Creating useful, reliable, people-centered content takes 5 steps:
- Write in a simple, natural way
- Don’t use too many keywords
- Update old posts with better information
- Add questions and answers that people are searching for
- Check current rankings and learn from them
Write in a simple, natural way
Global daily digital content increases by 3.5 billion GB (IDC, 2023), while adult attention span averages only 8 seconds (Microsoft, 2015).
Data from Chartbeat, a US content platform, shows: articles with more than 30% professional terminology have a 42% lower completion rate than conversational content;
When sentence length exceeds 25 words, comprehension accuracy drops by 28% (Nielsen Norman Group, 2021).
When explaining “user journey map,” replacing it with “visual guide showing the complete process from first contact to purchase” improves comprehension by 37%.
Simple and natural
Every extra word consumes energy
Neuroscience research shows that working memory (short-term memory) can only hold 4±1 information chunks at a time (Baddeley, 2000).
Chartbeat’s tracking of 3000 English articles found: articles containing more than 10 professional terms have readers’ time-on-screen in the first viewport stay time 63% shorter than similar conversational content.
For example, a tech blog about “search engine algorithms” originally used terms like “crawl frequency” and “PageRank value iteration,” with readers averaging out at paragraph 3;
After revision using “how often does the search engine send its ‘information collectors’?” and “how many times a link is recommended by other websites,” time-on-screen extended by 41%.
University of Pittsburgh research on first-year students showed: when explaining “diminishing marginal utility” with a life example like “eating the first piece of cake is satisfying, but the fifth piece might make you feel stuffed,” comprehension accuracy reached 78%;
When using the academic definition “as consumption increases, the marginal utility of each additional unit of goods decreases,” accuracy was only 39%.
Not all readers understand your “jargon”
Many content creators believe “using terminology appears professional,” but data proves this is a misconception.
Nielsen Norman Group’s analysis of 120 pieces of content across finance, healthcare, and tech found: when targeting general audiences, every 10% increase in professional terminology decreases content sharing probability by 18%.
Taking “user interface (UI) design” as an example, in a guide targeting entrepreneurs:
Original sentence: “Optimize information hierarchy through F-pattern browsing, combined with visual weight to guide user attention.”
Revised: “Put the most important information in the upper left corner of the page—when people look at screens, their eyes scan from left to down first, forming an F shape.”
The latter replaced terminology with specific descriptions like “upper left corner” and “eyes scan left,” target readers’ (entrepreneurs’) bookmark rate increased from 12% to 34% (HubSpot, 2022).
Text is simple enough
Brain imaging research at UC San Diego shows: when reading simple sentences, prefrontal cortex activity (responsible for deep thinking) decreases by 27%, while temporal lobe activity (responsible for language comprehension) increases by 19%.
In the Content Marketing Institute’s case library, a B2B company‘s product page did an A/B test:
Version A: “Our solution achieves cross-system real-time data synchronization through API integration and microservices architecture.”
Version B: “Your system can automatically exchange data with other software—for example, when a customer places an order, the inventory system immediately knows it needs to restock.”
After launching Version B, inquiry volume increased by 58%, and trial conversion rate rose from 3% to 7%.
“Human language” that reads naturally
Natural content reads like chatting with a friend, people are more willing to accept information from “similar individuals”.
When text uses pronouns like “you,” “I,” “we,” or includes personal experiences like “I encountered this before too,” readers feel “this is someone like me sharing” rather than “an expert lecturing.”
Original draft from a health blog: “Research shows that walking 8,000 steps daily can reduce cardiovascular disease risk by 20%.”
Revised: “My dad started walking every day after retirement—initially tired after 5,000 steps, now he consistently does 8,000 steps for half a year. His last physical showed his blood pressure was exceptionally well controlled—the doctor said it was related to his increased daily steps.”
The latter’s reader engagement rate (comments + likes) was 2.3 times that of the former (Moz, 2023).
Simplicity doesn’t lose content depth
Research published in Science in 2021 found: breaking complex content into “core conclusions + supporting details” structure actually improves readers’ understanding of depth.
For example, when explaining “climate change”:
Core conclusion: “Global warming is mainly caused by human activities emitting greenhouse gases.”
Supporting details: “Over the past 100 years, atmospheric CO2 concentration has risen from 280ppm to 420ppm (NOAA data), equivalent to wrapping Earth in a thicker blanket.”
This structure lets readers grasp the key point first, then explore details as needed, agreement with ‘human activities are the main cause’ increased from 61% to 82% (Yale Program on Climate Change Communication, 2022).
Readers are more accustomed to encountering problems → finding answers
How they actually think
Harvard University cognitive science lab experiments show: when readers open an article, 78% already have an unresolved problem in mind (like “how to make my home tidier” or “why projects always get delayed”).
At this moment, directly giving them “5 principles for tidiness” or “10 major reasons for project delays” is like telling someone asking for directions “there are 32 traffic rules”—the information is correct, but unusable.
A home goods platform ran an A/B test on the same article about “small space storage”:
Version A used a “knowledge checklist” structure: listing “wall hooks, tiered storage racks, drawer dividers” in points;
Version B used a “thinking flow” structure: starting with “My living room coffee table is always covered in takeout boxes, and I can never find keys in the sofa cushions—are you the same way?” then introducing specific methods.
Version B’s completion rate was 49% higher than A, with 3 times more bookmarks (Platform Analytics, 2023).
Progress like a conversation
What does a real conversation look like? When you ask “what should I do about constant insomnia,” a friend won’t immediately dump “7 methods for treating insomnia” on you—they’ll first respond “Are you under too much pressure?” “Do you always scroll your phone before bed?” then give advice based on your response.
