[{"data":1,"prerenderedAt":1921},["ShallowReactive",2],{"i-lucide:search":3,"i-lucide:instagram":8,"i-lucide:linkedin":11,"page-en-\u002Frecursos":13,"resources-list-en":34,"i-lucide:building-2":1913,"i-lucide:graduation-cap":1915,"i-lucide:landmark":1917,"i-lucide:arrow-up-right":1919},{"left":4,"top":4,"width":5,"height":5,"rotate":4,"vFlip":6,"hFlip":6,"body":7},0,24,false,"\u003Cg fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\">\u003Cpath d=\"m21 21l-4.34-4.34\"\u002F>\u003Ccircle cx=\"11\" cy=\"11\" r=\"8\"\u002F>\u003C\u002Fg>",{"left":4,"top":4,"width":5,"height":5,"rotate":4,"vFlip":6,"hFlip":6,"body":9,"hidden":10},"\u003Cg fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\">\u003Crect width=\"20\" height=\"20\" x=\"2\" y=\"2\" rx=\"5\" ry=\"5\"\u002F>\u003Cpath d=\"M16 11.37A4 4 0 1 1 12.63 8A4 4 0 0 1 16 11.37m1.5-4.87h.01\"\u002F>\u003C\u002Fg>",true,{"left":4,"top":4,"width":5,"height":5,"rotate":4,"vFlip":6,"hFlip":6,"body":12,"hidden":10},"\u003Cg fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\">\u003Cpath d=\"M16 8a6 6 0 0 1 6 6v7h-4v-7a2 2 0 0 0-2-2a2 2 0 0 0-2 2v7h-4v-7a6 6 0 0 1 6-6M2 9h4v12H2z\"\u002F>\u003Ccircle cx=\"4\" cy=\"4\" r=\"2\"\u002F>\u003C\u002Fg>",{"id":14,"title":15,"body":16,"description":23,"draft":6,"extension":24,"image":25,"meta":26,"navigation":10,"path":27,"seo":28,"seoDescription":29,"seoImage":25,"seoTitle":30,"service":31,"stem":32,"__hash__":33},"pages_en\u002Frecursos.md","Blog",{"type":17,"value":18,"toc":19},"minimark",[],{"title":20,"searchDepth":21,"depth":21,"links":22},"",2,[],"Articles and guides on social media trends. What recent research says, how to analyse Instagram and TikTok, and with which method.","md","\u002Fog-image.svg",{},"\u002Frecursos",{"title":15,"description":23},"A blog on social media trends. What recent research says, how to analyse Instagram and TikTok, and with which method. By the Trendia team.","Social media trends blog | Trendia",null,"recursos","hPKCv73WroCRMl5cBlrzRH-WjFFcT-s_tTNxWbp7-kQ",[35,179,305,449,554,652,784,900,1112,1211,1320,1507,1695],{"id":36,"title":37,"author":38,"authorRole":39,"body":40,"category":163,"date":164,"description":165,"draft":6,"extension":24,"featured":6,"image":166,"meta":167,"navigation":10,"path":168,"seo":169,"seoDescription":170,"seoImage":166,"seoTitle":171,"service":31,"stem":172,"summary":173,"tags":174,"__hash__":178},"resources_en\u002Frecursos\u002Fwhat-is-social-listening.md","What is social listening, and what changes when you analyse the video","David Villarreal","Co-founder & CEO",{"type":17,"value":41,"toc":157},[42,46,49,54,89,93,120,124,127,130,133,137,140,143],[43,44,45],"p",{},"Social listening is the analysis of what people say on social media about a brand, a product, a topic or a person. It does not stop at your own profiles. It listens to the public conversation, the one that happens even when nobody has tagged you.",[43,47,48],{},"It is different from your account statistics. Statistics tell you how what you posted performed. Social listening tells you what is being talked about, who is saying it and how that conversation changes.",[50,51,53],"h2",{"id":52},"what-it-is-for","What it is for",[55,56,57,65,71,77,83],"ul",{},[58,59,60,64],"li",{},[61,62,63],"strong",{},"Tracking a brand's reputation."," How much it is talked about, in what tone and when that changes.",[58,66,67,70],{},[61,68,69],{},"Spotting trends."," Which topics, formats or products start gaining weight in a category.",[58,72,73,76],{},[61,74,75],{},"Understanding the audience."," What worries people, what they ask for and what they criticise, in their own words.",[58,78,79,82],{},[61,80,81],{},"Watching competitors."," What is said about other brands and what content works for them.",[58,84,85,88],{},[61,86,87],{},"Anticipating a crisis."," Seeing a conversation heat up before it reaches the press.",[50,90,92],{"id":91},"how-it-works-in-four-steps","How it works, in four steps",[94,95,96,102,108,114],"ol",{},[58,97,98,101],{},[61,99,100],{},"Define the question."," \"What is being said about my brand?\" is not the same as \"Why has the conversation about my category dropped?\".",[58,103,104,107],{},[61,105,106],{},"Collect the conversation."," Posts, videos and comments that fit the question, on the agreed networks and period.",[58,109,110,113],{},[61,111,112],{},"Classify it."," Topics, tone, emotion, format. Today this is done with AI, but AI needs criteria and someone to check them.",[58,115,116,119],{},[61,117,118],{},"Read it."," Move from data to a conclusion that helps you decide something.",[50,121,123],{"id":122},"what-classic-social-listening-misses","What classic social listening misses",[43,125,126],{},"Social listening tools were built to read text: tweets, posts, news and forums. They work with keywords. If someone writes your brand name, it shows up.",[43,128,129],{},"But on TikTok and Instagram a large part of the conversation is not written. It is in what you see and hear. A video can talk about your product without naming it in the caption. It can use your category as the backdrop for a joke. It can show rejection with a gesture. None of that appears in a keyword search.",[43,131,132],{},"And even when the mention does appear, counting it does not explain why it worked. Two videos that name your brand can travel in opposite directions because of the format, the first seconds or the emotion they carry.",[50,134,136],{"id":135},"what-changes-when-you-analyse-the-video","What changes when you analyse the video",[43,138,139],{},"At Trendia we do social listening from inside the piece. We read the format, pace, emotion and narrative of each video and cross it with how people responded in the comments. The models are our own, checked against hand-annotated data, and we work with official access to Meta and TikTok.",[43,141,142],{},"It does not replace a mention-tracking tool. They complement each other. One tells you how much people talk. The other explains which content drives it.",[43,144,145,146,151,152,156],{},"To see how we apply it to a brand, here is ",[147,148,150],"a",{"href":149},"\u002Fen\u002Fbrands","social listening for brands",". And if you are choosing a tool, our guide to ",[147,153,155],{"href":154},"\u002Fen\u002Frecursos\u002Fsocial-listening-tools","social listening tools"," may help.",{"title":20,"searchDepth":21,"depth":21,"links":158},[159,160,161,162],{"id":52,"depth":21,"text":53},{"id":91,"depth":21,"text":92},{"id":122,"depth":21,"text":123},{"id":135,"depth":21,"text":136},"Methodology","2026-10-07","Social listening tracks what people say about a brand or a topic on social media. But on TikTok and Instagram the conversation is no longer only in the text. It is inside the video.","\u002Fblog\u002Fwhat-is-social-listening.png",{},"\u002Frecursos\u002Fwhat-is-social-listening",{"title":37,"description":165},"What social listening is, what it is for, and what it misses when the social media conversation moves from text to video.","What is social listening | Trendia","recursos\u002Fwhat-is-social-listening","Listening to social media started with counting mentions. Today the conversation runs through video, and that means looking inside every piece.",[175,176,177],"Social listening","Social media analysis","Video","CGn9r6QogC0ZnE8k6Kgg4wFFdjrdcbgOrFOJtvIIOdQ",{"id":180,"title":181,"author":38,"authorRole":39,"body":182,"category":289,"date":290,"description":291,"draft":6,"extension":24,"featured":10,"image":292,"meta":293,"navigation":10,"path":294,"seo":295,"seoDescription":296,"seoImage":292,"seoTitle":297,"service":31,"stem":298,"summary":299,"tags":300,"__hash__":304},"resources_en\u002Frecursos\u002Fsimulating-a-social-media-campaign-before-launch.md","Simulating a social media campaign before launch: what OranSim promises and what it does not",{"type":17,"value":183,"toc":283},[184,187,191,194,220,223,227,230,233,237,240,244,247,253,259,265,273],[43,185,186],{},"What if you could test a social media campaign before spending the budget? That is the idea behind OranSim, a system presented on 23 September by a team