When research gets cheap enough that it stops hurting

A friend of mine in consumer goods was venting to me the other day.

He said, "Run Liu, we wanted to do a competitive analysis. We pulled three people together and spent two weeks grinding away at it — and by the day the report came out, the market had already moved."

I asked him, "Why aren't you using AI?"

He said, "I am. I use it to write copy, schedule posts, generate scripts. It's great."

You see, that's exactly the problem.

The real blockage isn't at the outlet

When most people think of AI, their first reflex is "output".

Drafting documents, building spreadsheets, auto-replying, batch-generating. Everyone's busy optimizing how much water comes out of the faucet.

But think about it — no matter how big the faucet, if the intake pipe is only as thick as your finger, how much water can actually come out?

What's the intake pipe?

It's research.

It's who your buyers really are, what your competitors are doing, what customers are complaining about, what kind of hook will move people. This is the intelligence that feeds strategy, feeds positioning, feeds every line of copy.

It's just that few people ever call it "research". Everyone calls it "looking a few things up", "feeling out the landscape", "seeing what's happening in the industry".

Because it's too expensive, you can't go deep. Because you can't go deep, every time you just cobble something together at the last minute.

Time is tight, so you look at two fewer competitors. Budget is tight, so you skip that niche audience. You're rushing to get out the door, so you skim a hundred reviews in a single glance.

The result: the foundation under your strategy is thin. And when the foundation is thin, whatever you build on top wobbles.

A week shrinks to an afternoon — and then what

I once heard a real-world account that really stuck with me.

A head of research said that in the past, running one market scan, working through the competitors, and reading a few hundred customer reviews used to take her a full week.

And now?

An afternoon.

You see, that's the key. When most people hear this, the first thing they think is "Wow, four and a half days saved."

But saving time is just the surface.

What actually changes is behavior.

What does "behavior changes" mean?

When research is fast and cheap, it stops hurting to do it.

In the past, building a persona for one niche segment took three days. You'd hesitate — is it worth it? Now it takes twenty minutes. You don't hesitate.

That hypothesis you always wanted to test but never dared spend the time on — you validate it now, almost in passing.

A second competitor, a third packaging option, a fourth channel — these used to be "forget it"; now they're "might as well check".

The same team, the same budget, and the volume of research more than doubles. The intelligence behind every quarter's decisions is far thicker than before.

This isn't saving time.

This is turning "I can't bear to spend it" into "it doesn't even hurt to spend it".

Four jobs, all redefined

I thought about it myself. The research that a marketing team's daily survival depends on really comes down to about four jobs.

The first is figuring out who your buyers are.

In the past, public information was scattered and thin, and piecing together a persona was like assembling a puzzle with missing pieces. Now, fragmented clues can quickly be assembled into a picture that's coherent — one you can actually act on.

The second is watching your competitors.

It used to be "once a year", and once it was done the report got locked in a drawer, because nobody had time to maintain it. Now it can become an always-on radar that keeps updating — because it's cheap enough to keep running.

The third is listening to what the market is saying.

You couldn't even capture the full voice of a single channel, let alone all of them. Now, reviews, posts, live flying comments （danmaku — real-time comments that fly across the video screen）, and DMs can be kneaded together until the genuine sentiment comes through.

The fourth is finding the angle for your content.

What kind of framework will break through to a wider audience, what kind of hook will really bite. This used to depend on inspiration and luck; now it depends on gathering signals scattered everywhere and pulling them into focus.

You see, what do these four jobs have in common?

None of them are new tasks invented by AI. They are old tasks that AI has re-priced.

Once the price changes, behavior changes. And when behavior changes, the quality of decisions changes.

The real turning point

I thought about this some more afterwards, and the more I thought about it, the more interesting it seemed.

The moment someone starts to see themselves as "a person who does research", the world in front of them widens.

They start asking: is this judgment of mine backed by data? Can I spend ten minutes validating this hypothesis? When was the last time I updated this competitor profile?

Research is no longer that errand you're forced to run before you can write the proposal.

It becomes a habit, a kind of muscle memory.

And all good strategy, at its core, starts with asking the right question and finding the right evidence.

My friend came back to me later.

He said, "I did what you said — I turned that competitive workflow into always-on monitoring. It runs automatically once a week, and all I do is look at what changed."

I asked him how it felt.

He thought for a moment and said, "It feels a little unreal. I used to feel that research was a luxury; now it's as cheap as tap water."

Yes.

When intelligence is no longer scarce, the scarce thing becomes whether you're willing to bend down and draw one more bucket.