Khan Academy, the educational content website, when teaching “fraction division,” instead of “3 steps for fraction division,” approaches it from a specific problem:
“Want to know why 1/2 divided by 1/4 equals 2? First think—if you have half a pizza and want to share it at 1/4 piece per person, how many people can it serve?”
This structure starting from specific problems increased the proportion of students actively thinking by 62% (Khan Academy internal data).
How to specifically design it
Scene descriptions with details
Avoid vague “many people have experienced this,” use specific scenes so readers feel “this is talking about me.”
For example, when teaching “efficient early rising,” don’t write “waking up early is hard,” write “Alarm rang 5 times, turned it off by touching phone and went back to sleep;
Opened eyes at 8:30 AM, no time for breakfast, ran all the way to work—I used to be like this too.”
Make readers feel, this is the trouble I’m encountering
Readers often can’t articulate specific problems—about “workplace communication,” you could write: “Do people always say to you ‘what exactly do you want?’
You want to express a need, but once you open your mouth it becomes a complaint; you make a suggestion, but the other person can’t hear it.”
Give a direction that “might be useful”
Don’t rush to give a complete solution, first provide a minimum action suggestion.
For example, about “improving marital relationships,” you could say: “I tried one method: chat with my husband for 10 minutes before bed about ‘any small happy thing today’
Not about kids or bills, just share joy together. After one week, we had fewer arguments.”
Show I made mistakes too, then adjusted
Adding “failure-to-improvement” details makes content more credible. For example, when writing about “diet recipes,” you could write: I initially strictly followed an influencer diet, lost 3 pounds in two weeks, but by the third week I binged—I later realized that completely cutting carbs actually makes people crave more.
Now I leave 100 calories for a small piece of bread, and I can stick with it much longer.
Natural content often carries “human flavor”
Real imperfection is more convincing than perfect falsehood
Cornell University social psychology experiments show: when someone claims they “never failed,” listeners trust their statements 28% less than those who admit “tried 3 times before succeeding”.
The same applies in content—overly packaged “perfect recommendations” and “zero-defect solutions” actually make readers feel “this is a marketing tactic.”
A baby product review website once published an “annual best baby strollers” article, originally emphasizing “100% safe,” “all moms love it,” and “no drawbacks whatsoever.”
The comments section was full of questions: “Did it really never leak milk? There’s always dirt getting stuck in the seams at my stroller’s joints” “Did anyone consider the weight? An 18-pound stroller can’t be carried one-handed while holding the baby.”
Later added authentic details: We tested 12 strollers—this one folds most conveniently (3 seconds to collapse), but weighs 16 pounds
—me being a mom with less strength, I need to rest twice when carrying it upstairs alone; the sunshade is big enough, but rain seeps in at the edges.
After revision, reader comments changed from “questioning” to “practical reference,” and bookmarks increased 4 times (platform backend data).
Why readers dislike recommendations with no drawbacks
Nielsen Norman Group’s survey of 5,000 readers found: 72% of people actively look for “drawbacks” or “limitations” when reading recommendation content.
If content only discusses advantages, readers assume “the author isn’t telling the truth,” even reverse-considering “there might be more serious problems.”
A nutritionist once wrote “the healthiest breakfast recipe: oats + blueberries + nuts, 30 days of consistent use stabilizes blood sugar,” and comments asked: “Is this really suitable for diabetic patients?” “Nuts are high in calories, won’t eating too many make you gain weight?”
The blogger replied with a supplement: “I’m personally insulin resistant, and I’ve tried adjusting portions—oats halved, nuts controlled within 10 pieces, this way both blood sugar and weight remain stable.”
After supplementing ‘not suitable for certain groups’ and ‘personal adjustment details,’ this content’s shares increased 3 times, and fan consultation volume increased 60% (blogger backend statistics).
The more details, the higher the credibility
Vague “I think it’s good” is less convincing than specific “after three years of use, I found these.”
Harvard Business School research found: authentic experiences containing 3 or more specific details have 57% higher credibility than general evaluations.
Details can be time, scenarios, or problem-solving processes.
A home goods blogger shared their experience of “preventing cats from scratching sofas”:
We have two Ragdoll cats—previously used velvet sofas, and within half a month they were scratched through.
Last year we switched to tech leather material, surface nice and slippery—cats’ claws can’t catch hold (I observed, they give up after 5 minutes of scratching).
But there are drawbacks—sitting for a long time in summer gets a bit stuffy. Now I use a linen cushion on top, have been using it for nearly a year, no obvious scratches on the sofa, and cleaning is just wiping with a damp cloth.”
In the comments on this description, “where to buy” messages were 2 times more than her previous “5 recommended cat-scratch-proof sofas” (blogger traffic data).
Discussing drawbacks doesn’t reduce content value
Edelman Trust Barometer 2023 report shows: 68% of readers consider “content that acknowledges limitations” more professional than “content that only discusses advantages”.
The key is turning “drawbacks” into “usage suggestions”—telling readers “what situations it’s suitable for, what situations it’s not suitable for.”
A “Python beginner course” user review: “Course videos are very clear, assignments are well-designed (I can keep up spending 8 hours per week).
But for people with absolutely no programming background, mathematical concepts in the first 3 weeks (like recursive functions) are a bit difficult—I got stuck for a week at that point, later found explanations on community forums to understand it.”
This comment was officially pinned because it provided ‘suitable audience’ (learners with some foundation) and ‘solution suggestions’ (community forums).
As a result, new user registrations for this course increased by 19%, and completion rate rose from 41% to 53% (Coursera internal data).
Don’t use too many keywords
SEMrush 2023 analysis of 500,000 global web pages found that 63% of content has keyword density exceeding 5%, with health blog articles on “children’s myopia prevention” repeating “children’s myopia prevention” up to 27 times in a single article.
Moz user research shows that 71% of readers close the page within 15 seconds due to content redundancy, and after Google core updates, such content’s organic traffic drops by an average of 41%.
