A draft can be clear, well organized, and still not give a skeptical buyer enough reason to act. If you are a marketing lead, you have probably read a page like that and known something was missing without being able to say what. This guide walks you through the psychology of persuasive content as a nine-step revision you can run on one page at a time. If you've been wondering how to write content that converts, this is the order that tends to work. You'll go from the decision your reader is trying to make, to an argument that supports it, to a next step they can actually take.
The short answer to "what persuasion techniques work in written content" is a set of persuasive writing techniques that fit together. You make a relevant claim you can defend, back it with reasons and concrete evidence, show why the reader should trust you, connect the consequences to what they care about, use social proof only when it's real, answer the most serious objection, and ask for one specific, doable action. Classical rhetoric groups the first three appeals as ethos, pathos, and logos, meaning credibility, emotion, and argument. They are ways to build an argument, and none of them is a guaranteed conversion trick, so you'll want to check whether each one helps your particular reader.
This guide covers argument and persuasion. It doesn't cover how to structure a story or how to format a page for easy reading, since those are separate topics.
How the pieces fit together
Rhetoric in content marketing is easier to use when you treat the three classical appeals as a checklist for your draft. Ethos is about the communicator: who is speaking, and why should the reader believe them. Pathos is about the reader's emotional state, meaning the consequence they care about. Logos is the argument itself, where a specific claim is tied to evidence that supports it. Aristotle drew these distinctions in his Rhetoric, and they still work as a way to look at a piece of copy and see what it's missing.
You don't need all three in every paragraph. Think of ethos pathos logos content as a way to diagnose the whole page. A page about a large purchase may need a lot of evidence even if the opening only briefly names the stakes.
Two other ideas will help along the way. The first is the elaboration likelihood model, which Petty and Cacioppo published in 1986. It says that when people think carefully about a message, the quality of the argument matters most, and when they aren't thinking hard, surface cues like popularity or a familiar badge matter more. For a consequential buying decision, that's a good reason to give the reader reasons they can inspect. It isn't a reason to sort your visitors into "logical" and "emotional" types, because the same person can be either depending on the day.
The second idea is BJ Fogg's behavior model, B=MAP, which says a behavior happens when motivation, ability, and a prompt come together at the same moment. You can use it later as a check on your call to action.
Before you start, it helps to write five lines on a scratch page.
- For this audience and decision...
- The claim is...
- The strongest evidence we have is...
- The most serious fair objection is...
- The right next action is...
This worksheet is an editing tool we're suggesting, not a validated scientific instrument. It gives you something to test each step against.
1. Define the reader's decision and the action you want
Choose one reader, one decision, and one realistic action for the page. Write down what the reader already believes, what evidence they don't have yet, what could stop them, and how big the decision is for them. For a marketing lead, a real decision might be whether a content production system can keep product facts and brand voice accurate while taking some manual review off their plate. The next action in that case could be looking at a workflow, and not buying right away.
How to tell you're done. Your team can fill in this sentence: "A marketing lead deciding ___ needs to believe ___, verify ___, and then do ___." Also write down the one assumption you most need to check against real customer language. If you haven't done that research yet, audience research before you write is a good place to start. A B2B buyer persona template can help you keep the reader specific.
Where people go wrong. Writing for "everyone in marketing" is the usual one. Another is asking for a trial before the reader has enough information to want one. A third is treating the fears and buying triggers in a persona document as measured statistics about a whole population. Those figures are worth investigating for your own organization, but they aren't general benchmarks, so don't publish them as research findings.
If you use DeepSmith, Deep IQ stores your Buyer Persona, Products & Services, and Brand Voice context, and that context shapes each draft. You still have to decide which objection and which action matter on this particular page. AI Visibility can show you the tracked questions and answers around that decision, though its mention and citation metrics don't measure a reader's motivation or your conversions.
2. State one defensible claim and connect it to a buyer consequence
Write a claim with a clear subject, an action, and a benefit you can stand behind. Then put the reasoning right after it, so the reader sees why the capability affects their decision. It helps to separate what something can do from what it has been measured to do. "A workflow that carries approved product context into each article reduces the need to re-brief writers on the same facts" describes a mechanism, and that's a fair thing to say. "It cuts editing time in half" is a measured result, and you'd need evidence for it.
How to tell you're done. Another editor can underline your claim, point to the reason that supports it, and say exactly which part would need evidence. The claim speaks to a decision, and it isn't just a nice adjective like "effortless."
Where people go wrong. Claiming a business result because a feature exists is common, and so is stacking several unrelated promises in one paragraph. If your organization hasn't measured a time saving, don't put one in.
Common mistake. A confident claim with no reason next to it reads as an assertion, and skeptical readers tend to treat assertions as noise. If you can't say why the claim is true in one sentence, the reason is what's missing, and more emphasis won't fix that.