of six researchers, with the code published. It is a preprint: it has not been peer reviewed yet, so it is best read as a serious proposal and not as a settled result.",[50,188,190],{"id":189},"what-the-simulator-does","What the simulator does",[43,192,193],{},"OranSim links four decisions every social media marketing team makes to what happens next:",[55,195,196,202,208,214],{},[58,197,198,201],{},[61,199,200],{},"The creative",": which piece is published.",[58,203,204,207],{},[61,205,206],{},"The creator",": who publishes it.",[58,209,210,213],{},[61,211,212],{},"The targeting",": who sees it first.",[58,215,216,219],{},[61,217,218],{},"The budget",": how much is invested.",[43,221,222],{},"From there it simulates who sees the campaign, how they respond and how that response spreads across 60 population segments. To compare two scenarios it always starts from the same initial population, so the difference comes from the change you made and not from chance.",[50,224,226],{"id":225},"the-figures","The figures",[43,228,229],{},"The models that predict engagement were trained on 39,000 posts from RedNote, the Chinese social network, and tested on a further 12,154. They explain between 56% and 62% of the variation in engagement (R² of 0.56 to 0.62, on a log scale).",[43,231,232],{},"The most striking experiment is the budget one. When it is doubled, reach nearly doubles and the cumulative 14-day response rises to 1.96 times the baseline. But the average fit between the content and the people who see it drops, and so does the probability that each person engages. The extra budget reaches people who care less about the piece.",[50,234,236],{"id":235},"why-we-think-it-is-a-trend-worth-following","Why we think it is a trend worth following",[43,238,239],{},"Much of the work with simulated audiences consists of asking what a type of person would think. This is a different step: simulating the whole journey of a social media campaign, from first exposure to spread.",[50,241,243],{"id":242},"how-we-read-it-at-trendia","How we read it at Trendia",[43,245,246],{},"With interest, and with three cautions.",[43,248,249,252],{},[61,250,251],{},"It is for comparing, not for predicting."," A model that explains 60% leaves 40% unexplained. Saying \"scenario A performs better than B\" is reasonable. Promising a reach figure is not.",[43,254,255,258],{},[61,256,257],{},"The data comes from a single network."," Nothing guarantees that what was learned on RedNote carries over as is to Instagram or TikTok.",[43,260,261,264],{},[61,262,263],{},"The creative goes in as just another input."," And it is the hardest part to turn into data. What format a video has, what emotion it conveys and what story it tells is exactly what we analyse at Trendia, on real videos and with human review. A simulator will only be as good as what it knows about the piece it simulates.",[43,266,267,268,272],{},"If you want to know what is working in your category today before the next campaign, ",[147,269,271],{"href":270},"\u002Fen\u002Fque-quieres-saber","tell us what you need to know",".",[43,274,275,276,282],{},"Source: ",[147,277,281],{"href":278,"rel":279},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.28388",[280],"nofollow","OranSim: Simulating Social Media Marketing",", arXiv, 23 September 2026.",{"title":20,"searchDepth":21,"depth":21,"links":284},[285,286,287,288],{"id":189,"depth":21,"text":190},{"id":225,"depth":21,"text":226},{"id":235,"depth":21,"text":236},{"id":242,"depth":21,"text":243},"Trends","2026-10-06","A new study simulates how the reach and response of a social media campaign change with the creative, the creator, the targeting and the budget. We read it calmly.","\u002Fblog\u002Fsimulating-a-social-media-campaign-before-launch.png",{},"\u002Frecursos\u002Fsimulating-a-social-media-campaign-before-launch",{"title":181,"description":291},"OranSim simulates a social media campaign before launch: creative, creator, targeting and budget. What the study says and where its limits are.","Simulating a social media campaign before launch | Trendia","recursos\u002Fsimulating-a-social-media-campaign-before-launch","Doubling the budget nearly doubles reach, but it reaches people the content fits less well. That is what a newly published campaign simulator says.",[301,302,303],"Campaigns","Social media marketing","Simulation","QL6T36PRZBmskNWQ4v0n3RO-3Bph7Fo6VEyXOzhbNDA",{"id":306,"title":307,"author":38,"authorRole":39,"body":308,"category":434,"date":435,"description":436,"draft":6,"extension":24,"featured":6,"image":437,"meta":438,"navigation":10,"path":439,"seo":440,"seoDescription":441,"seoImage":437,"seoTitle":442,"service":31,"stem":443,"summary":444,"tags":445,"__hash__":448},"resources_en\u002Frecursos\u002Fmisinformation-fake-news-and-public-opinion.md","Fake news and misinformation: how to measure their real weight in public opinion",{"type":17,"value":309,"toc":428},[310,313,316,320,323,343,346,350,390,394,414,418,421],[43,311,312],{},"Every week some hoax circulates about a public body, a policy or a candidate. The usual reaction is to respond quickly. Sometimes that is right. Other times, responding to something that barely moved gives it the visibility it never had.",[43,314,315],{},"The difference is measurement. Not whether there is fake news, because there almost always is, but how much it weighs in public opinion.",[50,317,319],{"id":318},"misinformation-hoaxes-and-opinion-are-not-the-same","Misinformation, hoaxes and opinion are not the same",[43,321,322],{},"It helps to separate three things:",[55,324,325,331,337],{},[58,326,327,330],{},[61,328,329],{},"Misinformation:"," false or misleading content that circulates, with or without intent.",[58,332,333,336],{},[61,334,335],{},"A hoax, or fake news:"," one specific piece of misinformation, usually in a recognisable format. A video taken out of context, a doctored screenshot, an invented figure.",[58,338,339,342],{},[61,340,341],{},"Opinion:"," a judgement, however harsh or uncomfortable. It is not misinformation.",[43,344,345],{},"Mixing them up leads to mistakes in both directions: treating legitimate criticism as a hoax, or letting a hoax through because it looks like one more opinion. That is why, in a public opinion study, what counts as misinformation is agreed before the work starts.",[50,347,349],{"id":348},"five-questions-to-measure-its-weight","Five questions to measure its weight",[94,351,352,358,368,374,380],{},[58,353,354,357],{},[61,355,356],{},"How far has it circulated?"," Volume is the starting point, not the conclusion.",[58,359,360,363,364,272],{},[61,361,362],{},"Where has it circulated?"," A closed group of like-minded accounts is not the same as the general conversation. We have written about why ",[147,365,367],{"href":366},"\u002Fen\u002Frecursos\u002Fbots-and-coordinated-accounts-on-social-media","detecting bots does not prove they had influence",[58,369,370,373],{},[61,371,372],{},"Has it left its bubble?"," The key moment is when a hoax reaches people who were not looking for it.",[58,375,376,379],{},[61,377,378],{},"Has the conversation changed?"," If new topics, new words or a different tone appear in the comments after the hoax, it had an effect. If not, it was noise.",[58,381,382,385,386,272],{},[61,383,384],{},"Which format carried it?"," More and more, misinformation travels as video. A short video with a clear emotion spreads differently from text. Some studies suggest that falsehood moves ",[147,387,389],{"href":388},"\u002Fen\u002Frecursos\u002Fwhy-misinformation-spreads-faster-on-x","faster than truth on some networks",[50,391,393],{"id":392},"what-to-do-with-the-answer","What to do with the answer",[55,395,396,402,408],{},[58,397,398,401],{},[61,399,400],{},"If the hoax has not left its bubble",", responding can amplify it. Often, monitoring it is enough.",[58,403,404,407],{},[61,405,406],{},"If the conversation has changed",", a response is needed, and the conversation itself shows which format and tone to use.",[58,409,410,413],{},[61,411,412],{},"If it is part of a longer narrative",", debunking