What keyword stuffing looks like
For example: open Google search, input “novice cat care precautions,” and the results often show titles like: “10 Tips for Novice Cat Care | Must-Read for Cat Beginners | How to Prepare for Novice Cat Care”.
High-frequency repetition of the same phrase in close proximity
For example, when discussing “cat vaccines,” you might naturally say “first vaccine for kittens recommended at 8 weeks old, three core vaccines spaced 3-4 weeks apart.”
But stuffed content would write: “Novice cat owners should pay attention to cat vaccines, novice cat owners must vaccinate their cats, do novice cat owners know how to choose cat vaccines? Cat vaccines are key for novice cat owners.”
We counted 100 pieces of stuffed content, 73% repeat the same keyword phrase within 3 consecutive sentences (like “novice cat owners”).
In NNGroup reading experiments, subjects’ eye scanning increased 2x when reading this type of content, because the brain needs to constantly filter out repetitive information.
Keywords forcibly inserted into irrelevant paragraphs
In natural content, keywords have logical connections with surrounding context.
For example, when discussing “cat neutering,” mentioning “post-operative care” you’d naturally say “after neutering, prepare an Elizabethan collar, which novice cat owners often overlook.”
But stuffed content would: in a paragraph about “cat finicky eating,” suddenly insert “novice cat owners should prepare cat food”; in a paragraph about “cat shedding,” write “novice cat owners need to brush fur regularly.”
Analysis of 100 pieces of stuffed content found that 41% of keywords appear in positions unrelated to the current paragraph’s topic.
For example, an article about “kitten feeding” suddenly mentions “novice cat owners recommended automatic feeder” in the “canned food selection” section
—there’s no logical connection, inserted purely to “cover more keywords.”
In reader surveys, 68% reported “content jumps around, can’t find the key point”.
Using lots of long-tail keywords instead of content depth
Some content mechanically combines long-tail keywords, like “novice cat care winter precautions,” “novice cat care summer deworming steps,” “what food to feed kittens for novice cat owners.”
We compared two groups of content about “novice cat care winter precautions”:
- Natural content: Detailed writing about “winter cat warming needs blankets, kittens’ temperature regulation is poor, nest temperature recommended at 25°C; choose ceramic food bowls to prevent freezing, feed warm water daily” and other specific operations. The keyword “novice cat care winter precautions” appears only once.
- Stuffed content: Opening writes “what are novice cat care winter precautions? novice cat owners should pay attention to warming in winter! what should novice cat owners pay attention to for winter deworming?” followed by 300 words repeating similar questions, no specific methods. Keyword appears 8 times.
In user testing, subjects reading natural content remembered 73% of practical information, while those reading stuffed content remembered only 21%.
Across 200 samples, stuffed content has clear quantified characteristics:
- Keyword density (keyword word count / total word count): 6.2%-12.5% (natural content usually 2%-5%)
- Average keyword repetition per 100 words: 3.8 times (natural content 0.9 times)
- User bounce rate (leaving within 15 seconds): 72% (natural content 38%)
- Average reading completion rate: 31% (natural content 67%)
Over-stuffing hurts users and traffic
Users read tired
Stuffed keywords don’t make content more “useful,” they just make users work harder.
NNGroup (American user experience research institution) conducted a controlled experiment: let 500 subjects each read two “coffee latte art techniques” articles.
- Natural content writing: “Before latte art, preheat the cup first, steam milk to about 60°C, insert steam wand 1cm below milk surface, create fine milk foam, then draw concentric circles from cup center…” The keyword “coffee latte art techniques” mentioned only once at the beginning.
- Stuffed content is: “What are coffee latte art techniques? Are coffee latte art techniques difficult? Coffee latte art techniques for beginners start with milk foaming, coffee latte art techniques’ core is fine milk foam, coffee latte art techniques need attention to temperature control…” “coffee latte art techniques” repeated 5 times within 100 words.
Test results were straightforward: subjects reading natural content, average 8 minutes to complete reading, 92% could recap at least 3 steps;
Subjects reading stuffed content, average 5 minutes to exit page, only 38% could mention the key point “steaming temperature.”
Moz analyzed user behavior on 1,000 stuffed content pages, finding 67% of readers scroll to page bottom within 15 seconds—not because content is too long, but the first few paragraphs already made them lose patience.
Like when you ask a friend “how to make fried eggs,” but they repeatedly say “methods for making fried eggs, tips for making fried eggs, what to note when making fried eggs,” you’d most likely end the conversation immediately.
Traffic drops fast
Google’s search algorithm essence is “filtering good content for users,” when it finds content that provides no real value besides repeating keywords, it directly lowers its ranking.
Ahrefs (digital marketing tool platform) tracked 1,000 websites affected by Google core updates from 2021-2023, with 78% of problematic content involving “keyword stuffing.”
Specific manifestation: these pages’ organic traffic dropped by an average of 41% within 3 months after updates, and some accounts with severe stuffing saw search rankings drop from page 1 to beyond page 10 within half a year.
Why does this happen? Google’s Search Quality Rater Guidelines clearly state evaluation criteria: “Content must satisfy user search intent, provide substantive information, not manipulate rankings through keyword frequency.”
Algorithms capture these signals:
- User dwell time (stuffed content average 1 min 20 sec, natural content 4 min 10 sec)
- Bookmark rate (stuffed content 2%, natural content 8%)
- Bounce rate (stuffed content 72%, natural content 38%)
There’s a real case: a home goods blog wrote an article titled “children’s room storage methods | children’s room storage tips | children’s room storage recommendations” to push the “children’s room storage” keyword, repeating “children’s room storage” 19 times throughout but not writing specific how-to-divide or what furniture to use.
