How to Truly Turn Tobacco Content User Relationships into Assets That Sustainably Generate Profit

I have always believed that there is an easily misjudged issue in content operations: Having many followers does not mean you have many customer assets.

An account with 100,000 followers looks valuable; but if you do not know why these 100,000 people followed you, how many have returned in the past six months, who only reads headlines, who is willing to ask specific questions, and who continuously engages with you, then those 100,000 people may just be a nice number.

True customer assets are relationships that can be continuously understood, continuously engaged, and continuously generate value within compliance boundaries.

This is especially true for tobacco-related content.

This is not an area where you can simply copy the private-domain playbook of ordinary consumer goods. China's Advertising Law imposes clear restrictions on tobacco advertising, including a ban on tobacco ads in mass media, public places, public transportation, and outdoors, and a prohibition on sending any form of tobacco advertising to minors.

Therefore, the "profit" I discuss is not about teaching people to bypass regulations to sell tobacco, but rather addressing a broader operational question: How can user relationships formed around tobacco industry knowledge, policy information, history and culture, health risk education, and industry services become long-term, stable, and compliant business assets.

I value three things most: data, trust, and engagement.

I. Customer Assets Are Not Follower Count, But Your Understanding of Users

Many content teams, when they first start user management, typically have only one table:

User nickname, Follower count, Contact information.

In my view, this has almost no operational value.

Truly useful data should at least answer seven questions:

Where did they come from?

Why did they follow?

What do they read?

How often do they return?

Do they proactively ask questions?

What life cycle stage are they in now?

Suppose a content project has 10,000 users.

After 30 days of observation, it was found that:

• 10,000 followers;
• 6,200 people have read at least 3 pieces of content;
• 1,800 people engage at least 2 times per month;
• 420 people have actively participated in discussions for 3 consecutive months;
• 110 people have become stable consultation or long-term communication targets.

At this point, what is truly worth operating is not "10,000 people," but the last 420 or even 110 people.

This is not to say other users have no value, but that different users should be managed with different cost levels of operation.

For a user who reads an article once a month, you do not need to assign manual customer service follow-up.

For a user who has proactively asked professional questions for three consecutive months, and each question is highly relevant to the same field, their relationship value is clearly different.

Therefore, I divide user data into four layers:

Layer 1: Identity Data

For example, source channel, first follow time, age group compliance, geographic location, device, content entry point.

Layer 2: Behavioral Data

Reading, bookmarking, commenting, sharing, searching, private messaging, returning visits.

Layer 3: Interest Data

Is the user consistently interested in policies, industry history, health risks, product knowledge, or career-related information?

Layer 4: Relationship Data

This is the most easily overlooked layer.

For example:

What questions has the user previously asked?

Have you replied?

Did they acknowledge your answer?

Did they come back later to continue the discussion?

This type of data is actually more valuable than "like counts."

Because a like only proves a single action, while continuous discussion proves a relationship.

II. The Most Important Role of Data Is Not Statistics, But Helping You Make Decisions

Many teams end up turning user data into a giant Excel spreadsheet.

Hundreds of columns, looks very professional, but in reality, no one uses it.

I prefer to keep only the data that will change operational actions.

For example, a user has been inactive for 60 consecutive days.

The real value of this data point is not to tell you "the silence rate is 37%," but to prompt an action:

Why is this user silent?

If further investigation reveals that silent users mainly come from a certain content entry point, then the problem may not be the users themselves at all, but a change in content quality.

Consider a simulated case:

In March 2026, a content team had 24,000 historical followers.

After re-organizing the past 90 days of behavior, they found:

• 24,000 people in the database;
• 8,300 people had no behavior in the past 30 days;
• 4,200 people only read headlines;
• 1,900 people frequently bookmarked but rarely commented;
• 670 people commented consistently;
• 210 people proactively asked questions.

The team's original plan was to send unified content to all 24,000 people.

I believe this was a mistake.

