Writing well is a superpower. Great writers are great thinkers, capable of using the written word to impose structure on nebulous ideas and simplify the complex. They're persuasive, able to create ironclad arguments to defend their beliefs. They're interesting, masters in the art of winning attention. They can reach across oceans and continents to create real, personal connections with their readers.
And good writing is even more rare in the era of AI slop.
The volume of content being produced is incalculable, and most of it is — I don’t think there’s any other way to say this — bad. If you want people to read your content, so you can convince those same people to buy your product or service, you need to know how to make people want to read it.
We published an earlier version of this guide in 2021 and have updated it to help you create content that stands out in the AI-generated flood zone.
Step 1: Develop Your Idea
The first step is to identify what you’re writing about, and “develop” is the operative word. You can't manufacture a good idea the moment you sit down to write. They need time to grow — something AI can't give you. This is the most important step because great writing can't cover for bad ideas.
Build an idea farm — a single document that collects every nascent thought, opinion, or idea you have during the day. It could be a Google doc, a Trello board, or any database that stores raw "seeds." These are thoughts from sales conversations, unexpected customer feedback, a colleague's point in Slack, or a quote from a book you're reading.
At first, these seeds won't look impressive. They're nowhere near ready to become blog posts. But as your doc fills up, you'll spot commonalities or unexpected angles that grow into ideas worth harvesting.
Good ideas need depth and perspective. You might find that perspective from academic methods like keyword research and audience analysis, or simply by getting input from a peer. Colleagues help identify your most compelling ideas, ask questions to further your thinking, and add their unique experience and perspective into the mix.
When you're ready to validate an idea, make sure it passes these tests:
Impact: Will this move the needle for your business?
Originality: Are you adding something new to the conversation?
Credibility: Do you have the evidence and expertise to back up your claims?
Timeliness: Is now the right time to publish this?
If you can answer yes to all four, you have an idea worth writing about.
Step 2: Find Your Angle
The next step is to consider how you frame the idea. A good topic becomes a great post when you frame it with the right angle.
Anyone can ask ChatGPT, "What is content marketing?" and get a competent answer. The angle — your contrarian take, your specific experience, your unique framing — is what makes someone choose your article over an AI-generated blurb.
Unique angles matter for SEO, too. When every top-ranking article says roughly the same thing, you stand out by adding something new to the conversation. Google calls this information gain, and it's how you compete when AI can summarize the basics instantly.
There's a simple exercise to turn your validated ideas into unique angles:
List 20–30 truisms about your industry ("retention is king," "hire for culture," "build things people want").
Grab a partner to bounce ideas off of, and record the session. ChatGPT will work as a partner, too.
For each truism, ask: Do I have reason to believe the opposite, or can I imagine edge cases where this doesn't hold?
Skip liberally — you won't have much to say for most items, and that's fine.
The one or two truisms where you find yourself passionate about reversing convention are your angles.
Another way to bring a unique angle: Interview subject matter experts. Even great writers don't have deep expertise in every topic they cover. Experts add credibility and fresh insights that transform generic advice into specific, actionable guidance.
When you interview experts, ask "dumb" questions that your readers would ask. Let the conversation flow naturally beyond your prepared questions. Focus on their lived experience — that's the part AI can't replicate.
Step 3: Structure Your Outline
Writing a blog post outline is like planning a road trip. You don't pick stops before you choose your destination. Same with outlines — you can't build one before you figure out where it's leading.
That destination is your thesis: a one-sentence expression of the argument you intend to make.
Once you have your thesis, define all the key points that support it. These become your article headings, structured logically. As you put them in a doc, you'll see gaps in your thinking — rearrange and rewrite them until the logic holds.
Great writers are great thinkers.
You're about 10% into your article now. Next, add supporting evidence to each point: explanations, quotes, stats, and examples that support your argument. This takes you to 30%.
Almost done with the outline. You just need to verify it against the MECE framework.
Each section should cover distinct ground without overlap — that's mutually exclusive. Together, they should cover your topic without obvious gaps — that's collectively exhaustive.
For example, if you're writing about building a content pipeline, don't structure it around your research process ("why pipelines matter," "what makes a good pipeline," "how to build one"). Instead, use a logical framework like process stages: planning content, producing content, and analyzing results.
Step 4: Write Your First Draft
You can write your introduction before or after you draft the main body — either way, you'll refine it later. The chances you landed on a catchy hook and a sleek thesis on your first pass are low.
Your introduction should grab readers with something unexpected, clearly state your thesis in one or two sentences, and show concrete benefits so readers know what they'll gain.
As you draft the body, use BLUF (Bottom Line Up Front) in every paragraph and section. Lead with your most important point, then add context and supporting details. Readers get your insight immediately instead of hunting for it. When answer engines pull excerpts from your content, they look for the insights right at the top of each section.
Don't just present information; lead readers through the reasoning process so they arrive at your conclusion themselves. Use the thesis-antithesis-synthesis structure: present the status quo (thesis*), articulate the problems with it (antithesis), then share your better solution (synthesis).