piece by piece is less useful than understanding the story that ties the pieces together.",[50,415,417],{"id":416},"public-bodies-and-political-parties","Public bodies and political parties",[43,419,420],{},"For a public body, misinformation affects trust in a service or a policy. For a political party, especially during an election campaign, it affects how a candidate or a proposal is perceived. The method is the same: public data, agreed criteria and a documented reading with its pieces and its limits.",[43,422,423,424,272],{},"At Trendia we do not rule on what is true. We identify the pieces that shift the narrative, document how they circulate and measure whether they moved public opinion. Here is how we do it for ",[147,425,427],{"href":426},"\u002Fen\u002Fpublic-sector","public bodies, institutions and political parties",{"title":20,"searchDepth":21,"depth":21,"links":429},[430,431,432,433],{"id":318,"depth":21,"text":319},{"id":348,"depth":21,"text":349},{"id":392,"depth":21,"text":393},{"id":416,"depth":21,"text":417},"Public opinion","2026-10-04","A hoax circulating does not mean it changed what people think. For a public body or a political party, the useful question is not whether there is fake news, but how much it weighs.","\u002Fblog\u002Fmisinformation-fake-news-and-public-opinion.png",{},"\u002Frecursos\u002Fmisinformation-fake-news-and-public-opinion",{"title":307,"description":436},"How to measure the weight of fake news and misinformation in public opinion on social media. Five questions for public bodies and parties.","Fake news, misinformation and public opinion | Trendia","recursos\u002Fmisinformation-fake-news-and-public-opinion","Detecting a hoax is the first step. Knowing whether it moved the conversation, and whose, is what lets you decide whether and how to respond.",[434,446,447],"Disinformation","Fake news","JFfJs-7wZ6g3fa08NKYLQtNmd9vhGZJZInAtclH42l4",{"id":450,"title":451,"author":38,"authorRole":39,"body":452,"category":289,"date":540,"description":541,"draft":6,"extension":24,"featured":6,"image":542,"meta":543,"navigation":10,"path":544,"seo":545,"seoDescription":546,"seoImage":542,"seoTitle":547,"service":31,"stem":548,"summary":549,"tags":550,"__hash__":553},"resources_en\u002Frecursos\u002Fsocial-media-trend-shifts-the-conversation-turns-first.md","The conversation turns before the data: how to spot a trend shift on social media",{"type":17,"value":453,"toc":534},[454,457,461,464,468,476,479,483,486,488,491,513,516,519,526],[43,455,456],{},"People who follow social media trends tend to arrive late for the same reason: they watch the indicator that confirms the change, not the one that announces it. A study published as a preprint on 24 September gives a clue about where to look earlier.",[50,458,460],{"id":459},"what-it-analyses","What it analyses",[43,462,463],{},"The authors study 79 markets on Polymarket, a platform where people bet money on the outcome of future events. Each market leaves two trails: the buy and sell trades, and the participants' comments. The study compares which of the two gives warning first.",[50,465,467],{"id":466},"what-it-finds","What it finds",[55,469,470,473],{},[58,471,472],{},"Attention in the comments anticipates periods of high activity over the following 18 hours almost as well as the trading history.",[58,474,475],{},"In 23 of the 27 markets where it could be assessed, the stance of the comments moved against the leading option before that option lost the lead.",[43,477,478],{},"The authors speak of predictive association, not cause. The comments signal the turn; it is not shown that they bring it about.",[50,480,482],{"id":481},"why-it-matters","Why it matters",[43,484,485],{},"A prediction market is not Instagram or TikTok. But it is one of the few places where you can compare, almost in real time, what people say with what people do. And the result fits the starting idea of any trend study: the discourse moves before the numbers.",[50,487,243],{"id":242},[43,489,490],{},"Volume alone says little. A topic can have more and more conversation and be losing support at the same time. To read a trend shift it helps to look at once at:",[55,492,493,498,503,508],{},[58,494,495],{},[61,496,497],{},"How much is being said.",[58,499,500],{},[61,501,502],{},"Which way the conversation leans.",[58,504,505],{},[61,506,507],{},"How much disagreement there is.",[58,509,510],{},[61,511,512],{},"How fast it is changing.",[43,514,515],{},"The most interesting case is the narrative that is still in the majority but has already started to give way. In the count it still wins. In the comments it no longer convinces in the same way. For a brand, that can be a format that keeps adding views while the comments fill up with fatigue.",[43,517,518],{},"Three cautions: 27 markets are few, the study has not yet been peer reviewed, and it is about betting, not consumption. We take it as a well-posed hypothesis, not as a rule.",[43,520,521,522,272],{},"If you want to know where the conversation in your category is heading, see ",[147,523,525],{"href":524},"\u002Fen#como-funciona","how a trend study works",[43,527,275,528,533],{},[147,529,532],{"href":530,"rel":531},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.28965",[280],"The Conversation Turns First: Crowd Discussion and Price Reversals in Prediction Markets",", arXiv, 24 September 2026.",{"title":20,"searchDepth":21,"depth":21,"links":535},[536,537,538,539],{"id":459,"depth":21,"text":460},{"id":466,"depth":21,"text":467},{"id":481,"depth":21,"text":482},{"id":242,"depth":21,"text":243},"2026-10-02","A study of 79 prediction markets finds that comments change sides before the outcome does. What it means for anyone following social media trends.","\u002Fblog\u002Fsocial-media-trend-shifts-the-conversation-turns-first.png",{},"\u002Frecursos\u002Fsocial-media-trend-shifts-the-conversation-turns-first",{"title":451,"description":541},"The conversation changes before the indicators do. A study of 79 prediction markets shows how comments anticipate a social media trend shift.","How to spot a trend shift on social media | Trendia","recursos\u002Fsocial-media-trend-shifts-the-conversation-turns-first","In 23 of 27 cases, the comments turned against the favourite before it lost the lead. The conversation gives warning.",[289,551,552],"Conversation analysis","Early signals","5-pIoMbfyEkFjwfxbQQ9GUBpcDTfNtcyyWdGh2TdrmE",{"id":555,"title":556,"author":38,"authorRole":39,"body":557,"category":289,"date":637,"description":638,"draft":6,"extension":24,"featured":6,"image":639,"meta":640,"navigation":10,"path":641,"seo":642,"seoDescription":643,"seoImage":639,"seoTitle":644,"service":31,"stem":645,"summary":646,"tags":647,"__hash__":651},"resources_en\u002Frecursos\u002Fwhy-misinformation-spreads-faster-on-x.md","Why misinformation spreads faster on X: the algorithm adds all engagement into a single score",{"type":17,"value":558,"toc":631},[559,562,566,569,572,576,579,583,586,588,591,611,614,617,624],[43,560,561],{},"That misinformation on social media travels faster than verified information has been known for years. What was missing was the mechanical reason. A study by Pan Li and Shuang Gao, published as a preprint on 24 September, looks for it inside X's own recommendation algorithm, which they examine component by component.",[50,563,565],{"id":564},"the-mechanism-all-engagement-goes-into-the-same-sum","The mechanism: all engagement goes into the same sum",[43,567,568],{},"The algorithm predicts what each user will do with a post (like it, repost it, reply or quote it) and combines those predictions into a single score, a weighted sum. The authors call this engagement fungibility: for the score, one reaction can be swapped for another.",[43,570,571],{},"The consequence is that a post can be amplified through instant reactions even when hardly anyone stops to reply. And misinformation attracts that kind of reaction above all.",[50,573,575],{"id":574},"what-they-tested","What they tested",[43,577,578],{},"Using X data on the 2024 US elections, they found that adjusting the weight of each