3. Build the logos: support the claim with reasons and evidence
Match each important claim to the strongest evidence you really have. That might be a documented product behavior, a demonstration you describe accurately, a study whose population and outcome fit what you're saying, or a customer result you have permission to share along with its context. Then say out loud how the evidence leads to the conclusion, because readers won't always draw that line themselves. When you're unsure how much detail to give, a specific, checkable description usually beats vague praise. Describing which product-context fields shape a draft says more than saying an AI system "knows your brand."
How to tell you're done. A skeptical reviewer can trace each important statement back to its evidence, can tell what was observed from what was inferred, and can see the limits that apply. If a number is missing, it stays missing. Pages that add original insight tend to have an easier time here, because they have something of their own to point to. The same goes for pages built on original research, since the evidence is yours to describe.
Where people go wrong. An invented statistic that looks precise is the worst version. Citing a study about a different behavior as proof of a conversion lift is another, and so is presenting a hypothetical example as if a customer had said it. Even the advice to "be concrete" has limits. Researchers who studied concrete versus abstract advertising found that concrete descriptions were more appealing under uncertainty in the settings they tested, across four studies, three in a lab and one in the field. That's a useful finding, and it doesn't say concrete language beats abstract language everywhere.
If a draft came from an AI tool, this is the step where a fact-check pass on the draft pays off.
4. Establish ethos with credibility the reader can verify
For each claim that matters, name the real source of authority. It might be a subject-matter review, product documentation, a method you explain openly, or an endorser who actually used the thing. If a limitation affects the reader's decision, say so. For a product passage, ask the owner to check the exact capability, the plans it's available on, and the wording of any performance claim. Credibility comes from the match between what you assert and what you can show, and forceful wording doesn't add to it.
How to tell you're done. The reader can tell who stands behind an important claim and what they can check for themselves. Endorsements come from real experience, and any connection that could change how credible the endorser looks is disclosed.
Where people go wrong. Inventing a customer quote, implying an endorser used a product when they didn't, hiding a material relationship, or using a testimonial to suggest a result the company couldn't back up on its own. The FTC's guides on endorsements and testimonials are worth reading for the factual context they set out around advertising claims. They don't promise that adding testimonials will raise sales. If your page needs to show expertise, the ideas in E-E-A-T for AI search carry over well. Using expert reviewers with real credentials is one practical way to do it.
Deep IQ holds your approved positioning, the claims you want to make and avoid, your product profiles, and your brand voice. Content Studio's Writer uses that context to produce an article, and Produced Content lets a person review and edit it before it goes live. That's a way to carry approved context through production, and it doesn't prove that every sentence a tool produces is accurate or persuasive, so the review is still your job. If you want to see how that grounding works, there's more on grounding content in brand voice and product facts.
5. Make the stakes felt without manufacturing fear
Turn the claim into a consequence your reader recognizes. A marketing lead might care that the same factual corrections keep eating review time, or that inaccurate content could wear down trust. State the concern in proportion, then explain the believable path from your proposed solution to a better decision. Pair the emotional point with a reason and some evidence, since pathos works best next to logos and doesn't replace it. In ethos pathos logos content, the emotional appeal is the piece most likely to be overdone, so it deserves a careful second look.
How to tell you're done. The reader can say why this matters for their actual responsibilities, and a reviewer can't find an invented crisis, a countdown, or a guaranteed outcome anywhere in the copy.
Where people go wrong. Assuming loss-framed wording always beats benefit-framed wording is one trap. Borrowing an extreme story and presenting it as typical is another, and so is stretching fear past what the facts support. "This is worth checking before you scale production" is honest advice. "You will lose every AI citation if you wait" is not something we have support for.
Pro tip. Read your emotional paragraph out loud and ask whether you'd say it that way to the reader's face. If it sounds like a sales pitch when spoken, tone the stakes down to what you can actually back up.
6. Use influence cues only when they carry real information
Most persuasive writing techniques that people borrow from psychology work best when they are treated as information for the reader. Choose a cue, if you need one, that helps this reader judge the claim. Robert Cialdini's framework lists social proof, authority, reciprocity, liking, scarcity, commitment and consistency, and unity. In a business article, social proof could be a customer account you have permission to use, with its limits stated. Authority could be a qualified expert explaining the method. Reciprocity could be a useful decision checklist offered before any ask. Shared identity and past commitments have to be genuine to count. Mention scarcity only for a real constraint, with accurate terms.
How to tell you're done. If you took the cue out, the reader would lose some useful, true information, and would not just lose a little pressure. A testimonial's audience, circumstances, and result can't be mistaken for what everyone will get.
Where people go wrong. Made-up testimonials, popularity claims that don't apply to this reader, borrowed authority, and artificial deadlines are the usual ones. Treating Cialdini's seven principles as seven guaranteed ways to lift conversions is another. You don't have to use all of them. Pick the two or three that fit this buyer's decision and leave the rest.
Common mistake. "Add social proof" doesn't mean drop in an anonymous quote. Ask what the person actually experienced, whether their account matches the claim you're making, and whether a connection to your company needs to be disclosed.