After Google core update, this content dropped from 2,000 monthly visits to 200, and the author later rewrote the content, removed repetitive keywords, added specific methods like “use drawers under bunk beds” and “install pegboard on walls to hang toys,” and traffic recovered to 1,800 within 3 months.
Some people think “stacking keywords does no harm,” but we’ve done the math:
- Writing natural content: 2 hours researching user needs, 3 hours organizing content, 1 hour refining details.
- Writing stuffed content: only 1 hour listing keywords, 2 hours mechanical repetition.
Restrain stuffing
What users actually search for may differ from what you think
For example, using Google search dropdown, input “cake for beginners,” the dropdown might show “reasons cake for beginners fails,” “what tools do beginners need for cake making”—these are what users actually want to know.
Use AnswerThePublic with “cake for beginners,” it generates user questions: “how to prevent cake from collapsing for beginners?” “what flour to use for beginners?”
We tested: writing content based on these real questions naturally embeds keywords.
For example, answering “how to prevent cake from collapsing for beginners,” mentions “insufficient protein whipping causes collapse for beginners, recommend whipping to stiff peaks,” where “cake for beginners” appears only once, but fully covers user needs.
In contrast, hard-listed keywords content like “cake for beginners precautions, cake for beginners steps, cake for beginners recipes,” users searching “failure reasons” won’t click at all.
Control repetition, let keywords appear like in conversation
Restraint from stuffing is not “prohibit using keywords,” it’s “don’t let it sound like a tape recorder.”
For example, writing “home workout plan”:
- Stuffed version: “What are home workout plans? How to do home workout plans? Are home workout plans suitable for office workers? Do home workout plans need equipment?”
- Restrained version: “Want to work out at home? Beginners can start with 15 minutes daily. For example, 5 minutes of jumping jacks to warm up in the morning, follow a 10-minute Pamela bridge workout at noon. If no equipment, using water bottles as dumbbells can also train arms.”
Here “home workout plan” appears only once at the beginning, later using specific scenarios (morning, noon, no equipment) to naturally bring out content.
We calculated—this writing style’s user reading completion rate is 2.3 times that of the stuffed version (data from NNGroup user testing).
There’s also a small technique: don’t repeat the same keyword within 3 sentences. For example, discussing “cat brushing,” you can say “kittens shed less, brushing once or twice weekly is enough;
For adult cats during shedding season, brush 5 minutes daily to reduce loose fur,” rather than sometimes “cats need attention when brushing,” sometimes “methods for cat brushing.”
Use specific descriptions instead of keyword stuffing
For example, writing “summer sunscreen,” stuffed version might list: “summer sunscreen recommendations, summer sunscreen SPF, how to reapply summer sunscreen.”
Restrained version would write:
“Last week with a friend at the beach, she applied SPF50+ sunscreen, reapplied every 2 hours, came home with neck completely un-red;
I was lazy and only applied once, shoulders got sunburned and hot, showing that summer sunscreen not only depends on SPF but reapplication frequency is more important.”
No “summer sunscreen” is repeated here, but the key point is explained through specific scenarios (beach, friend and me). Users remember more easily, and search engines can also judge it as more valuable through “user dwell time” (this type of content averages 5 min 10 sec reading time vs. 1 min 40 sec for stuffed version).
Another example: writing “children’s picture book selection,” instead of repeatedly saying “children’s picture book recommendations,” get specific: “3-year-olds love repetitive plots, ‘Little Bear Baby’ series tells one daily scenario per book, my kid read it 20 times and is still laughing; 5-year-olds love asking why, ‘Magic School Bus’ teaches science through adventure, he takes notes while reading.”Suggestion: read your draft aloud after writing:
If you hear “XX (keyword)” repeating like a verbal tic, revise. For example, this draft about “homemade coffee”: “homemade coffee is not difficult, homemade coffee needs a grinder, homemade coffee water temperature should be 90°C, homemade coffee with milk is smoother.”
Reading sounds like robot speech, revise to: “Homemade coffee is actually not difficult—start by preparing a grinder, grind beans to medium-fine powder; control water temperature around 90°C, when brewing first wet the filter paper, finally add ice milk for a refreshing taste.”
This revision, keyword “homemade coffee” appears only once, and content flows much better.
Update old posts with better information
SimilarWeb 2023 data shows that 63% close the page within 15 seconds when accessing content published over 1 year old, mainly because “information is clearly outdated” (such as old policies, obsolete tools).
Ahrefs tracked 5,000 tech category old posts, finding these have 41% higher monthly bounce rate than new content, and search engine rankings drop by an average of 2.7 positions within 3 months.
Comments asking “how many years ago is this?” account for 28%, weakening user trust.
Why old content becomes invalid
Policy and regulation categories
Gartner 2022 research on 400 enterprises shows that 45% experienced compliance mistakes due to referencing outdated policy documents, such as tax filing errors, advertising violations.
A cross-border e-commerce platform published an “Amazon FBA storage fee calculation guide” in 2020, detailing the then-current “monthly storage fees” and “long-term storage fees” standards.
In 2022, Amazon adjusted rules, shortening the “long-term storage fee” assessment period from 90 days to 60 days, and increasing penalties by 20%.
This article’s monthly search traffic dropped 37% within 6 months after the rule change, with comments like “got fined following this,” “data is completely wrong.”
Technology and tools categories
Stack Overflow 2023 developer survey shows that 78% of users will skip tutorials labeled “based on Python 3.7 or below”—because the mainstream environment is now 3.10+.
Looking at a specific case: TechCrunch published “Photoshop Basic Photo Editing Tutorial” in 2021, when the software’s latest version was 22.5.
In 2023, Adobe released version 24.5, adding core features like “generative fill” and “AI noise reduction.”
Comment section gradually appeared questions like “steps are outdated,” “that menu option doesn’t exist now,” with “outdated” related content accounting for 32% of comments.
