After I figured out why a website's RPM differs, the next question arose naturally: what should I write about to make advertising earn more? The answer "write about expensive niches" sounds simple, but in practice, it leads either to a correct content plan or to losing Google's trust. This article explains how niche economics work, how to assess a topic's potential independently, without made-up "RPM tables," and how to combine mass traffic with profitable articles without sacrificing expertise.
About the method. As in the previous article, I do not provide figures from my advertising account or publish "average RPMs by niche": they depend on the network, geography, season, and format, so any such table is misleading. Instead, below are the mechanics, relative levels, and tools you can use to check your niche with real data. My own observations are marked separately.
📚 Table of Contents
💎 Anatomy of a Profitable Niche: What Advertisers Pay Most For
An advertiser buys not an impression or a click, but a chance to get a customer. Therefore, the maximum bid they are willing to pay depends on two things: how much one customer will bring them and how likely it is that your reader will become that customer.
Customer LTV: Why Cloud Providers Fight for Every Customer
LTV (Lifetime Value) is the amount a customer brings to a company over the entire period of cooperation. Compare two cases:
- Mobile game: a player may make a few small purchases and uninstall the app within a week.
- Cloud provider: a team that deploys production on its infrastructure pays monthly for years. Moving to another cloud is expensive and risky, so a customer, once acquired, usually stays.
The cause and effect here are direct: the higher the LTV, the more an advertiser can spend to acquire one customer and still be profitable. If a customer brings thousands of dollars over several years, the company can afford expensive advertising, and it will outbid advertisers with cheaper products at auction.
Why B2B and DevTools Buy Advertising More Expensively Than Mass Market
In B2B, a high LTV is complemented by three additional factors:
- Narrow audience. There are thousands of times fewer people choosing a platform for deploying Java applications than those reading about movies. When the audience is rare, competition for each contact with it is higher.
- A person makes decisions for the company. A developer choosing a database or monitoring system is not spending their own money but the team's budget. One such reader can bring the advertiser a corporate contract.
- Long sales cycle. B2B solutions are rarely bought on the first contact, so advertisers pay to be present at every stage: from tutorials to tariff comparisons.
The mass market works the opposite way: the audience is huge, the check is small, and there are so many impressions available that there is no point for the advertiser to overpay for each one.
💡 In simple terms:
An expensive niche is not a "difficult topic" but a topic whose readers can become customers with a large and long-term check. Clouds, DevTools, and B2B services pay more because one of their customers is worth more, and there are few such readers.
📊 Niche Levels by RPM Potential
Below are relative levels, not dollars. The actual RPM of each level depends on the audience's geography, season, and advertising network: a high-level article with regional traffic in January might earn less than a mid-level article with traffic from the USA in December. Therefore, the table shows direction, not a guarantee.
| Level |
Niches |
Who Reads |
Who Advertises |
RPM Potential |
| 1 | Entertainment and general content | Broad audience, no purchase intent | Mass market, FMCG, games | Low |
| 2 | Tutorials and educational guides | Beginners and students | Online courses, educational platforms | Below average |
| 3 | Engineering tools: AI models, OpenRouter, OpenCode | Practicing developers | DevTools, SaaS, AI services | Medium |
| 4 | B2B, Cloud, DevOps, PaaS: deploying Java/Spring, Render, Railway, AWS | Developers and tech leads choosing infrastructure | Clouds, hosting, monitoring, databases | High |
| 5 | Fintech, acquiring, cybersecurity | Teams integrating payments or protecting production | Payment services, enterprise security | Very high |
Pay attention to two things. First: the boundary between levels is not determined by the "seriousness" of the topic but by the reader's intent. A tutorial "how to deploy Spring Boot on Railway" is closer to level 4 than level 2 because its reader is already choosing a platform. Second: levels 4-5 mean not only higher RPM but also higher expertise requirements (more on this in the section on choosing a niche).
🔍 How to Evaluate Niche Potential Independently
Instead of trusting others' RPM tables, you can check how much advertisers are willing to pay for your topic's audience using real Google data.