They eventually changed to:

For the 8,300 silent users, reduce manual investment;

For the 1,900 high-bookmark users, increase in-depth material content;

For the 670 high-engagement users, establish a question bank;

For the 210 proactively consulting users, focus on maintaining long-term relationships.

Three months later, the team found a very interesting result:

The people who generated truly high-quality feedback were not those who commented the most, but those who bookmarked for a long time, occasionally asked questions, and came back after some time to continue the discussion.

This confirmed my view even more:

**User value cannot be simply measured by engagement frequency; it should be judged by whether engagement forms a continuous relationship.**

III. The Second Layer Asset: Trust

If data solves the problem of "I know who the user is," then trust solves:

Why is the user willing to continue listening to you?

Tobacco content is particularly prone to this problem.

Because after reading an article, users may ask:

"Did you get paid?"

"Are you promoting a certain product?"

"Is your data reliable?"

"Who produced this conclusion?"

Once users begin to doubt the source of the content, all subsequent commercial value rapidly declines.

Therefore, I believe content operators should truly build not a "marketing persona," but a track record of credibility.

For example, every time policy-related content is published, cite the source whenever possible.

For health risk data, do not invent numbers.

For industry data, note the statistical caliber and timeframe.

When errors are discovered, make public corrections.

If a conclusion cannot be confirmed, simply tell the user:

"I currently have not found sufficiently reliable data to support this claim."

This statement may seem to reduce the certainty of the content, but it actually increases trust.

Because users will gradually discover:

You do not rush to give an answer every time.

This kind of trust is very slow to build.

But once formed, its value is also very high.

IV. I Particularly Oppose a Kind of "Pseudo-Trust"

Some accounts create a sense of authority through extensive technical jargon, complex charts, and expert photos.

It works in the short term.

But it is very dangerous in the long term.

Because trust is not visual packaging.

If an account tells users every day for six months:

"This method is very effective."

"This product is very excellent."

"This trend will definitely happen."

Users will eventually realize that it actually has no judgment capability, it is just continuously outputting conclusions.

A truly professional account should allow itself to say:

"This matter currently cannot be judged."

And should even allow itself to say:

"My judgment last month was wrong."

This kind of correction itself is a customer asset.

I divide user trust into three stages:

Stage 1: Content Trustworthiness

The user feels that what you say is generally reliable.

Stage 2: Judgment Trustworthiness

The user begins to believe that your analytical approach is worth considering.

Stage 3: Relationship Trustworthiness

The user is willing to continue coming back to discuss with you even if they disagree with your conclusions.

Stage 3 is what truly approaches a long-term customer asset.

V. The Third Layer Asset: Engagement

Engagement is not "replying to comments."

This is where I believe many content teams waste money most easily.

An account that replies "Thanks," "Nice," "Bookmarked" 500 times a day looks very busy.

But these engagements may not necessarily form assets.

Truly valuable engagement should be able to change the next content production.

For example, a user asks:

"What exactly will the recent policy changes affect?"

After you answer, add this question to the question bank.

Then you find that 87 people have asked similar questions in a month.

At this point, it is no longer a user question, but has become a content demand.

So you can produce:

"Timeline of Relevant Policy Changes Over the Past Year"

The user's question becomes content.

Content generates new questions.

Questions enter the database again.

This is when a true cycle forms:

User → Question → Content → Feedback → Data → New Content.

I believe this is what engagement assets truly mean.

VI. Establish a Six-Level User Relationship Model

If I were to rebuild a content project, I would not directly divide users into "regular fans" and "VIPs."

I would first divide them into six levels.

Level 1: Browsing Users

Only read content.

System auto-records only, no extensive manual intervention.

Level 2: Light Engagement Users

Like, bookmark, share.

Focus on observing content preferences.

Level 3: Commenting Users

Start expressing their own judgments.

At this stage, pay attention to what users truly care about.

Level 4: Proactively Consulting Users

Users begin to proactively seek your help.