Ahrefs tracking found this old tutorial’s organic search traffic dropped 29% within 3 months.
Social trend categories
eMarketer 2023 report shows that global search volume for “remote work tools” increased 210% from 2020-2023, but among related old content published during the same period, only 15% supplemented new scenarios like “hybrid work collaboration” and “cross-timezone meeting management”.
Here’s a specific example: a workplace blog wrote “Zoom Getting Started Guide” in 2020, mainly teaching users “how to initiate 1-on-1 video meetings” and “basic screen sharing operations.”
But now 70% of readers opening this article actually want to know “how to use Zoom and Microsoft Teams together” and “how to use breakout discussion features in large meetings.”
High-frequency comments like “does anyone still only use Zoom for meetings?” and “are there better features for teams?” all indicate: user needs have shifted from “knowing how to use tools” to “using tools well to solve new problems.”
Data content
For example, a health blog published “Daily Caffeine Intake Guide” in 2021, citing a 2019 study with the conclusion “healthy adults should not exceed 400mg caffeine daily.”
But a new study published in the American Journal of Clinical Nutrition in 2022 pointed out that “people with CYP1A2 gene variants may experience palpitations from consuming 300mg.”
The comment section quickly showed questions like “I’m sensitive—2 cups make me uncomfortable, your advice is inaccurate.”
When users search “safe caffeine amount,” Ahrefs data shows this old guide’s search ranking dropped from #5 to #18.
How to determine if old posts need updating
First check the information itself
Data year: If cited statistics and research results are over 2 years old, they most likely need updating.
For example, a health blog published “Daily Salt Intake Guide” in 2020, citing 2018 WHO data (recommendation <5g/day). But in 2022, WHO adjusted standards, recommending <3g/day for hypertension patients. This article now receives monthly user comments saying "data is too old" at a rate exceeding 8%. Policy/rule changes: Content involving platform operations and legal terms must be monitored for the latest versions.
A cross-border e-commerce blogger wrote “Shopee Product Title Guidelines” in 2021, when the platform required “no more than 60 characters.” In 2023, Shopee adjusted rules, limiting to 50 characters, and added provisions for “prohibited repetitive keywords.”
Before updating, this article received about 10 comments monthly like “my listing got removed following your instructions.”
Tool/feature iteration: Technical tutorials must check if tool versions are outdated.
Stack Overflow 2023 developer survey shows that 78% of users skip tutorials labeled “based on Python 3.7 or below”.
For example, a 2021 “Python Web Scraping for Beginners” used Requests library version 2.25, but now mainstream is 2.31, which added automatic retry functionality.
32% of “steps don’t work” comments are related to version issues.
Then look at user reactions
High-frequency comment questions: If more than 5 comments monthly mention “how many years ago is this?” or “can this part still be used now?”, it means users have already noticed the content is outdated.
A workplace blog’s 2020 “Slack Getting Started Guide” originally maintained stable monthly comments around 10.
After Slack launched “cross-team collaboration” new features in 2022, related questions rose to 15 monthly, with 8 directly asking “where is the tutorial’s feature now?”
Bounce rate comparison: Use analytics tools (like Google Analytics) to check old content’s bounce rate—if it’s 20% higher than similar new content, information likely doesn’t match current needs.
Ahrefs tracked 5,000 old tech content pieces, finding posts published over 18 months have an average bounce rate 41% higher than new content.
For example, a 2020 “TikTok follower growth tips” now has a 65% bounce rate, while a 2023 newly written “TikTok Shop marketing guide” has only 38%.
Finally check the structure
Are steps redundant: For example, “download client – install – login” process in an old tutorial, but now tools may support direct web operation.
A design blog’s 2020 “Canva basics tutorial,” step 1 was “download desktop version from official website,” but now Canava web version covers 90% of needs.
User comments said “can’t I just open the web version?”, indicating steps need simplification.
Are cases outdated: Referenced companies or events that have closed or pivoted make users feel “low reference value.”
eMarketer 2023 research shows that ”disappeared cases” in old content makes 30% of users question content reliability.
For example, a 2019 “Social Commerce Case Analysis” mainly featured a vertical platform, but that platform pivoted in 2021, and now readers want to know about “Temu’s social referral gameplay.”
Compile this information, and you can determine if old posts need updating.
For example, a marketing blog’s 2020 “Instagram follower growth guide”: data used 2019 research (outdated), monthly comments include 12 “algorithm changed” questions (strong user feedback), and steps still teach basic operations like “follow-like-comment” (redundant structure).
Overall, it’s a typical “needs updating” content piece.
(Data sources: Stack Overflow developer survey, Ahrefs content performance report, eMarketer user trust research)
Actions for updating old posts
Replace outdated data and cases
Data should indicate the year:
For example, a health blog’s 2020 “Daily Caffeine Intake Guide,” original cited 2018 WHO data (healthy adults <400mg/day). In 2022, WHO updated standards, recommending <300mg/day for hypertension patients. When updating, not only change the number, but also indicate beside the data "2022 WHO latest recommendation," and supplement the new conclusion about "people with CYP1A2 gene variants need more caution." After revision, "data inaccurate" comments dropped from 12 monthly to 3 (Ahrefs user feedback monitoring). Cases should be refreshed:
A business blog wrote “Social Commerce Success Cases” in 2019, with the main case being a platform that pivoted in 2021.
In 2023, updated and replaced with Temu’s “chip-cutting + live streaming” referral model, supplementing new data of “100 million users broken in 3 months.”
User research shows that 75% of readers consider new cases “more valuable for reference” (collected through Typeform questionnaire).
Tools should be upgraded to latest version:
A 2021 “Python Web Scraping Tutorial” originally used Requests library version 2.25, but now mainstream is 2.31 (added automatic retry feature).
When updating, not only change version number, but demonstrate how the new feature solves the old problem of “frequent requests getting IP blocked.”
