How Pure Online Tobacco Health Content Can Control Customer Acquisition Costs and Achieve Stable Profits

Since the second half of 2021, I have moved almost my entire advertising budget from feed ads to "content assets": long-form articles on my site, answers on Zhihu, WeChat official account serials, plus a private domain on WeCom (Enterprise WeChat). The vertical is tobacco-related health science content——quit-smoking pathways, oral/nasal damage, nicotine dependence, household secondhand and thirdhand smoke cleanup——we don't sell cigarettes; we sell information structures, toolkits, and light consulting.

In these four years, my deepest conclusion is only one sentence: whether pure online tobacco health content can make money does not depend on whether you can produce viral hits, but on whether you can squeeze customer acquisition cost (CAC) into a predictable range and use a funnel to turn reading into compound-interest assets. Traffic fluctuation is the norm; whether the ledger stays stable depends on whether SEO accumulation, private-domain tiering, and the conversion funnel are managed by you as "production costs."

Below I write in the order I actually operate: first SEO to lower CAC, then the private domain to raise LTV, and finally the funnel numbers to cap profit.

1. Why this vertical is most likely to "look like it has traffic but has no profit when you do the math"

Tobacco health topics naturally have search volume: reasons for failed quit attempts, gum recession and smoking, how long until breath improves after quitting, secondhand smoke and children, whether e-cigarettes are safer……users arrive with specific problems, and intent is stronger than in general health.

But there are three hard constraints that most people hit in their first year:

1. Narrow advertising channels. Tobacco, nicotine, and simulated-smoking categories face heavy restrictions in mainstream search and social advertising. A common saying in the e-cigarette export circle: when hard ads on Google/Meta etc. are restricted, you can only rely on SEO, content, influencers/affiliate paths to acquire customers. If a health-science account flirts with the edge by "selling harm-reduction devices," bans and ad rejections will directly blow up CAC.
2. Public-domain distribution gets more expensive. As e-commerce and content platforms raise monetization rates, brands that rely on buying traffic generally feel "customer acquisition costs rising year by year"; in medical-health private-domain practice, some organizations also use private domains to compress acquisition cost substantially and push repurchase above public-domain advertising——this isn't motivational talk, it confirms that "public domains are only for testing samples, private domains are for cultivating profit."
3. Tension between content compliance and conversion. Write too softly, conversion drops; write too much like selling, trust collapses and account risk rises. In March 2022 I tested "21-day quit-smoking check-in + external-link mini-store toolkit" on a short-video platform: burned about 4200 yuan in ads in 7 days, got 186 people to add WeChat, closed 11 orders at 99 yuan per order, which worked out to about 38 yuan customer acquisition cost per customer and ROI below 0.3. The problem wasn't the product, it was that the promise at the top of the funnel didn't match delivery at the back end, and refunds plus bad reviews raised subsequent acquisition difficulty.

So the operating principles I set for myself are:

• Public domain is only for "precipitable entry points" (search, long-form, Q&A), minimize buying "one-time exposure."
• The profit center sits in the private domain and repeatable services, not single-article read counts.
• Review all channels with the same funnel metrics, otherwise you'll always feel like "you're just barely missing it."

2. SEO: the real way to squeeze marginal acquisition cost toward "near zero"

2.1 What I'm actually optimizing is not keyword density but the "problem-intent hierarchy"

For tobacco health SEO, keywords can be roughly divided into four tiers (I've maintained them in a table since 2022, updating search performance monthly):

My view: the easiest mistake in the early stage is stacking the whole site with L1. L1 traffic looks good but converts terribly. From Q4 2021 to Q1 2022, 60% of my site's articles were broad science pieces like "smoking and disease X," and organic search monthly UV once reached 18,000, but monthly WeChat adds were under 90, and CAC calculated against my time cost (at 80 yuan/hour of creation) was actually not low.