Step 1. Advertiser Bids in Keyword Planner
In Google Keyword Planner, each keyword has two metrics: "Top of page bid (low range)" and "Top of page bid (high range)". According to Google's definition, these are historical advertiser bids for appearing at the top of search results, approximately at the 20th and 80th percentiles. Access requires a Google Ads account with filled-in payment details, but running ads is not mandatory.
How to use it:
- List 10–20 keywords for the topic you are planning.
- Compare bids with keywords from another topic, for example, "render vs railway" with "history of paper money."
- Look at the ratio, not the absolute numbers.
Important disclaimer: these are bids for clicks in search ads, not your RPM from banners. They do not directly translate into website revenue. However, they honestly show how willing advertisers in that niche are to pay for an audience, and it is this willingness that is later reflected in banner ad rates.
Step 2. Demand in Search Console
A high bid is useless if no one searches for the topic. In Search Console, check which queries your site already appears for in search results: these are topics where Google already considers you relevant. A new article in a related, high-value topic has a better chance than an article in a completely new niche for the site.
Step 3. Ad Network Reports
The most accurate source is your own dashboard. After a few months of work, compare the RPM for individual pages: this will show you which topics are actually earning from your audience, not just on average for the market.
🧪 Case Studies from WebsCraft Practice
In my blog, I see two types of articles that earn differently. I don't provide numbers for the reasons explained at the beginning, but the difference in behavior is noticeable.
Case 1. Volume: Model Reviews and AI News
Articles about new AI models and tools attract the most visitors. Here are a few examples from the blog:
The reason for popularity is clear: the topic is fresh, and many people are searching for it. However, the reader of such an article is mostly looking for a free solution or is simply following the news, so the income per visitor here is moderate.
Role in the blog: traffic, new readers, brand awareness, and internal links to other articles.
Case 2. Yield: Deployment, Databases, and RAG Infrastructure
In-depth technical materials on hosting, deploying Java applications, databases, and RAG infrastructure are read by significantly fewer people. However, in my dashboard, such articles show noticeably higher income per visitor. I attribute this to the fact that they are read by developers who are already choosing paid infrastructure, although I haven't measured the reason separately: geography and reading duration may also contribute.
Role in the blog: income. A small number of such articles can provide a significant portion of earnings.
💡 Conclusion in simple terms:
"Magnet" articles attract people, and "cashier" articles earn money. A blog needs both types, and the main question is in what proportion.
⚖️ 70/30 Strategy: Volume vs Yield
A disclaimer upfront: this is my approach to content planning, which I developed while running my own blog, not a proven rule or industry standard. I do not claim to be the absolute truth. For your website, niche, and audience, the optimal proportion may be completely different, so consider it a starting point that is worth testing with your own data.
Here's how I distribute topics:
- 70% — Volume: topics with high demand and moderate RPM. AI news, model reviews, free tools. They attract mass traffic and new readers.
- 30% — Yield: topics with narrow traffic but high commercial value. Deployment, PaaS, CI/CD, databases, monitoring, payment integration. They generate income.
Why Not 100% Yield
It might seem logical to write only high-paying articles. However, niche topics have low search demand, and it's difficult for a new site to rank for them until it has a sufficient number of visitors and authority in the topic. In my opinion, Volume articles help precisely here: they attract readers, some of whom return, share materials, and link to them. Moreover, Volume articles are a natural place for internal links to Yield materials: a reader of a free model review might well click through to an article about deploying their application.
How to Adjust the Proportion for Yourself
After 2–3 months, review the proportion based on Search Console data and ad network reports. I focus on the following indicators:
- Traffic is falling, income is stable — it's worth adding Volume articles: there's a lack of new readers.
- Traffic is growing, income is stagnant — it's worth adding Yield articles and internal links to them.
- Yield articles are not ranking high — most likely, the site lacks authority in the topic; strengthen it with related Volume materials.
Perhaps for your project, a 50/50 or 80/20 split will work better, or perhaps you prefer a different planning principle altogether. The main thing is to make decisions based on your own numbers, not on others' proportions, including mine.
🤝 Affiliate Programs + Banners
Yield articles have another advantage: affiliate links fit naturally into them. A reader of a tutorial on deploying to a PaaS platform will still choose a platform, and a referral link to it is not intrusive advertising but part of the instructions.