metric changes very little. What does work, according to the study, is a prior condition: content is not amplified until the system predicts a minimum of thoughtful engagement. Across 46 robustness checks, that gate reduced exposure to low-credibility content without reducing exposure or engagement for mainstream content.",[50,580,582],{"id":581},"why-it-matters-beyond-x","Why it matters beyond X",[43,584,585],{},"Because almost everyone measures engagement the way that algorithm does: by adding it up. And two posts with the same total can be very different things. One prompted a thousand passing likes. The other, two hundred replies from people who stopped to write.",[50,587,243],{"id":242},[43,589,590],{},"When we analyse a trend we look at what kind of response it gets, not only how much. It helps to separate three things:",[55,592,593,599,605],{},[58,594,595,598],{},[61,596,597],{},"Quick reaction",": like, view, share without comment.",[58,600,601,604],{},[61,602,603],{},"Thoughtful response",": a comment with an argument, a quote, a reply video.",[58,606,607,610],{},[61,608,609],{},"Amplification",": how much new exposure follows.",[43,612,613],{},"That is why comments are part of the analysis and not an appendix. In the project with the UOC on mental health on TikTok, the corpus included more than 2,200 videos and more than 67,000 comments.",[43,615,616],{},"One caution: the study is about X and a specific electoral context. The TikTok and Instagram algorithms are not public, so it cannot be assumed they work the same way. What does carry over is the question: is this engagement peak interest, or is it reflex?",[43,618,619,620,272],{},"You can see ",[147,621,623],{"href":622},"\u002Fen\u002Fprojects","how we work on a real project",[43,625,275,626,533],{},[147,627,630],{"href":628,"rel":629},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.28947",[280],"Why Does Misinformation Propagate Faster? An Algorithmic Perspective on X",{"title":20,"searchDepth":21,"depth":21,"links":632},[633,634,635,636],{"id":564,"depth":21,"text":565},{"id":574,"depth":21,"text":575},{"id":581,"depth":21,"text":582},{"id":242,"depth":21,"text":243},"2026-09-30","A study examines X's recommendation algorithm piece by piece and finds the mechanism. Content can be amplified by quick reactions alone, without anyone stopping to reply.","\u002Fblog\u002Fwhy-misinformation-spreads-faster-on-x.png",{},"\u002Frecursos\u002Fwhy-misinformation-spreads-faster-on-x",{"title":556,"description":638},"X's algorithm adds all engagement into a single score. A study explains why that speeds up misinformation on social media.","Why misinformation spreads faster on X | Trendia","recursos\u002Fwhy-misinformation-spreads-faster-on-x","Two posts with the same engagement may have nothing in common. A study of X's algorithm explains why.",[648,649,650],"Engagement","X algorithm","Misinformation","CP7rQ8rgtnGjhPuT7LBEORaCu_Kq0TaAE5en56bBet0",{"id":653,"title":654,"author":38,"authorRole":39,"body":655,"category":163,"date":770,"description":771,"draft":6,"extension":24,"featured":6,"image":772,"meta":773,"navigation":10,"path":774,"seo":775,"seoDescription":776,"seoImage":772,"seoTitle":777,"service":31,"stem":778,"summary":779,"tags":780,"__hash__":783},"resources_en\u002Frecursos\u002Fmarket-research-with-social-media.md","How to do market research with social media",{"type":17,"value":656,"toc":760},[657,660,663,667,670,681,684,688,691,695,698,702,705,709,712,716,719,723,743,746,750],[43,658,659],{},"Classic market research asks: surveys, interviews, focus groups. Market research on social media observes. It collects what people post and comment about a category when nobody is asking them anything.",[43,661,662],{},"Both are useful and neither replaces the other. A survey gives you a representative sample and answers to closed questions. Social media gives you people's real language, the topics you had not thought to ask about and changes as soon as they start.",[50,664,666],{"id":665},"step-1-a-business-question","Step 1. A business question",[43,668,669],{},"\"What is being said about my category\" is too broad. Something specific works better:",[55,671,672,675,678],{},[58,673,674],{},"What stops people from trying this kind of product?",[58,676,677],{},"What do they expect from a brand in my sector that they do not find today?",[58,679,680],{},"Which market trends are gaining weight in my category this year?",[43,682,683],{},"The question decides everything else: which networks, which period and which content goes in.",[50,685,687],{"id":686},"step-2-where-and-when-to-look","Step 2. Where and when to look",[43,689,690],{},"Each network has its own audience and its own way of talking. TikTok and Instagram now hold a large share of the conversation about consumption, especially among younger audiences. The period matters as much as the network. A window that is too short mistakes a spike for a trend, and one that is too long mixes different moments.",[50,692,694],{"id":693},"step-3-what-content-goes-in","Step 3. What content goes in",[43,696,697],{},"Define the corpus: which accounts, hashtags, topics or searches. Include content from brands, creators and ordinary users, because each tells part of the story. And note what is left out. An honest study says what it did not look at.",[50,699,701],{"id":700},"step-4-classification-criteria","Step 4. Classification criteria",[43,703,704],{},"Before analysing, agree on how you will classify what you find: topics, motivations, barriers, emotions, formats. This is what turns thousands of videos and comments into something comparable. If you use AI, these criteria are what you need to teach it, and part of the sample has to be checked by hand.",[50,706,708],{"id":707},"step-5-from-data-point-to-consumer-insight","Step 5. From data point to consumer insight",[43,710,711],{},"A data point is \"30% of comments mention the price\". A consumer insight explains what is behind it: perhaps it is not the price, but that nobody explains why it costs what it costs. Getting there means reading the pieces, not just counting them. With video, it also helps to look at which format and emotion come with each topic.",[50,713,715],{"id":714},"step-6-a-conclusion-you-can-act-on","Step 6. A conclusion you can act on",[43,717,718],{},"The study is done when someone can make a decision with it: what to communicate, what to launch, what to stop doing. Back it with real examples and with the limits of the data.",[50,720,722],{"id":721},"what-a-study-like-this-cannot-tell-you","What a study like this cannot tell you",[55,724,725,731,737],{},[58,726,727,730],{},[61,728,729],{},"It is not representative of the whole population."," It speaks for those who post and comment, not for those who only watch.",[58,732,733,736],{},[61,734,735],{},"It does not measure purchase intent."," It measures conversation. Sometimes the two go together and sometimes they do not.",[58,738,739,742],{},[61,740,741],{},"It depends on data access."," You can only analyse what is public and what each platform allows you to query.",[43,744,745],{},"That is why it works better as a complement to traditional market research than as a replacement.",[50,747,749],{"id":748},"how-we-do-it-at-trendia","How we do it at Trendia",[43,751,752,753,756,757,272],{},"We analyse TikTok and Instagram videos from the inside, with our own models checked by people and official access to Meta and TikTok, and deliver the reading with its data and its limits. See how we approach ",[147,754,755],{"href":149},"market research and consumer insights for brands",", or a ",[147,758,759],{"href":622},"real project from start to finish",{"title":20,"searchDepth":21,"depth":21,"links":761},[762,763,764,765,766,767,768,769],{"id":665,"depth":21,"text":666},{"id":686,"depth":21,"text":687},{"id":693,"depth":21,"text":694},{"id":700,"depth":21,"text":701},{"id":707,"depth":21,"text":708},{"id":714,"depth":21,"text":715},{"id":721,"depth":21,"text":722},{"id":748,"depth":21,"text":749},"2026-09-29","What people post and comment about a category is a source