In May 2022 I did a structural surgery:

• Stopped updating half of the L1 topics;
• Forced 1 L2 self-check + 1 L3 action page at the end of every L1 article;
• Built 12 new L4 "downloadable/executable" pages (quit-smoking record forms, oral self-checks, home ventilation de-smoke steps).

By November 2022 the result: organic search monthly UV dropped to about 11,000 (-39%), but WeChat adds rose to 240+/month, and the material-pack conversion rate went from 0.4% to about 1.6%. Traffic dropped, and money began to look right. This was the first time I confirmed from numbers: in tobacco health, the KPI for SEO should be "effective-intent traffic," not UV.

2.2 Choosing topics within the compliance boundary: better slow than gambling on gray keywords

For e-cigarette/harm-reduction related pages, I stick to three hard rules:

• Titles and body never promise "harmless," "cures disease," or "replaces medical treatment";
• Comparison articles must clearly state exposure pathways and evidence limitations (heat-not-burn, vaping, and traditional cigarettes are not a simple "one is better than the other");
• Conversion hooks land on "quit-smoking support, oral care, behavior recording, home environment," not on "helping you keep smoking comfortably."

In 2023 a peer used "harm-reduction gadget reviews" to get tens of thousands of reads in two weeks, then three months later the account was throttled and external links were all dead. I did the math: that kind of play can have short-term CAC below 5 yuan, but content assets depreciate extremely fast, turning SEO into feed advertising. I'd rather spend 4–8 weeks raising one L2/L3 long-form article that can rank for 1–2 years.

2.3 Update cadence and "compound-interest writing"

My current on-site cadence (executable by one person):

• 1 L2 or L3 article per week (2000–3500 words, with checklists/timelines);
• Revise 1 old article every two weeks (add data, internal links, new consensus from 202x);
• Monthly internal-link audit: topics in the same problem cluster (e.g. "after quitting, oral") become a hub page + subpages, reducing pages competing for the same keyword.

In time cost, a long article that can rank takes about 6–10 hours from topic selection to publishing. If within 12 months it brings 80 WeChat adds, of which 12 people pay (mixed order value 199–699), the acquisition cost per article will be far below advertising. SEO's profit is not "payback in the same month" but "marginal cost collapse in months 4–12." A pure-online operator without more than 3 months of cash and patience will misjudge SEO as "useless"——actually the observation window was too short.

The industry side confirms this too: in tobacco-adjacent categories where ads are hard to buy, foreign-trade and independent-site players generally treat SEO as a long-term asset rather than a traffic switch; the logic transfers to Chinese-language health science as well——what you can control is content structure and update discipline, not the platform algorithm's mood.

3. Private domain: turning "read-and-leave" into countable user assets

3.1 Traffic hooks: I only use three kinds that are "high-intent, low-annoyance"

From public to private domain, I tried a dozen scripts and finally kept three (in primary use since 2023):

1. Self-assessment result interpretation: after the oral/dependence self-assessment, prompt "full scoring logic + personalized points of note in WeChat." The add-WeChat rate is stably 8%–15% in L2 articles (counted by end-of-article clicks).
2. Staged timeline PDF: e.g. "24 hours–12 months after quitting: oral and smell changes." It converts better than "add me and get materials" because the user is already in L3 intent.
3. 7-day record sheet + review voice message: suited for people who have already decided to act; within 48 hours after adding WeChat, the first usable form must be delivered, otherwise churn is extremely fast.

What I clearly don't like: raffles, pure-porridge check-in groups, and unfiltered "quit-smoking mutual-help groups of 10,000." In August 2022 I built a free group of 500 people; two weeks in, activity dropped to 3%, consultation conversion was nearly zero, and it consumed 40 minutes every evening. Free big groups are often negative assets in this vertical.

