next feature sprint like the research never happened.

This isn't because your research is weak.

It's because you're not speaking the language stakeholders actually respond to.

Sales teams learned this long ago.

Here are five psychological techniques that work in research, too.

1 · The Framing Effect.

It's not what you say. It's how you say it.

People react differently depending on the frame.

Instead of "Users struggle with the payment flow"…

…try "We're losing users at the exact moment they're ready to give us money."

Same fact. Different emotion.

The second makes stakeholders feel the urgency.

2 · Anchoring Bias.

The first number sets the mental tone.

In pricing, marketers show the expensive option first — so the cheaper one feels like a deal.

In research, start with the worst pain point first.

Everything else feels manageable when the anchor is high.

3 · Affordability Illusion.

$350 a year feels heavy.

Less than $1 a day feels chill.

Break big costs into smaller, relatable bites.

"Fixing this flow could save 5 seconds per user. That's 200 hours of user time a month."

Suddenly your recommendation feels valuable.

4 · The Contrast Effect.

Nothing is expensive until you compare it.

Show a bad design version first. Then the preferred one.

Your brain loves context.

The preferred option instantly feels smarter — and more premium.

5 · The Rule of 3.

Give people three options.

They almost never pick the cheapest.

People instinctively go for the middle one — it feels safe.

Whether it's feature tiers, pricing plans, or design options…

…how you structure choices shapes perception.

It's not manipulation.

It's understanding how people actually decide.

Your research is solid.

Now make it impossible to ignore.

Adapted from Beyond Research posts: BR125.

← Back to all nuggets

Here's a food critic analogy.

You try a new ramen shop.

You think: "The broth is cold. The noodles are soggy. The place is half-empty at lunch."

Those are findings.

Just facts.

They don't change the ramen shop's future.

Why?

Because knowing "the noodles are soggy" doesn't tell the chef what to do.

So you dig deeper.

You ask the chef why.

Turns out: supplier delivers 2 hours late every morning.

Staff rush the prep to catch up.

Noodles overcook.

Now that is your insight.

One is a symptom.

The other is the root cause that unlocks action.

Most research decks are actually finding decks.

Full of "Users didn't click the CTA."

Light on "Users say the CTA feels like a commitment they're not ready for."

One looks clean in a dashboard.

The other makes your team sweat and ship better.

The difference matters.

Business can only fix what they truly understand.

A finding is an observation.

An insight is an explanation.

One gets noted.

The other gets acted on.

Your job isn't just to collect findings.

It's to connect the dots until the why becomes clear.

Insights without explanation are just data points.

Insights with explanation unlock decisions.

Adapted from Beyond Research posts: BR106, BR102, BR8, BR129.

← Back to all nuggets

You've seen this in research conversations.

"Our research shows…"

…followed by what sounds like a casual chat with two people.

That might be useful.

But it's not the same credibility tier as a mixed-method study with 50+ participants, fresh data, and documented rigor.

The problem?

Researchers often can't articulate why one study is more credible than another.

Here's a simple scoring system that works.

Rate your research on four dimensions.

Each scored at three levels.

This isn't to shame your smaller studies.

It's to help you — and your stakeholders — understand what you can reasonably trust.

And where you need to do more work.

Adapted from Beyond Research posts: BR60, BR136, BR137, BR151.

← Back to all nuggets

You've felt lost deciding what to research.

The topic is real.

But you're not sure what kind of research it needs.

A survey? Interviews? Analysis?

This confusion usually traces back to not naming the gap you're trying to close.

There are five common research gaps.

Each one tells you what method works best.

1 · Knowledge Gap.

Something isn't known yet.

Example: online transport research is thick, but why do people trust a service before they've tried it?

What to do: run a study to gather comprehensive context.

2 · Theoretical Gap.

Existing theory doesn't explain a new phenomenon.

Example: work motivation theory explains productivity — but not why Gen-Z resigns despite good salaries.

What to do: retest, compare competing theories, or build a better framework.

3 · Methodological Gap.

The method doesn't fit the question.

Example: customer satisfaction studied only via surveys — when the experience needs interviews or observation.

What to do: mixed methods. Combine qual + quant.

4 · Empirical Gap.

Not enough real-world evidence.

Example: research says remote workers are more productive — but no Indonesian companies were sampled.

What to do: field data with locally relevant samples.

5 · Evidence Gap.

Existing findings are inconsistent or weak.

Example: one study says social media harms teen mental health; another says it depends on how they use it.

What to do: follow-up studies, larger samples, or a systematic review.

Name your gap.

And your research knows where to go.

Adapted from Beyond Research posts: BR176.

← Back to all nuggets

McKinsey CEO Bob Sternfels spoke at CES 2026 about which skills AI can't touch.

His framing:

AI is powerful at information-finding, data summarizing, pattern-spotting.

That's now a commodity.

So researchers have a choice.

Lean harder into what you do that machines can't.

He named three skills.

1 · Aspiration.

AI doesn't have aspirations.

It can predict. It can't decide what's worth fighting for.

A machine tells you what's possible.

You decide what's important.

This is where your judgment about business direction, user needs, and ethical priorities matters.

Machines optimize for patterns. Humans optimize for purpose.

2 · Judgment.

AI generates lots of options.

It can't choose between them with values.

You can see the full context. Understand the risks. Make ethical calls.

You know which choice aligns with who you want to be.

AI expands options.

Judgment is the uniquely human act of choosing.

3 · True Creativity.

AI works with patterns it's seen before.

It guesses "what's next?" based on what already is.

Humans can break patterns.

Imagine things that don't exist yet.

Pioneer new directions.

That's creativity.

Not just novel — directional.

It opens new paths instead of optimizing existing ones.

Sternfels also noted a shift.

Academic degrees are less dominant. Real skills are more valued.

The implication is clear:

Stop hiding behind credentials.

Build a portfolio that shows aspiration, judgment, and creative thinking.

If you're worried AI will replace you, you're competing in the wrong arena.

Compete on depth of judgment, clarity of aspiration, and boldness of creativity.

That's where you're irreplaceable.

Adapted from Beyond Research posts: BR152, BR140, BR161.

← Back to all nuggets

You've watched this happen.

The deck is polished.

The data is sound.

You present with confidence.

And then someone asks:

"So… what are we supposed to do with this?"

That's not a data problem.

It's a storytelling problem.

Here's what lands.

Not "Users had low task success rate."

But "One user said, 'I almost gave up, but I've come too far to quit' — and we almost lost them at that exact moment."

One is a metric.

The other is drama.

One gets filed.

The other gets quoted in the lunch room.

The TRS Framework.

Beyond Research practitioners found three levers that matter when presenting research.

T · Timing.

Is your insight arriving when decisions are being made?

If you present after the roadmap is locked, it arrives as opinion.

Know your product timeline.

Show up early.

Be part of the conversation before the bets are placed.

R · Relevance.

Does your insight connect to what stakeholders actually care about?

Your finding might be true.

But if it doesn't link to the KPI they're measured on, it bounces off.

Translate your insights into their language.

S · Storyline.

Can stakeholders remember your finding in a conversation — without the deck?

"Users felt the CTA was a commitment, not an invitation." — memorable.

"The CTA had low affinity scores." — not.

Wrap data into a story that's easy to hold, repeat, and act on.

The dream moment.

It isn't when someone asks for the full deck.

It's when someone quotes your insight without it.

Adapted from Beyond Research posts: BR116, BR98, BR101, BR104, BR92, BR115, BR123.

← Back to all nuggets