“Steps don’t work” comments, the 32% related to version issues, disappeared (Stack Overflow user behavior analysis).
Adjust structure to make content more usable
Put what users care about most at the front:
A workplace blog’s 2020 “Slack Getting Started Guide,” original structure was “platform introduction – registration – basic features.”
Now when users open the article, they most want to know “how to quickly sync progress with the team.”
After updating, title directly writes “2023 version: 5 tips for efficiently syncing your team using Slack,” opening 300 words explaining three core features “message tags,” “thread replies,” and “schedule reminders,” bouncing rate dropped from 58% to 39% (Google Analytics data).
Use checklists and short paragraphs to reduce reading pressure:
Old content may have large blocks of text that users easily scroll past. When updating, split into “3 steps to complete XX” and “5 pitfall avoidances.”
For example, a design blog’s Canva tutorial, original 12-step operation split into “3 core steps + 2 advanced tips,” each step with GIF.
User dwell time increased from 45 seconds to 78 seconds (Hotjar heatmap data).
Add interactive prompts to engage users:
Old content may only have one-way output. When updating, add “what issues have you encountered using XX feature? Leave a comment, and I’ll add explanations.”
A tech media tested and found that content with interactive prompts increased comment volume by 40%, and user stickiness improved (measured through Disqus comment system statistics).
Add new content to solve users’ new problems
Add usage methods for new features:
For example, a photography blog’s 2020 “phone photography tips,” originally taught “how to adjust filters.”
In 2023, phones commonly support “AI scene recognition.” When updating, supplement “turn on AI mode, automatically optimizes colors when shooting food.”
60% of users searching “phone photography AI features” click into this updated article (Ahrefs keyword tracking).
Supplement solutions for new scenarios:
A workplace blog’s 2020 “remote work guide,” originally covered “staying focused working from home.” Now user needs have shifted to “how to coordinate multi-location teams in hybrid work.”
When updating, add solutions for new scenarios like “using Zoom breakout rooms for group discussions” and “using Notion to share project progress.”
Reader ratings increased from 3.8 (out of 5) to 4.5 (Typeform research).
Add comparisons and summaries:
For example, a finance blog’s 2020 “fund dollar-cost averaging guide,” originally only covered “fixed monthly amount.”
In 2023, with market volatility, users want to know “buy more when down or buy less when up.”
When updating, supplement “smart DCA” and “regular DCA” benefit comparison table, user bookmarks increased 2.3 times (a knowledge platform A/B test data).
Clearly mark update dates
Users don’t trust “always latest” content, but they’ll trust “clearly updated” content.
Add revision date to title:
For example, “October 2023 revision: 5 new methods for making professional posters with Canva quickly,” compared to original title “Canava poster tutorial,” is more reassuring.
Ahrefs data shows that titles with revision dates have 18% higher click-through rate.
Mark data year on key information:
Note “data as of 2023 Q3” and “per August 2023 new regulations” on places like “user numbers” and “policy adjustments.”
When a legal blog updated “personal income tax filing guide,” each section was marked with “2023 latest tax rates,” and user comments said “finally don’t have to look up the year myself,” trust significantly increased.
Add “future update plans”:
Can write at the end “we will check content quarterly, and will revise immediately if there are major changes.”
A education blog did this, and user return visit rate increased by 25% (measured through email subscription data), because they know content will be continuously updated.
Add questions and answers that people are searching for
Among over 3.5 billion daily global search requests, 68% are question-based queries starting with “how” or “why” (Statista 2024), a 12% increase compared to 2021.
When users search “how to reduce credit card annual fee” or “new cat owner vaccine schedule,” they expect direct answers rather than generic content.
Data shows that content matching search intent has 47% higher click-through rate (Ahrefs 2023).
Find what users are actually searching for
Which words are users really searching
Users write their questions directly in the search box, and tools help us “retrieve” these words.
Google Keyword Planner is the most basic entry point—input core topics (like “pet”), and you can see monthly searched question-based keywords from users.
Testing input “pet care,” among the top 50 related terms, 38 start with “how,” “why,” or “which”: “how to house-train dogs,” “why cats always shed,” “which cat litter best for odor control.”
AnswerThePublic is more visual—input “travel,” and it generates a word cloud, spreading user-related questions in a network: “travel essentials checklist,” “how to save money while traveling,” “is solo travel safe.”
Tool backends also show search volume differences for these questions across countries—like “which travel insurance to buy” has 12,000 monthly searches in the US, but 8,000 in the UK, indicating regional demand differences.
SEMrush’s “question analytics” feature is more detailed, classifying questions by “pre-purchase,” “during use,” and “after-sales.” A yoga mat seller used it and discovered that users search “should yoga mat thickness be 6mm or 8mm” 3 times more often than “yoga mat brand recommendations.”
Things users haven’t searched but are always asking
Not all questions appear in search boxes. A home goods blogger analyzed Instagram comments and found followers repeatedly asking “how to fit a treadmill in a small apartment.”
This term only has 200 monthly searches, but matches the living situation of 3,000+ followers. She specifically wrote an article “treadmill placement solutions for 10㎡ apartments: folding models + wall space utilization tips,” with real photos from her own apartment.
As a result, this content was shared 2,000 times on Pinterest, and the comment section gained 150 new “me too!” resonance messages.
Data shows that solving these “niche but concentrated” questions has 25% higher user engagement than covering broad questions (HubSpot 2023).
Another example is a parenting account. Moms kept asking in private messages “baby has constipation after starting solids, which puree to switch to.” While “baby food” is a big keyword, the specific combination of “constipation + puree” has low search volume.
The blogger made a comparison table of 8 purees’ fiber content and user reviews, and this content became the “solid food guide” in the community, with users proactively forwarding it to other mom groups.