Double Monetization of Articles on PaaS and Cloud
In an article about deployment, two income sources can work simultaneously:
- Network banners — earn on every impression, regardless of reader actions.
- Referral links — earn when a reader registers or pays. For example, Railway has a referral program that awards credits; check the service's website for similar programs on Render and their terms before publishing.
These sources do not compete: banners monetize all readers, while affiliate programs monetize those who actually use the advice.
Mandatory Rules
🧭 How to Choose a Niche Before Launching
Assessing Competition in Google Search Results
Before you start writing, enter your target query into Google and look at the first page. If it only contains official documentation and major media outlets, a new article will struggle. If small blogs and forums appear alongside them, the topic has a chance. The best opportunities are for queries where the results are outdated or superficial, and you can provide fresh practical experience.
Why You Shouldn't Chase Expensive Niches Without Expertise
The most common mistake after learning about RPM is to start writing about what's more expensive, not what you understand. Google has a direct answer to this. In its guidelines for helpful content, among the signs of content created "for search engines, not for people," are the questions: "Is the content primarily intended to attract visitors from search?" and "Do you mostly rehash others without adding value?".
For the most expensive niches, there's another problem. Finance, security, and law fall under "Your Money or Your Life" (YMYL) topics, for which, according to the same guidelines, Google places special emphasis on E-E-A-T: Experience, Expertise, Authoritativeness, and Trustworthiness. A technical blog that suddenly starts writing about taxes or legal issues competes with lawyers and financial publications without any advantage, while simultaneously diluting its own subject matter. I've analyzed typical mistakes that cause such sites to lose traffic separately: TOP 10 Mistakes of YMYL Sites and How to Fix Them.
The right approach is to find the intersection of an expensive niche with your own experience. A developer who has integrated a payment system into their project can write a detailed technical article about connecting acquiring in Spring Boot: the topic is at level 5, and their expertise is genuine. A developer without such experience, who writes "a review of the best payment systems," gets an expensive topic and an empty article.
💡 My advice:
I would choose not the most expensive niche, but the most expensive among those you actually work in. From my experience, the chain looks like this: expertise helps get search positions, positions bring traffic, and traffic in an expensive topic yields a high RPM. If the first link is missing, the rest usually doesn't work. But this is my perspective, not a universal rule: test it on your own site.
❓ Frequently Asked Questions (FAQ)
Which niche gives the highest RPM from advertising?
Usually, those where the advertiser's client has a large and long-term contract: fintech, cybersecurity, B2B services, cloud, and DevOps. However, the specific RPM depends on the audience's geography, season, and advertising network, so there's no universal figure.
Can you find out the RPM of a niche before launching a website?
Precisely, no. But you can estimate the relative potential: compare advertiser bids in Google Keyword Planner for key queries of different topics. These are bids for search advertising, not your RPM, but they show well how much advertisers value the audience of a topic.
Can affiliate links be combined with ads?
Yes, and it works well in technical articles about infrastructure. Be sure to disclose affiliate links to readers, mark them with the `rel="sponsored"` attribute, and check the rules of your advertising network.
Is it worth a technical blog writing about finance or law for the sake of RPM?
Only if you have real expertise. These are YMYL topics where Google pays special attention to the author's experience and authority. It's better to write about the technical side: for example, how to integrate payments, rather than how to choose a bank.
✅ Conclusions
- 🔹 An expensive niche is determined not by the complexity of the topic, but by the advertiser's client LTV and the reader's intent: cloud, DevTools, B2B, and fintech pay more because their client is more valuable.
- 🔹 "RPM by niche" tables in dollars are misleading: actual income depends on geography, season, and network. Compare relative levels.
- 🔹 The potential of a topic can be checked independently through bids in Keyword Planner, demand in Search Console, and reports from your own advertising network.
- 🔹 A blog needs both Volume articles for traffic and Yield articles for income; a 70/30 proportion is a starting point, not a rule.
- 🔹 The best niche is the intersection of an expensive topic and your real experience. Without expertise, an expensive topic will yield neither positions nor income.
📖 Read Also
📎 Sources
Material is current as of October 2026. Advertising network and affiliate program terms change, so check them before making decisions.