At this point, trust has already emerged.

Level 5: Continuously Engaging Users

Users come back to continue discussing previous issues.

This is a very important relationship asset.

Level 6: Long-Term Value Users

Not just one engagement, but continuously generating content feedback, industry information, professional needs, or compliant business cooperation opportunities.

Six-level users cannot be managed with the same operational approach.

The biggest mistake is:

Using the cost of Level 1 users to maintain Level 6 users.

And vice versa.

If you arrange manual customer service for every casual browsing user, profits will soon be eaten up by labor costs.

VII. Customer Assets Must Ultimately Enter the Profit Model

If user relationships cannot generate any sustainable value, then they are merely an operating cost.

I like to use a very simple framework:

Customer Asset Value = Effective User Scale × Activity Level × Trust Level × Sustainable Value Exchange Capacity × Lifecycle

The most easily overlooked term in this is:

Lifecycle.

A user who contributes 100 yuan today does not mean they are worth 100 yuan.

If they only exist for one day, their lifecycle value may be very low.

Conversely, a user who maintains stable engagement for three consecutive years, even if their single-period value is not high, may form greater long-term value.

For example, a compliant content project:

In the first year, there are 5,000 effective users.

If each user's average annual contribution value is 80 yuan, the theoretical annual value is 400,000 yuan.

But if only 30% of users stay in the second year, and the cost of acquiring new users is high, then the first year's operational results have not truly solidified.

So one of the most important indicators of customer assets is not:

"How much money did we earn this year?"

But:

**"How much of the relationships established last year remain this year?"**

VIII. A More Practical Profit Judgment: Customer Assets Cannot Be Measured by Revenue Alone

Assume two users:

User A generates 1,200 yuan in value per year but requires 20 hours of manual customer service.

User B generates 600 yuan in value per year but only requires 2 hours of manual work.

If you only look at revenue, A is more important.

If you look at profit, B may be far superior to A.

Therefore, customer assets should include an indicator:

Relationship maintenance cost.

A simple calculation:

User Annual Contribution Value − User Annual Maintenance Cost = User Net Relationship Value

Once this indicator is introduced, many so-called "high-value users" will be reordered.

The most common mistake I have seen is:

Mistaking the hardest-to-serve person for the most important person.

In reality, it is not.

A truly excellent long-term customer often has three characteristics:

Their needs are clear.

Their trust is stable.

They do not require extensive repeated education.

Such users are more likely to form sustainable profits.

IX. Case Study: How a 20,000-Follower Account Reorganized User Assets

The following case is a constructed operational review scenario for illustrative purposes and does not correspond to a specific real project.

Assume that in April 2026, in Shanghai, a tobacco industry knowledge content team reorganized an account with 20,000 historical followers.

In the past year, the team mainly looked at two indicators:

Follower growth and single-article readership.

Starting April 10, they spent 9 days reorganizing the past 180 days of data.

They finally obtained:

Originally, the team believed their core users were the 4,900 people.

After reanalysis, they decided to focus resources on the 860 and 96 people.

They specifically did three things.

First, establish a user question bank.

Break down the past 6 months of consultations into:

Policy questions 31%

Industry questions 24%

Health risk questions 19%

History and culture questions 14%

Other 12%.

Second, establish a content feedback form.

After each piece of content is published, not only look at readership, but also:

Bookmark rate, comment rate, return visit rate, question generation rate.

Third, establish a trust record.

Whenever data errors, policy updates, or previously made judgments need revision, record them.

By July, the team noticed a change:

The number of single-article hits did not increase significantly.

But the continuous user return rate notably increased.

More importantly, the team no longer relied on finding a "hit topic" every week.

Because they already had a continuously growing question bank.

This is the signal that customer assets are beginning to break away from platform traffic.

X. What Is Truly Valuable Is Not the User Database, But "User Memory"

I believe this is the most underestimated part of customer assets.

A database can only tell you:

Zhang San has read 10 articles.