3.2 Tiering: private-domain profit comes from "who you shouldn't serve"

I divide private-domain users into four tiers (WeCom tags, strictly enforced since 2024):

My view: the key to stable profit is not turning all A into D, but stopping yourself from wasting time on A. In Q2 2023 I tallied: 60% of customer-service time was consumed by A-tier "casual questions," contributing less than 10% of revenue. After that the rule changed: A-tier only gets automated replies and articles; manual service only for B/C/D. Monthly labor hours dropped about 35%, while consultation volume barely fell.

In public data on medical-health private domains, a common narrative is that repurchase and first-order conversion beat pure public-domain advertising; in my case, what really opens the gap is the "manual leverage after tiering", not a few extra Moments posts.

3.3 Reach frequency and content supply

• Moments: 4–5 posts per week, roughly 2 experience breakdowns : 1 anonymous replay of a user question : 1 product/slot (hard ads no more than 20%).
• Community (only B/C small groups, max 80 people): 2 themed Q&A sessions per week, no idle chatter flooding.
• 1v1: C/D use a fixed intake form (years of smoking, daily cigarettes, oral symptoms, past quit attempts, goal), turning consultation from "chatting along" into "structured diagnosis."

From January to June 2024, my private-domain pool was around 3,200 people (including historical silent ones), with 40–55 active consultations per month, 18–28 closed orders, and monthly service income mostly fluctuating between 12,000 and 24,000 yuan. The fluctuation mainly comes from seasonality (quit intent rises after Chinese New Year, summer is slightly flat) and whether I steadily produce L3 content, not from a particular viral hit.

4. Conversion funnel: a set of numbers I use to manage profit

Below is a comparable funnel model (the numbers come from the magnitude range of the 2024 site + Zhihu + WeChat official account data, for reference, not a promise):

4.1 The most common bottlenecks at each layer and how I break them down

Bottleneck 1: high reading, low clicks

The reason is almost always: the body talks about harm, but the hook sells an "quit-smoking course now." The user is still at L1, and you're charging at L4.

Action: change the hook to "self-assess which stage you're at," and postpone the deal.

Bottleneck 2: high adds, low 7-day conversation

In October 2023 my add-WeChat welcome message was "Hello, do you need quit-smoking help?" and the activation rate was under 20%. After switching to a three-choice menu (oral problems / repeated quit attempts / just want materials), the activation rate reached about 38%. Choice costs matter more than enthusiasm.

Bottleneck 3: intent exists, deals low

Mostly a messy price anchor. I now fix three tiers: material pack → single diagnosis → periodic accompaniment, and publicly show the difference (deliverable list written in stone). After cutting "quote on case-by-case basis," the close rate went from about 30% to around 45%.

Bottleneck 4: deals exist, no repurchase

For tobacco health users, "success" is often graduation (they quit), so repurchase can't rely on "buying the same course again." I design repurchase as: home de-smoke solutions, oral care stage packs, 90-day relapse prevention, secondhand-smoke communication scripts for spouses——the same person at different life stages, rather than repeatedly charging IQ tax.

4.2 The accounting view of stable profit (four numbers I must compute every month)

1. CAC = (content creation time cost + tool subscriptions + minimal advertising) / new paying customers

I estimate my creation at 80–120 yuan/hour. In healthy months, CAC is mostly 60–150 yuan per paying customer.

1. First-order gross margin = order value − payment channel − delivery time cost

Material packs have high margins but don't scale; consulting margins depend on whether you use forms to cut time.

1. 90-day LTV = first order + repurchase + referrals (record if any)

I require LTV / CAC ≥ 3 in healthy months. Below 2, stop expanding and fix the funnel first.

1. Time capacity cap

Running solo, when C+D tiers serve more than 25 people at once, delivery quality collapses and bad reviews push CAC back up. Profit sometimes requires raising prices + cutting headcount, not acquiring more traffic.

5. Patching the three blocks into a "controllable profit" operating system

5.1 The growth order I endorse (90 days)

Days 1–30: only fix the supply structure

• Take stock of existing articles, mark L1–L4;
• Pick 1 problem cluster to go deep on (e.g. "quitting × oral" or "secondhand smoke × family"), complete the hub page + 4 sub-pages + 1 tool;
• Unify the add-WeChat hook and the welcome-menu copy.