7. Answer the strongest fair objection
Good rhetoric in content marketing usually includes a straight answer to the doubt a reader already has. Say the objection a thoughtful reader would raise, then give a proportionate answer or an honest boundary. For a marketing lead, a fair one is "Will more output mean generic copy or wrong product claims?" A believable answer describes the approved brand and product context, the point where a person reviews the work, and what still needs human judgment. If you don't have the evidence to settle the concern, say that, and don't claim it's solved.
How to tell you're done. The draft names the real objection in words the reader would recognize, and gives them a reason to update their view without pretending the remaining risk is gone.
Where people go wrong. Listing objections and never answering them is one way. Picking a trivial objection that's easy to knock down is another. A third is announcing that two-sided copy has been proven to beat one-sided advertising. A meta-analysis of one-sided and two-sided messages, published in 1999, found benefits for refutational two-sided messages on nonadvertising topics. In its advertising comparisons, those messages didn't show a significant advantage in credibility or persuasiveness over one-sided ones, and the author noted there were few relevant studies. So the reason to answer a real objection here is to make this particular argument more complete and more honest, and not to promise a lift.
8. Ask for a next step the reader can actually take
Fit the action to how ready the reader is. Name the action and say what it's for. "Review the workflow against your approved claims" tells the reader more than "Get started." If a trial is the right next step, say what they can evaluate during it and make the conditions of the offer clear. Then use Fogg's model as a check. Does the reader have a reason to act, a practical way to do it, and a prompt they can't miss?
If the reader can't yet tell whether your product suits them, a lower-commitment action is usually the better fit, and pushing harder on the appeal won't change that.
How to tell you're done. A reader can answer "What happens next, and why should I do it now?" and the answer doesn't lean on a made-up deadline. The action matches the evidence you've given and the level of commitment you're asking for.
Where people go wrong. Repeating the call to action with more urgency when the real problem is an unanswered objection, thin proof, or an action that takes more effort than the reader is ready to spend. If a page isn't converting, look at steps 2 through 7 before you change the button.
9. Check the argument and measure the intended action
Before you publish, run a short audit. List every claim, its evidence, any qualifications, the strongest objection you haven't answered, and the action you're asking for. Have a product owner review product claims, and have an editor decide whether the emotional appeal is fair.
If you want to know whether a change to the copy helped, pick one main behavior that fits the page, such as qualified trial starts or completed demo requests. Set a baseline, write down the variant and the audience, and estimate the sample size and how long the test needs to run before you start. Look at the quality of what comes after the click too, and not just the click. A controlled A/B comparison is better than declaring a win from a story or from a before-and-after trend that hasn't been adjusted for anything else. Optimizely's help center explains how long to run an experiment and points to a sample-size calculator for estimating the traffic you'd need. If you're just getting started with testing, a 90-day content experiment is a manageable way to build the habit.
How to tell you're done. You have an approved argument you can trace, and you've recorded how you'll define the outcome before you see any results. If your traffic is too low for a useful test, write that down and collect the reader's objections directly, and don't call a winner without the numbers to support it.
Where people go wrong. Treating pageviews or AI citations as if they were conversions is a common one, and AI visibility metrics measure something different from a reader deciding to act. Changing several claims and the offer at once and then crediting one sentence is another. So is stopping a test early because the first numbers look good. How long a test takes and how much traffic it needs depend on your metric and your design, and there's no universal threshold for sample size or lift.
A quick before and after
Here's a small example of the steps working together. It's an editorial illustration, and not a quote, a measured result, or approved product copy.
Before: "Our powerful AI transforms your content and guarantees growth. Start now." The outcome isn't supported, there's no mechanism, nothing speaks to a particular buyer, and the guarantee can't be justified.
After, the claim and logos: "DeepSmith's Writer uses the product and brand context stored in Deep IQ while producing an article. Review the resulting claims against your approved product information before publishing." This describes a real capability and keeps the checking step in, and it doesn't assert any growth.
For ethos, you'd point to the product owner or the documentation behind an important capability, and you would not make up a customer testimonial. For pathos, you could say: "If your team keeps correcting the same product claims, keeping approved context in the production workflow addresses a real review burden." The "if" matters, because that burden hasn't been measured for every reader. For the objection, you might write: "Will the output still need editorial judgment? Yes. Review the strategic argument, the product claims, and whether the requested action suits the reader before publishing." And for the prompt, offer a specific way to evaluate once the reader has seen enough evidence, without promising a time saving you can't back up.
What to do next
Pick one existing page that matters to you and fill in the five-line worksheet from the start of this guide. Then check the strongest claim on that page against its evidence, and decide how you'll measure the action you want. That's a good afternoon of work for anyone learning how to write content that converts, and it's a much better use of time than rewriting ten pages.
When you're ready to see whether a workflow that carries your approved context through production fits your team, you can start a 7-day free trial of DeepSmith and judge the drafts against your own claims and your own review process.