of market research that does not depend on a survey. Here is how to set it up, step by step, and where its limits are.","\u002Fblog\u002Fmarket-research-with-social-media.png",{},"\u002Frecursos\u002Fmarket-research-with-social-media",{"title":654,"description":771},"How to do market research with social media in six steps: question, sample, analysis and consumer insights. With its strengths and its limits.","How to do market research with social media | Trendia","recursos\u002Fmarket-research-with-social-media","Six steps to turn the TikTok and Instagram conversation into consumer insights, and what a study like this can and cannot tell you.",[781,782,176],"Market research","Consumer insights","zSMVJpt9YsVx-vawx6MPP0p8FALLOyfyWqqOf0nhqNM",{"id":785,"title":786,"author":38,"authorRole":39,"body":787,"category":163,"date":885,"description":886,"draft":6,"extension":24,"featured":6,"image":887,"meta":888,"navigation":10,"path":889,"seo":890,"seoDescription":891,"seoImage":887,"seoTitle":892,"service":31,"stem":893,"summary":894,"tags":895,"__hash__":899},"resources_en\u002Frecursos\u002Fai-and-human-validation-in-social-media-analysis.md","AI and human validation: how to analyse discourse on social media without trusting the model blindly",{"type":17,"value":788,"toc":879},[789,792,796,799,802,814,816,819,823,826,828,831,856,859,862,872],[43,790,791],{},"Classifying thousands of messages with a language model is easy and cheap. Knowing whether it did it well is another matter. A study published as a preprint on 23 September by independent researcher Shahan Ahmed is a good example, small and reproducible, of how to answer that second question.",[50,793,795],{"id":794},"what-it-does","What it does",[43,797,798],{},"It analyses how seven English-language outlets in Bangladesh headlined the country's 2026 measles outbreak. It gathers 403 headlines and asks a language model to classify each one on two dimensions: sentiment and stance, the latter with four closed categories.",[43,800,801],{},"It then compares what the model says with what human reviewers say on 153 headlines. Agreement is measured with Cohen's kappa, which runs from 0 (the agreement you would get by chance) to 1 (total agreement):",[55,803,804,809],{},[58,805,806],{},[61,807,808],{},"Stance: 0.89.",[58,810,811],{},[61,812,813],{},"Sentiment: 0.75.",[50,815,467],{"id":466},[43,817,818],{},"Negative headlines went from 56% to 88% as the outbreak progressed, and those amplifying risk from 44% to 84%. Blame, by contrast, appeared rarely, in around 9% of headlines. And in 86% of those cases it was attributed to systemic failures, not to named political figures.",[50,820,822],{"id":821},"the-detail-that-interests-us","The detail that interests us",[43,824,825],{},"Press headlines are not social media, but the method is the one needed to analyse discourse on TikTok or Instagram. And one figure deserves attention: people and the model agree more on stance (0.89) than on sentiment (0.75). Sentiment is more ambiguous than it seems. That is why a \"positive sentiment\" percentage with no validation behind it deserves little trust.",[50,827,243],{"id":242},[43,829,830],{},"The order matters:",[94,832,833,839,845,851],{},[58,834,835,838],{},[61,836,837],{},"Closed categories",", defined before classifying.",[58,840,841,844],{},[61,842,843],{},"A sample annotated by hand"," by people.",[58,846,847,850],{},[61,848,849],{},"Measure the agreement"," between the people and the model.",[58,852,853],{},[61,854,855],{},"Only then, apply AI to the whole set.",[43,857,858],{},"This is the approach we work with. In the project with the UOC on mental health on TikTok there was a taxonomy of 15 labels, 350 videos annotated by hand as a reference and, after that, the classification of the full sample, more than 2,200 videos.",[43,860,861],{},"AI lets you reach thousands of pieces. Human validation is what lets you trust the result.",[43,863,619,864,867,868,272],{},[147,865,866],{"href":622},"the project with the UOC"," or how we work with ",[147,869,871],{"href":870},"\u002Fen\u002Funiversities","research teams",[43,873,275,874,282],{},[147,875,878],{"href":876,"rel":877},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.28362",[280],"Threat Amplified, Blame Restrained: LLM-Assisted Media Framing Analysis of the 2026 Bangladesh Measles Outbreak",{"title":20,"searchDepth":21,"depth":21,"links":880},[881,882,883,884],{"id":794,"depth":21,"text":795},{"id":466,"depth":21,"text":467},{"id":821,"depth":21,"text":822},{"id":242,"depth":21,"text":243},"2026-09-28","A study classifies 403 headlines with a language model and checks the result against human reviewers. The order in which it does so is the lesson.","\u002Fblog\u002Fai-and-human-validation-in-social-media-analysis.png",{},"\u002Frecursos\u002Fai-and-human-validation-in-social-media-analysis",{"title":786,"description":886},"Discourse analysis and sentiment analysis with AI: a study validates a language model against human reviewers and shows how to do it properly.","AI and human validation in social media analysis | Trendia","recursos\u002Fai-and-human-validation-in-social-media-analysis","Categories first, then human review, and only then AI at scale. A small study does it well and shows why sentiment analysis calls for more caution.",[896,897,898],"AI","Human validation","Sentiment analysis","r7SlTlRxUyK-kogNIqCV8ZoIJKYDfVwyuKTLilPaYbM",{"id":901,"title":902,"author":38,"authorRole":39,"body":903,"category":1098,"date":1099,"description":1100,"draft":6,"extension":24,"featured":6,"image":1101,"meta":1102,"navigation":10,"path":1103,"seo":1104,"seoDescription":1105,"seoImage":1101,"seoTitle":1106,"service":31,"stem":1107,"summary":1108,"tags":1109,"__hash__":1111},"resources_en\u002Frecursos\u002Fsocial-listening-tools.md","Social listening tools: what each one measures and which you need",{"type":17,"value":904,"toc":1091},[905,908,911,915,918,938,941,945,948,965,969,972,989,993,996,1016,1019,1023,1026,1078,1081],[43,906,907],{},"Search for a social listening tool and you will find names like Brandwatch, Talkwalker, Meltwater or Sprout Social. They all help you understand social media, but they do not look at the same thing. Comparing prices or features without knowing what you need to answer often ends with a tool nobody on the team opens.",[43,909,910],{},"This guide sorts them by what they measure.",[50,912,914],{"id":913},"_1-mention-tracking","1. Mention tracking",[43,916,917],{},"These are social listening tools in the strict sense. They track what is published on social networks, news sites, forums and blogs based on keywords, and summarise it as volume, tone and reach.",[55,919,920,926,932],{},[58,921,922,925],{},[61,923,924],{},"They answer:"," how much people talk about my brand, where and in what tone.",[58,927,928,931],{},[61,929,930],{},"They work well with:"," text. Written posts, headlines and comments.",[58,933,934,937],{},[61,935,936],{},"They miss:"," what is not written. A video about your category that never names it will not show up in a keyword search.",[43,939,940],{},"Brandwatch, Talkwalker and Meltwater belong to this family, each leaning more towards news media or social media depending on the case.",[50,942,944],{"id":943},"_2-native-analytics-on-each-network","2. Native analytics on each network",[43,946,947],{},"Instagram and TikTok give professional profiles their own statistics.",[55,949,950,955,960],{},[58,951,952,954],{},[61,953,924],{}," how what I posted performed. Reach, views, interactions and followers.",[58,956,957,959],{},[61,958,930],{}," your own accounts. They are free and come straight from the source.",[58,961,962,964],{},[61,963,936],{}," everything that is not yours. There are no statistics for competitors or for the wider conversation.",[50,966,968],{"id":967},"_3-social-media-management-suites","3. Social media management suites",[43,970,971],{},"Tools for scheduling posts, replying to messages and pulling the statistics of several profiles into one dashboard. Some, such as Sprout Social, also include a listening