Goal: add-WeChat activation rate >30%, don't think about scaling yet.

Days 31–60: only fix conversion

• Fix the three product tiers and delivery list;
• Review funnel numbers weekly (even hand-filled in a spreadsheet);
• Cut channel updates that don't produce B/C.

Goal: the paid conversion path is replicable, LTV/CAC trial calculation ≥2.5.

Days 61–90: slightly add supply

• 1 new L2/L3 per week + 2 old-article updates;
• Open one action group with ≤80 people;
• Settle high-frequency questions into sellable materials, reducing repeated 1v1 labor.

Goal: monthly paying-customer count is stable, not a single-month spike.

5.2 Hard rules after stepping on the pits

1. Don't apply feed-thinking to SEO. Viral hits are fine, but a hit must link back to your L2/L3 assets, otherwise it's noise.
2. Don't raise unfiltered big groups. The value of a private domain is relationship density and tags, not headcount screenshots.
3. Don't turn consultations into emotional chat. A 1v1 without an intake form makes CAC retaliate through your sleep costs.
4. Don't make "harm-reduction myths" that can't restate the evidence boundary. If the account dies once, SEO and private domain may clear together.
5. The profit-first metrics are LTV/CAC and delivery cap, not follower count. In 2024 I deliberately stopped two columns that read okay but produced no consultations; the books got cleaner.

5.3 When should you spend money on buying traffic?

Only when all of the following hold at once:

• Some L3 content converts organically and stably for 4 consecutive weeks;
• Private-domain delivery doesn't queue more than 5 days;
• Buying traffic is only used to verify "does the same hook hold on another platform," and there's a cap on weekly budget (my personal habit: single experiment ≤1500 yuan).

Otherwise buying traffic is just subsidizing funnel leaks. In the industry, low-cost acquisition talks about content, SEO, private domain, partnerships, and UGC as routine weapons——but in the tobacco health vertical, I define 80%+ of the budget as "content production and tools" rather than "exposure".

6. A one-page "profit control dashboard" you can copy directly (field-level)

I suggest you spend 20 minutes every Friday filling in these cells (Excel is enough):

• Number of new/updated articles this week (L2/L3/L4 separately)
• Hook clicks and WeChat adds brought by organic search
• 7-day activation rate after adding WeChat
• New-customer deals, order-value distribution (material/single/periodic)
• Manual hours, estimated CAC, rolling 90-day LTV
• This week's single improvement (write only one, e.g. "welcome message changed to three-choice")

Since I insisted on a weekly meeting (with myself) in 2023, the biggest change isn't a traffic explosion, but rarely being in a state of "busy for a month yet unable to say where the money came from and where the leak went". For a pure-online operator, this is closer to "stable profit" than gaining another 10,000 followers.

7. Wrapped up into executable judgments

For pure online tobacco health content, the essence of controlling acquisition cost is three things stacked together:

1. SEO is responsible for cheaply delivering "people with problems" to your door——use intent tiering and long-lifecycle pages, not piling up harm science.
2. The private domain is responsible for turning people at your door into serviceable, repurchasable assets——use tiering and structured delivery, not the hustle of a 10,000-person group.
3. The funnel is responsible for telling you whether to double down or stop——use CAC, activation rate, close rate, LTV, and your time cap, not feelings.

If you're currently in "have reads, no profit": don't expand platforms first. Pick one problem cluster, complete it in L2→L3→L4 order, change the welcome message, fix three product tiers, and watch the funnel for 8 consecutive weeks. The tobacco health vertical doesn't lack anxiety; it lacks people who can translate anxiety into executable steps and collect money within the compliance boundary. Those who keep acquisition cost pressed down will find profit turning from "occasional orders" into "systematic orders."