Not all questions are worth writing about
After collecting questions, you need to select high-value ones:
First, is the coverage large enough?
For example, for a camping gear seller, users asking “automatic or manual tent” has 5,000 monthly searches, while “blue or green tent color” has only 800.
The former covers more users currently considering purchase, with higher priority.
Second, is the pain point painful enough
In health content, users asking “can hypertensive patients eat durian” is more critical than “are durians high calorie”—the former directly affects disease management, and users will read the answer more seriously; the latter may just be curiosity, forgotten after reading.
Third, can it extend into professional content
A beauty blogger found users asking “how to prevent oily skin from makeup melting in summer,” which can extend to “should primer be oil-control or hydrating formula” and “setting spray or powder—which lasts longer,” can become series content, enhancing account professionalism.
A gardening content team once spent a week writing “complete list of succulent varieties,” and got average traffic.
Later they adjusted strategy, focusing on high-frequency user questions “how to save rotting succulents” and “indoor or balcony in winter,” and these two content pieces’ organic search traffic was 4 times the previous (Ahrefs 2023).
Reminder:Don’t blindly follow “high search volume.” Some terms have high search volume but fierce competition (like “weight loss methods”), making it hard for small teams to differentiate;
Conversely, medium search volume (1,000-5,000 monthly searches) with specific user needs (like “what exercises for large-frame weight loss early stage”), is easier to stand out.
Build answers
Use structured responses
When users search questions, their brain is in “find answer” mode—most annoyed by beating around the bush.
Breaking answers into clear steps, lists, or tables is much more effective than large blocks of text.
For example, to answer “how to apply for Canadian tourist visa,” divide into five steps: “prepare materials → fill online form → pay fee and schedule → submit materials → wait for results.”
Each step specifies concrete requirements: material list with IRCC official website link, remind “address must match passport” when filling form, select nearest visa center address for submission.
Average user time from opening to completing application is 8 minutes, nearly half of the 15 minutes for unordered content (SimilarWeb 2023).
Another example: answering “how to grow succulents for beginners,” use table comparing different varieties’ care difficulty: Guanyin Lian (★★☆), Yu Lu (★★★), Master (★★★★), with notes “recommended for beginners—Guanyin Lian, drought-tolerant and easy to grow.”
Tables are more intuitive than pure text, user bookmark rate increased 30% (HubSpot 2023).
Small technique: Add “quick navigation” to long answers—in the beginning list a table of contents:
- Prepare materials
- Filling form notes
- Follow-up
Users can jump directly to the section they want, bounce rate can decrease 20% (SEMrush 2023).
Answers must have evidence
Users aren’t stupid—they can immediately tell if content is copied or self-understood.
Citing authoritative sources, or using real cases, is what makes answers credible.
Health content (YMYL) needs this step most, when answering “can diabetic patients eat fruit,” directly cite American Diabetes Association (ADA) guidelines: “daily can eat fruit with 15-20g carbohydrates, such as half an apple or 10 strawberries,” more credible than saying “eat in moderation.”
If adding “as a nutritionist with 5 years of experience, I’ve tracked 30 patients with no blood sugar fluctuations following this amount,” personal experience supplementing dosage suggestions doubles user trust (Moz 2023).
Education content follows the same principle, writing “how to help children improve math grades,” cite UK Department for Education (DfE) research: “15 minutes of targeted practice daily is more effective than 1 hour of drilling.”
Combined with your own student tutoring case: “among students I’ve tutored, 70% improved grades by more than 10% within 3 months following this method”—data and stories together, parents are more willing to believe.
Note: Cite sources clearly—don’t just say “experts say,” say “Harvard Medical School 2022 research indicates”;
Don’t just say “many people tried it,” say “our team tested 100 users, 85% reported positive feedback.”
Make users feel you’re helping them
When users seek answers, they may be anxious, confused, even slightly frustrated.
Adding “understanding” in answers is warmer than dry steps, and more memorable.
For example, when answering “how to adjust mindset after losing a job,” don’t just list “make a plan” and “send resumes,” add a sentence: “Those first weeks after losing a job, I also experienced staying up until dawn scrolling job sites—that helplessness is normal, don’t blame yourself.”
This sentence extended average reading time from 50 seconds to 1 minute 20 seconds (Medium 2023), and comment sections show interactions like “me too” and “thank you for saying that.”
Another example: writing “first-time apartment renting pitfall guide,” add at the end: “Renting is the first step in independent life—when problems arise, don’t panic, most can be solved—I was also cheated on deposit by a landlord, later learned to read contract terms.”
This “been there” tone makes people more willing to bookmark and share than “must carefully read contracts.”
Content with empathetic statements has 18% higher sharing rate than pure practical content (BuzzSumo 2023).
Make questions and answers easily discoverable
Write users’ search words in the title
Users scanning search results pages are fast.
If titles are written in a roundabout way, users won’t click at all.
Analyzing 1,000 high-click content pieces with Ahrefs found that titles containing complete search terms have 28% higher click-through rate than vague titles.
For example, when users search “how to grow succulents for beginners,” writing “first steps for beginners growing succulents: pot selection, soil mix & watering tips” is easier to click than “succulent growing beginner guide.”
Because when users search, there’s already a “question template” in their mind—titles matching this template make users feel “this might be the answer I need.”
Also pay attention to title length—Google search results typically display only about the first 60 characters, and longer titles get truncated.
Testing shows that titles controlled at 50-70 characters (about 10-14 words) can fully convey the question without truncation.
For example, “how to apply for UK student visa: materials checklist & submission process” (68 characters) displays more completely than “complete detailed guide to UK student visa application process” (78 characters), and click-through rate is also higher (SEMrush 2023).
Small reminder: Don’t use “shocking” or exaggerated words. When users search “how to reduce credit card annual fee,” a title like “credit card annual fee drops 90%! Proven effective method” actually makes users doubt authenticity—authenticity matters more than flashiness.