module.",[55,973,974,979,984],{},[58,975,976,978],{},[61,977,924],{}," how I organise daily work and report on my profiles.",[58,980,981,983],{},[61,982,930],{}," teams that manage many accounts.",[58,985,986,988],{},[61,987,936],{}," the why. They show which pieces worked, but not what those pieces had in common.",[50,990,992],{"id":991},"_4-video-content-analysis","4. Video content analysis",[43,994,995],{},"The newest family, and the one that matches how TikTok and Instagram are consumed today. Instead of counting mentions, it analyses what happens inside each piece: the format, the first seconds, the emotion, the narrative, and how people respond in the comments.",[55,997,998,1004,1010],{},[58,999,1000,1003],{},[61,1001,1002],{},"It answers:"," why one piece of content works and another does not, for your brand, your competitors or a whole category.",[58,1005,1006,1009],{},[61,1007,1008],{},"It works well with:"," short video, which is where much of the conversation now takes shape.",[58,1011,1012,1015],{},[61,1013,1014],{},"It misses:"," the total volume of mentions in news and forums. That is what the first family is for.",[43,1017,1018],{},"This is what we do at Trendia, with our own models checked by people and official access to Meta and TikTok.",[50,1020,1022],{"id":1021},"how-to-choose","How to choose",[43,1024,1025],{},"Start with the question, not the tool:",[1027,1028,1029,1042],"table",{},[1030,1031,1032],"thead",{},[1033,1034,1035,1039],"tr",{},[1036,1037,1038],"th",{},"If you want to know…",[1036,1040,1041],{},"You need",[1043,1044,1045,1054,1062,1070],"tbody",{},[1033,1046,1047,1051],{},[1048,1049,1050],"td",{},"How much your brand is talked about, and in what tone",[1048,1052,1053],{},"Mention tracking",[1033,1055,1056,1059],{},[1048,1057,1058],{},"How your posts perform",[1048,1060,1061],{},"Native analytics",[1033,1063,1064,1067],{},[1048,1065,1066],{},"How to organise and report daily work",[1048,1068,1069],{},"A management suite",[1033,1071,1072,1075],{},[1048,1073,1074],{},"Why content works and what to repeat",[1048,1076,1077],{},"Video analysis",[43,1079,1080],{},"Many brands end up combining two. The most common pairing is one tool that counts and another that explains.",[43,1082,1083,1084,1088,1089,272],{},"If you are still getting the concepts straight, start with ",[147,1085,1087],{"href":1086},"\u002Fen\u002Frecursos\u002Fwhat-is-social-listening","what social listening is",". And to see video analysis applied to your brand, here is ",[147,1090,150],{"href":149},{"title":20,"searchDepth":21,"depth":21,"links":1092},[1093,1094,1095,1096,1097],{"id":913,"depth":21,"text":914},{"id":943,"depth":21,"text":944},{"id":967,"depth":21,"text":968},{"id":991,"depth":21,"text":992},{"id":1021,"depth":21,"text":1022},"Tools","2026-09-26","Not every social listening tool looks at the same thing. Some count mentions, some manage profiles and some analyse the content. Before choosing, it helps to know which question you want to answer.","\u002Fblog\u002Fsocial-listening-tools.png",{},"\u002Frecursos\u002Fsocial-listening-tools",{"title":902,"description":1100},"Social listening tools compared by what they measure: mentions, native analytics, profile management and video analysis. Which one you need.","Social listening tools compared | Trendia","recursos\u002Fsocial-listening-tools","Mention tracking, native analytics, management suites and video analysis. Four families of tools for four different questions.",[175,1098,1110],"Social media analytics","aNPGSeSw73uUvdn9WTckSMb0qA948fyflMkD3x5onZ4",{"id":1113,"title":1114,"author":38,"authorRole":39,"body":1115,"category":434,"date":1198,"description":1199,"draft":6,"extension":24,"featured":6,"image":1200,"meta":1201,"navigation":10,"path":1202,"seo":1203,"seoDescription":1204,"seoImage":1200,"seoTitle":1205,"service":31,"stem":1206,"summary":1207,"tags":1208,"__hash__":1210},"resources_en\u002Frecursos\u002Fbots-and-coordinated-accounts-on-social-media.md","Bots and coordinated accounts on social media: detecting them does not prove they had influence",{"type":17,"value":1116,"toc":1192},[1117,1120,1124,1127,1131,1145,1148,1150,1153,1155,1158,1172,1175,1178,1184],[43,1118,1119],{},"Every time a controversy grows on social media the same explanation appears: \"it was bots\". Sometimes there really were coordinated accounts. But their presence does not mean they were the ones that moved the conversation. A study published as a preprint on 21 September by Niccolò Di Marco, Matteo Cinelli, Shinichi Nakano and Andrea Frosini focuses on that second step: measuring influence after detection.",[50,1121,1123],{"id":1122},"what-they-propose","What they propose",[43,1125,1126],{},"Two complementary methods. The first reconstructs the diffusion cascades (who shares from whom) and calculates how much influence the coordinated accounts had against the maximum they could in theory have reached. The second works on the repost network and compares the coordinated accounts with the best-placed accounts in that network.",[50,1128,1130],{"id":1129},"what-they-find","What they find",[55,1132,1133,1139],{},[58,1134,1135,1138],{},[61,1136,1137],{},"2019 UK general election",": coordinated accounts achieved only a small fraction of the influence they could in theory have reached.",[58,1140,1141,1144],{},[61,1142,1143],{},"State-linked information operations in several countries",": large differences between campaigns. In several, the coordinated accounts held advantageous positions in the network and their influence was comparable to, or even greater than, that of the most central accounts.",[43,1146,1147],{},"The two methods agree, which suggests the result depends on the structure of each network and not on how it is measured.",[50,1149,482],{"id":481},[43,1151,1152],{},"Because answering \"were there bots?\" almost never closes the case. For a public body, an institution or a party, reacting to a campaign that moved nothing can give it the visibility it lacked. And not reacting to one that did carry weight is the opposite mistake.",[50,1154,243],{"id":242},[43,1156,1157],{},"In an analysis of public opinion on social media it helps to separate four questions that tend to get mixed up:",[94,1159,1160,1163,1166,1169],{},[58,1161,1162],{},"Was there coordination?",[58,1164,1165],{},"Where were those accounts in the conversation, at the centre or on the edge?",[58,1167,1168],{},"Would the content have circulated anyway without them?",[58,1170,1171],{},"How much of the spread can be attributed to them?",[43,1173,1174],{},"The third is the one most often forgotten, and it means looking at the content as well as the accounts: what was said, in what format and with what emotion. A narrative that already resonated with people does not need much help to grow.",[43,1176,1177],{},"The study does not offer a universal figure. Its contribution is a more demanding way of asking the question.",[43,1179,1180,1181,272],{},"If you work in a public body, an institution or a party, here is ",[147,1182,1183],{"href":426},"how we analyse public opinion on social media",[43,1185,275,1186,1191],{},[147,1187,1190],{"href":1188,"rel":1189},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.24398",[280],"Optimal and heuristic strategies for evaluating the influence of coordinated behavior in information cascades and retweet networks",", arXiv, 21 September 2026.",{"title":20,"searchDepth":21,"depth":21,"links":1193},[1194,1195,1196,1197],{"id":1122,"depth":21,"text":1123},{"id":1129,"depth":21,"text":1130},{"id":481,"depth":21,"text":482},{"id":242,"depth":21,"text":243},"2026-09-24","A study proposes measuring how much spread is really due to coordinated accounts. In the 2019 UK election, very little. In other campaigns, as much as the most central accounts.","\u002Fblog\u002Fbots-and-coordinated-accounts-on-social-media.png",{},"\u002Frecursos\u002Fbots-and-coordinated-accounts-on-social-media",{"title":1114,"description":1199},"Detecting bots or coordinated accounts on social media does not prove they changed the conversation. A study measures their real influence.","Bots and