Layer multiple types of evidence throughout: tackle the strongest counterargument head-on, tell stories that illustrate your point, use data to support claims, show social proof from experts or successful companies, and return to your core idea through different metaphors and angles. When you combine different types of arguments, you speak to different reader preferences.
AI can help you draft faster, but never let it fill your blank page. Start with your own thinking — an outline, a messy paragraph, a stream of thoughts. Use AI to get unstuck, generate variations, or connect disjointed sections, not to do your thinking for you.
Step 5: Revise Your Draft
First drafts are for you. Published drafts are for your reader.
That initial draft is a thinking tool. It challenges you to articulate ideas coherently, highlights gaps in your logic, and develops your understanding. But to communicate with your writing, you need to make it reader-centric.
Remove your "training wheels." These are the scaffolding you needed to write but readers don't need to read: the long-winded introductions that contextualize the topic for you, paragraphs that save the insight for the end, transitory information that doesn't serve the narrative.
Flip your structure. Sequential writing mirrors your thought process: "Here's the problem, here's background, here's research, here's my solution." Reverse it. BLUF your paragraphs if you haven't already.
Add shibboleths. The subtle hallmarks of experience that prove you're part of the in-group. Use idiosyncratic language correctly (developers say "JS," not "JavaScript"), weigh ideas appropriately (don't treat basic concepts as revelations when your audience already knows them), and share stories that sound authentically like a day in the life of your reader. This is where AI-generated content consistently falls short.
Use AI to clean up mechanical issues, like repetitive words, passive voice, and readability problems. But for everything else — understanding whether your argument is persuasive, whether your examples ring true, whether your logic holds — you need human editorial judgment. AI will tell you what you want to hear. A good editor (or your own critical eye) will tell you what you need to hear.
Don't waste your conclusion. Readers disproportionately remember how your article ends, so skip the rote recap. Instead, inspire them with what they can do next, reframe your main point through a fresh metaphor, or share a bonus insight that didn't fit elsewhere.
When you're done revising, your draft should read like you knew what you wanted to say from the start. Even if you didn't.
Step 6: Write Your Title
We say save the title for last. Great articles can be undone by weak titles, least of all because no one clicks into an uninteresting title, so it’ll never get read.
Like a good idea, a strong title needs time to bake. Certain phrases or angles for the hook will emerge as you write. Let them simmer before committing to one.
When you sit down to finalize your title, brainstorm multiple options using different approaches:
Help readers get promoted: Highlight the concrete benefit or result your article delivers ("How to Cut Your Sprint Planning Time in Half").
Reverse expectations: Take a truism and flip it ("You're a Content Marketer, Not a Writer").
Show your work: Own the effort behind your analysis ("Lessons from Analyzing 50 Welcome Emails").
Mine specifics for universal truths: Find the interesting detail buried in your piece and make it the star ("Why Your Facebook Ads Shouldn't Be Blue").
Your title serves two audiences: humans scrolling through search results or social feeds. Make it clear and benefit-driven for both.
Step 7: Polish and Publish
Before you hit publish, zoom out and check the overall visual of your piece. It should be skimmable and easy to digest.
Most readers won't read every word, so help them extract value quickly: Use descriptive H2s and H3s, break up text walls with bullet points or pull quotes, bold key statistics, and add a table of contents for longer posts.
AI has exposed the cheapness of content that was always cheap.
Check that your structure holds up. Your headers should ladder up to your title. Each section should feel like a natural part of answering the promise your title makes. Include your target keyword naturally in title and intro, add descriptive alt text to images, and link to relevant internal content.
Finally, read your draft out loud, even and especially if AI had a hand in it. Run it through Hemingway, Grammarly, or Claude to spot issues. Get a second pair of eyes for developmental edits, and the finishing touch: Proofread carefully. We run everything through our copyeditors for a reason — they catch what we miss.
Writing That Gets Read
It's true that AI has made the process of writing content feel cheap. It can write 1,000 words in 60 seconds, and we stand in awe as if word count is where the value is.
But it can't create what was always the differentiator for good content: an original, well-developed idea, compellingly expressed.
AI has exposed the cheapness of content that was always cheap — copycat content. After all, both human and machine can aggregate what already exists into a passable facsimile. One just does it faster. But that's not what makes good content.
Good content sticks. It changes how people think, gets saved for later, gets shared with colleagues. It comes from the thinking — spending weeks wrestling with an idea until it's ready, watching it shapeshift, interviewing experts and asking the dumb questions that lead somewhere interesting, layering evidence so readers feel like they arrived at your conclusion themselves.
We say: Don't worry about AI. It can accelerate your process, but it's only as original as the thinking you bring to it. Try something different: making something worth reading.
*Georg Hegel was a 19th-century German philosopher who developed the thesis-antithesis-synthesis framework. In this framework, "thesis" refers to the conventional wisdom (often what you're arguing against), while "synthesis" is your actual argument. Yes, this is confusing since we know "thesis" to mean "your main point."