User memory tells you:

Zhang San has been consistently interested in a certain type of policy issue for the past six months, asked a specific question 3 months ago, and later came back to follow up on related changes.

The operational value of the two is completely different.

Therefore, I recommend establishing "user relationship cards" rather than endlessly increasing fields.

A relationship card only needs to record:

Who the user is.

Why they followed.

What they care about.

What the most recent interaction was.

What important questions they have had in the past.

What relationship stage they are currently in.

What value should be provided next.

This way, user relationships transform from "data records" into "operational memory."

XI. The Most Dangerous Practice Is Treating All Users as Sales Targets

This is a very clear judgment I have about tobacco content operations.

If an account constantly seeks monetization opportunities from the moment a user first follows, the user relationship will quickly become commodified.

User just followed:

Push once.

User comments:

Push again.

User sends a private message:

Push again.

Eventually, the user will form a very clear expectation:

"As long as I interact with this account, I will be sold to."

Trust will decline.

Moreover, tobacco advertising itself has clear legal boundaries, so ordinary private-domain marketing logic cannot be directly applied to tobacco product promotion. China's current Advertising Law has very clear restrictions on tobacco advertising, and relevant international rules also emphasize broad restrictions on tobacco advertising, promotion, and sponsorship.

Therefore, I agree more with a sequence:

First provide value.

Then build trust.

Then form stable engagement.

Finally find legal and compliant ways of value exchange.

Rather than:

First add people in, then figure out how to make money.

XII. What Customer Assets Should Truly Resist Is "Traffic Fluctuation"

Platform traffic will inevitably fluctuate.

100,000 reads one month, maybe only 30,000 the next.

Recommendation algorithms will change.

Search rules will change.

Content preferences will change.

Platform policies will change.

If a content team's entire value is built on platform recommendation traffic, then it essentially has no customer assets.

True assets should be:

When the platform gives you traffic, you can identify which people are worth long-term operation;

When the platform does not give you traffic, you still have a group of stable relationships;

When content explodes, you can quickly understand which users have become interested;

When content is in a low period, you know which users are still willing to return.

This is what I understand as "long-term customer assets."

XIII. Where I Would Put 80% of My Operational Energy

If I were actually managing a content pool of 20,000 users, I would not distribute my time evenly.

I would put most of my energy into three places.

First, find truly active people.

Not the ones with the most followers, but the ones with continuous engagement.

Second, build a question bank.

The questions users ask are themselves demand signals for future content and services.

Third, maintain credibility.

Correct errors promptly, be transparent about sources whenever possible, and do not pretend to know things that cannot be confirmed.

Because data solves:

"I know you."

Trust solves:

"You are willing to believe me."

Engagement solves:

"We will continue to have a relationship."

Only when these three things come together can true customer assets be formed.

XIV. In the End, Only Four Numbers Need to Be Managed

If I could keep only four indicators, I would choose:

1. 30-Day Effective Activity Rate

Tells you whether the user relationship is declining.

2. 90-Day Continuous Engagement Rate

Tells you whether a real relationship has formed.

3. User Net Value

Tells you whether the value contributed by the user covers the maintenance cost.

4. User Lifetime Value

Tells you whether this is a long-term asset.

Follower count can grow very fast.

Readership can also suddenly explode.

But customer assets grow slowly.

They require users to come back again and again, ask questions again and again, verify your judgments again and again, and then gradually form stable expectations.

This is also what I believe is the most easily overlooked thing in content operations:

True customer assets are not how many users you have, but how many users are still willing to return after time has passed.

If a tobacco-related content project can solidify user data, content credibility, and continuous engagement while strictly complying with advertising and platform rules, then its commercial value will no longer depend entirely on a single hit article.

Traffic is responsible for bringing people in.

Data is responsible for getting to know people.

Trust is responsible for keeping people.

Engagement is responsible for sustaining the relationship.

And what can truly generate long-term profits is precisely this relationship itself.