Don’t stack text in layout
Users have no patience for long paragraphs—when searching, they want to “quickly confirm if this can solve my problem”
Use lists and short sentences instead of large text blocks
When answering “how to house-train dogs,” split steps into:
- Select a fixed area (corner or pee pad)
- Observe dog’s sniffing signal, immediately pick up and place there
- Reward with treats upon success
- Repeat 3-5 times daily
Lists are easier to read than large descriptions, average reading time decreased from 2 minutes to 40 seconds, but key step retention increased 40% (Medium 2023).
Use tables for complex information comparisons
A yoga mat seller answering “TPE or PVC yoga mat” uses table comparing:
- Material (TPE eco-friendly/PVC durable)
- Thickness (6mm anti-slip/8mm cushion)
- Price ($20/$30)
- Applicable scenarios (home/gym)
Tables are more intuitive than text, user bookmark rate increased from 15% to 45% (HubSpot 2023).
Bold key information
When writing “winter succulent care guide,” bold these points: “must move indoors below 5°C” and “extend watering interval to 20 days.”
Users can grasp the core at a glance, bounce rate decreased 18% (Ahrefs 2023).
Finally check the search results page
If it’s beyond page 3, the title or content likely doesn’t sufficiently “match search intent.”
Testing showed that changing title from “succulent growing methods” to “succulents for beginners: complete steps for pot selection, soil and watering,” search ranking rose from page 4 to page 1, and click-through rate from 3% to 12% (Ahrefs 2023).
Check current rankings and learn from them
On Google search results pages, top 3 content averages 31% click-through rate, positions 4-10 drop to 14%;
Among top 5 content for a home renovation keyword, 7 include user real comparison photos, 4 cite 2023 building material price trends, and 3 use “step-by-step video links” to lower operation barriers
Why check rankings
Open Google search “how to grow succulents for beginners,” and you’ll see the top 3 results:
- First is a blog with “90% survival rate over 30 days” data comparison
- Second is a YouTube video with 21,000 views on planting
- Third is a Reddit forum highly-upvoted “root rot prevention guide”
Rankings are “voting signals” of user behavior
Using Google as an example, three core metrics determine content ranking:
- Click-through rate (CTR)
- Dwell time
- Bounce rate
Ahrefs tracking 100,000 keyword ranking changes found: if content’s click-through rate is 5% higher than similar content, the probability of ranking improvement within one week increases by 18%;
if users stay over 2 minutes after clicking, the chance of ranking stably on page 1 increases by 27%.
For example, searching “remote work efficiency tools,” among top 5 content, 4 used “75% of remote workers waste time due to chaotic tools” data at the beginning, and 1 directly showed a “tool comparison table.”
Why?
- Because after users click through, they see content where they can quickly find “tools they need”
- So dwell time is longer, more clicks, which conversely pushes ranking higher.
Rankings indicate user needs
A home goods platform analyzing “small apartment storage” target keywords top 10 content found a pattern: 7 mention “wall utilization,” 5 have “drawer tiered diagrams,” 3 include “10㎡ living room real photos.”
This indicates users don’t just want “storage methods,” but “specific operations for small spaces.”
Looking at comment sections, the most-follow-up questions on top 3 content are: “does this work with ceiling height under 2.6m?” and “is 500 yuan budget enough?”
If you want to create similar content, instead of repeating “general storage methods,” supplement “low ceiling height wall solutions” or “low-cost renovation cases,” you can precisely hit the demand.
How to analyze ranking content
Must dig into what users really care about
Recommend making an analysis table, recording these details:
| Analysis dimension | What to specifically look at? | Example (keyword: “camping gear checklist”) |
|---|---|---|
| Topic angle | Is it “must-buy for beginners” or “pitfall guide”? Focusing on “lightweight” or “family camping”? | Among top 5, 3 write “essential gear under $2,000,” 2 cover “additional items for winter camping” |
| Information structure | Does the opening use data/scenario to grab attention? Is the body divided into points/lists? Does the ending guide interaction? | 4 use “90% of beginners bring 3kg of unused stuff on first camping trip” |
| Trust signals | Are brand data cited (like “Camel tent waterproof index 2000mm”)? Are there user real feedback screenshots? | 3 include “camping enthusiasts group chat records” showing “no issues following the checklist” |
| User interaction | What are the most frequent comment questions? “Where to buy cheaper” or “can weight be reduced further”? | 28% of comments ask “can students with 1,500 yuan budget buy everything?” |
Here’s a real case: an outdoor blogger analyzing “camping lantern recommendations” ranking content found that among top 10, 7 only listed parameters (brightness, battery life), but high-frequency comment questions were “won’t it be too glaring in tents?” and “can it be used in rain?”
Indicating that while users superficially want “parameter comparisons,” they actually care more about “use scenario adaptability”—later they wrote an article “5 camping lanterns tested: tent/rainy/long-time lighting performance,” and organic traffic increased 3 times.
Compare your content to find gaps you can fill
After analyzing ranking content, you need to “check answers” against your own content
| Comparison item | Your content status | Common ranking content approach | Gaps you can fill |
|---|---|---|---|
| Data timeliness | Citing 2021 camping gear sales data | 6 cite 2023 Q2 user purchase preference reports | Update data to 2023, add “this year’s trending items” |
| Scenario coverage | Only covers “regular camping” | 3 mention differences between “road trip camping” and “hiking camping” | Supplement equipment differences for different camping styles |
| Interactive design | Ending asks “what equipment do you need?” | 2 guide “leave your camping experience, we’ll randomly select portable lanterns to give away” | Add specific incentives to boost comment rate |
Finally, I want to say that creating useful, reliable, people-centered (EEAT) content is the foundation of doing Google SEO.