coordinated accounts on social media | Trendia","recursos\u002Fbots-and-coordinated-accounts-on-social-media","Counting bots is not measuring their effect. A study compares campaigns and finds everything from minimal influence to operations that weigh as much as central accounts.",[434,446,1209],"Bots","2vP9GZaPbHtn6gGB547L3lAsd_FOyTcJ2bFSoJ5iGPs",{"id":1212,"title":1213,"author":38,"authorRole":39,"body":1214,"category":163,"date":1306,"description":1307,"draft":6,"extension":24,"featured":6,"image":1308,"meta":1309,"navigation":10,"path":1310,"seo":1311,"seoDescription":1312,"seoImage":1308,"seoTitle":1313,"service":31,"stem":1314,"summary":1315,"tags":1316,"__hash__":1319},"resources_en\u002Frecursos\u002Ftime-windows-in-social-media-trend-analysis.md","Minutes, hours or weeks: the time window decides which social media trends you see",{"type":17,"value":1215,"toc":1300},[1216,1219,1223,1226,1228,1254,1257,1261,1264,1266,1269,1283,1286,1292],[43,1217,1218],{},"\"Trend\" does not mean the same thing if you look at the last hour or the last quarter. It sounds obvious, but few reports say which period they used and why. A paper by Georgios Panayiotou, Lorenzo Mannocci and Maurizio Tesconi, published on 18 September and accepted at the AIDEM 2026 workshop, puts numbers on that decision.",[50,1220,1222],{"id":1221},"what-they-study","What they study",[43,1224,1225],{},"How coordinated behaviour is detected on social media, that is, groups of accounts sharing the same thing in a synchronised way. To detect it you have to set a time window. Do two accounts count as coordinated if they share the same thing ten seconds apart? And a day apart? The authors test different windows on 14 datasets of verified information operations.",[50,1227,1130],{"id":1129},[55,1229,1230,1236,1242,1248],{},[58,1231,1232,1235],{},[61,1233,1234],{},"Short windows, from ten seconds to one minute",": they capture highly synchronised coordination, with high precision.",[58,1237,1238,1241],{},[61,1239,1240],{},"Long windows, from hours to days",": they bring out slower, more diffuse patterns, but also more noise.",[58,1243,1244,1247],{},[61,1245,1246],{},"The number of communities detected grows"," as the window gets longer.",[58,1249,1250,1253],{},[61,1251,1252],{},"Overlap between windows barely matters",": using consecutive, non-overlapping windows gives very similar results and costs less to compute.",[43,1255,1256],{},"The authors' conclusion is that window length is not a technical setting. It is a decision that determines which phenomenon you are observing.",[50,1258,1260],{"id":1259},"why-it-matters-beyond-bots","Why it matters beyond bots",[43,1262,1263],{},"Because the same happens with any social media trend analysis. A sound that explodes on TikTok for 48 hours, a format that grows over three months and a shift in values that takes years are different things, and each is only visible at its own scale. Looking only at the week turns every peak into a trend. Looking only at the year erases them all.",[50,1265,243],{"id":242},[43,1267,1268],{},"Two simple rules:",[94,1270,1271,1277],{},[58,1272,1273,1276],{},[61,1274,1275],{},"Look at more than one scale."," What appears in hours is momentum. What holds for weeks starts to be a trend.",[58,1278,1279,1282],{},[61,1280,1281],{},"Always say which window was used."," The period should be written in the report, in plain sight, like the source of the data.",[43,1284,1285],{},"Before trusting a trend report, ask which period it covers. If it does not say, half the answer is missing.",[43,1287,1288,1289,272],{},"You can see how we approach it in a ",[147,1290,1291],{"href":622},"real project",[43,1293,275,1294,1299],{},[147,1295,1298],{"href":1296,"rel":1297},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.21959",[280],"Setting the clock: Evaluating temporal window parameters for coordinated behavior detection",", arXiv, 18 September 2026.",{"title":20,"searchDepth":21,"depth":21,"links":1301},[1302,1303,1304,1305],{"id":1221,"depth":21,"text":1222},{"id":1129,"depth":21,"text":1130},{"id":1259,"depth":21,"text":1260},{"id":242,"depth":21,"text":243},"2026-09-22","A study of 14 verified campaigns shows that the length of the analysis window changes what is detected. It holds for coordination and for any trend analysis.","\u002Fblog\u002Ftime-windows-in-social-media-trend-analysis.png",{},"\u002Frecursos\u002Ftime-windows-in-social-media-trend-analysis",{"title":1213,"description":1307},"In social media trend analysis, the period you choose decides what you find. A study of 14 campaigns shows it and leaves a lesson in method.","Time windows in social media trend analysis | Trendia","recursos\u002Ftime-windows-in-social-media-trend-analysis","With windows of seconds one thing shows up. With windows of days, another. Choosing the period is not a technical detail.",[163,1317,1318],"Trend analysis","Coordinated behaviour","aa7OimVw7CSEBCQ6p-W-im5QwPjESAgcnN_b0rrT-W4",{"id":1321,"title":1322,"author":38,"authorRole":39,"body":1323,"category":1098,"date":1493,"description":1494,"draft":6,"extension":24,"featured":10,"image":1495,"meta":1496,"navigation":10,"path":1497,"seo":1498,"seoDescription":1499,"seoImage":1495,"seoTitle":1500,"service":31,"stem":1501,"summary":1502,"tags":1503,"__hash__":1506},"resources_en\u002Frecursos\u002Finstagram-analytics-tools.md","Instagram analytics tools: what each type does and which one you need",{"type":17,"value":1324,"toc":1481},[1325,1328,1332,1337,1340,1346,1352,1356,1359,1364,1369,1373,1376,1381,1386,1390,1398,1403,1408,1412,1415,1441,1444,1448,1455,1462,1466,1469,1475],[43,1326,1327],{},"Almost everyone lands on the same frustration: three tools open, none of them tell you what to publish on Tuesday, and you keep deciding on instinct. The problem is rarely the specific tool. It is that four different categories get bought as if they were the same thing.",[50,1329,1331],{"id":1330},"the-four-categories","The four categories",[1333,1334,1336],"h3",{"id":1335},"_1-management-suites","1. Management suites",[43,1338,1339],{},"They schedule posts, centralise the inbox across networks, manage team permissions and produce the monthly report. Metricool, for example, covers eleven platforms and has a free plan for one profile per network.",[43,1341,1342,1345],{},[61,1343,1344],{},"What they solve:"," operations. If the problem is that publishing takes half the morning or that you run eight client accounts and lose the thread, this is yours.",[43,1347,1348,1351],{},[61,1349,1350],{},"What they do not solve:"," judgment. They show you what happened, not why or what to do next.",[1333,1353,1355],{"id":1354},"_2-native-analytics","2. Native analytics",[43,1357,1358],{},"Instagram Insights is free and it is the primary source. Nobody will have more reliable data on your account than the platform itself.",[43,1360,1361,1363],{},[61,1362,1344],{}," the truth about your account. Reach, saves, video retention, audience data.",[43,1365,1366,1368],{},[61,1367,1350],{}," context. You do not know if 3% engagement is good or bad in your category, and you cannot see anyone else’s private data.",[1333,1370,1372],{"id":1371},"_3-social-listening","3. Social listening",[43,1374,1375],{},"They track brand mentions and public conversation, usually with sentiment analysis.",[43,1377,1378,1380],{},[61,1379,1344],{}," reputation and conversation. Knowing what is said about you and when something is going wrong.",[43,1382,1383,1385],{},[61,1384,1350],{}," content performance. Counting mentions does not explain why a Reel works.",[1333,1387,1389],{"id":1388},"_4-content-analysis","4. Content analysis",[43,1391,1392,1393,1397],{},"This is the youngest category. Instead of counting posts or mentions, they break down the content itself: what structure the hook has, how the video is cut, which patterns the pieces that perform in a category share. This is where ",[147,1394,1396],{"href":1395},"\u002Fen\u002Fcreadores","Trendia"," works.",[43,1399,1400,1402],{},[61,1401,1344],{}," the creative decision. Which format, which opening, which angle.",[43,1404,1405,1407],{},[61,1406,1350],{}," if what you need is scheduling across eleven networks, this is not your category.",[50,1409,1411],{"id":1410},"how-to-know-which-one-you-are-missing","How to know which one you are missing",[43,1413,1414],{},"A fast diagnosis is to look at where your team gets stuck:",[55,1416,1417,1423,1429,1435],{},[58,1418,1419,1422],{},[61,1420,1421],{},"“I cannot get posts out”"," → management suite.",[58,1424,1425,1428],{},[61,1426,1427],{},"“I publish but I do not know if it is going well”"," → native analytics, then a category comparison.",[58,1430,1431,1434],{},[61,1432,1433],{},"“I do not know what people say about us”"," → social listening.",[58,1436,1437,1440],{},[61,1438,1439],{},"“I do not know what to publish”"," → content analysis.",[43,1442,1443],{},"That last one is usually the gap, because none of the other three solve it and it is the most expensive one to leave open: it is the difference between publishing three times a week and publishing three times a week something that works.",[50,1445,1447],{"id":1446},"what-no-tool-can-give-you","What no tool can give you",[43,1449,1450,1451,1454],{},"It is worth saying because the market does not: ",[61,1452,1453],{},"from an account that is not yours, nobody has reach, saves or watch time."," Instagram does not expose that data. Any tool that gives you those figures for a competitor is estimating, and an estimate presented as a fact is worse than not having the number.",[43,1456,1457,1458,1461],{},"What is public and comparable: posting frequency, formats, themes, visible engagement, which pieces break each account’s own average, and how all of that evolves over time. That is enough to build a ",[147,1459,1460],{"href":1395},"solid competitive benchmark"," — and it is honest about what it measures.",[50,1463,1465],{"id":1464},"combine-instead-of-replace","Combine instead of replace",[43,1467,1468],{},"The right question is almost never “which of these five do I pick?”. It is “which layer am I missing?”. Many teams have operations solved and lack judgment; others have good analysis and cannot publish on time.",[43,1470,1471,1472,272],{},"If you already use a management suite and you are asking what Trendia adds on top, ",[147,1473,1474],{"href":1395},"see how the software works",[43,1476,1477,1478,272],{},"And if you would rather see it with your own accounts, ",[147,1479,1480],{"href":1395},"the trial is seven days with full access",{"title":20,"searchDepth":21,"depth":21,"links":1482},[1483,1490,1491,1492],{"id":1330,"depth":21,"text":1331,"children":1484},[1485,1487,1488,1489],{"id":1335,"depth":1486,"text":1336},3,{"id":1354,"depth":1486,"text":1355},{"id":1371,"depth":1486,"text":1372},{"id":1388,"depth":1486,"text":1389},{"id":1410,"depth":21,"text":1411},{"id":1446,"depth":21,"text":1447},{"id":1464,"depth":21,"text":1465},"2026-08-08","Management, analytics, listening and content analysis are four different tool categories. Choosing badly means paying for something that does not solve your problem.","\u002Fblog\u002Finstagram-analytics-tools.png",{},"\u002Frecursos\u002Finstagram-analytics-tools",{"title":1322,"description":1494},"Management, native analytics, social listening and content analysis: what each Instagram tool type solves and how to pick the one you are missing.","Instagram analytics tools: a guide by type | Trendia","recursos\u002Finstagram-analytics-tools","Four tool categories, four different problems. Which one you need depends on what is missing.",[1098,1504,1505],"Instagram","Analytics","u-joxNQoIoa3qAloFSs7nxkSuqWfccXmGmG-o7Dfz1s",{"id":1508,"title":1509,"author":38,"authorRole":39,"body":1510,"category":1680,"date":1681,"description":1682,"draft":6,"extension":24,"featured":10,"image":1683,"meta":1684,"navigation":10,"path":1685,"seo":1686,"seoDescription":1687,"seoImage":1683,"seoTitle":1688,"service":31,"stem":1689,"summary":1690,"tags":1691,"__hash__":1694},"resources_en\u002Frecursos\u002Fhow-to-analyze-instagram-competitors.md","How to analyse your Instagram competitors: method, metrics and limits",{"type":17,"value":1511,"toc":1672},[1512,1515,1519,1526,1529,1533,1540,1543,1547,1550,1588,1595,1598,1602,1605,1619,1626,1630,1637,1656,1659,1663,1666],[43,1513,1514],{},"Competitive analysis on Instagram usually ends in a folder of screenshots nobody opens again. Not because it was done badly, but because the last step is missing: turning what you saw into a decision. This is the method we use.",[50,1516,1518],{"id":1517},"step-1-define-the-category-not-the-sector","Step 1: define the category, not the sector",[43,1520,1521,1522,1525],{},"The most common mistake is picking business competitors. For content, your competition is ",[61,1523,1524],{},"the accounts competing for the same attention",", which is not always the same thing.",[43,1527,1528],{},"If you sell technical running shoes, your content competitor may be a trail-running creator with 40,000 followers, not the other shoe brand. A useful set is 5 to 12 accounts: fewer does not give a pattern, more becomes unmanageable.",[50,1530,1532],{"id":1531},"step-2-normalise-before-you-compare","Step 2: normalise before you compare",[43,1534,1535,1536,1539],{},"Comparing an account with 500,000 followers to one with 8,000 in absolute numbers says nothing. Each account should be measured ",[61,1537,1538],{},"against its own history",": which post breaks its average, not which one has more likes in absolute terms.",[43,1541,1542],{},"That frame change is what turns a data table into a reading. The interesting part is not that the large account has more reach: it is that one of its pieces did three times what it normally does, because that is where there is something to learn.",[50,1544,1546],{"id":1545},"step-3-measure-what-is-actually-public","Step 3: measure what is actually public",[43,1548,1549],{},"What you can compare with real data:",[55,1551,1552,1558,1564,1570,1576,1582],{},[58,1553,1554,1557],{},[61,1555,1556],{},"Frequency and cadence"," — how much each account publishes and whether it holds or goes in bursts.",[58,1559,1560,1563],{},[61,1561,1562],{},"Format mix"," — Reels, carousels, single image, and how the mix has changed.",[58,1565,1566,1569],{},[61,1567,1568],{},"Themes"," — what the category talks about and which territories are already taken.",[58,1571,1572,1575],{},[61,1573,1574],{},"Visible engagement"," — likes and comments, always relative to the account itself.",[58,1577,1578,1581],{},[61,1579,1580],{},"Standout posts"," — the pieces that break the average, and what they share.",[58,1583,1584,1587],{},[61,1585,1586],{},"Evolution"," — how the strategy has changed over the last six months.",[43,1589,1590,1591,1594],{},"And what you ",[61,1592,1593],{},"cannot"," know from someone else’s account, no matter what a tool shows you: reach, impressions, saves, shares and watch time. Instagram does not expose third-party data. If those numbers appear, they are estimates, and you should know that before you build a strategy on top.",[43,1596,1597],{},"Saying this out loud in front of a client is more credible than showing an invented number.",[50,1599,1601],{"id":1600},"step-4-look-for-the-pattern-not-the-piece","Step 4: look for the pattern, not the piece",[43,1603,1604],{},"This is where useful analysis splits from an inventory. Once you have the pieces that break each account’s average, the question is what they have in common.",[55,1606,1607,1610,1613,1616],{},[58,1608,1609],{},"Do they open with a question, a strong claim or a data point?",[58,1611,1612],{},"How long are the videos that work versus the ones that do not?",[58,1614,1615],{},"Is there a person on camera or not?",[58,1617,1618],{},"When does the brand appear?",[43,1620,1621,1622,1625],{},"That last point weighs more than it looks: retention changes noticeably depending on when the brand element comes in. It is exactly the kind of pattern that only appears when you analyse ",[147,1623,1624],{"href":149},"the first seconds systematically"," instead of watching videos one by one.",[50,1627,1629],{"id":1628},"step-5-turn-the-pattern-into-a-gap","Step 5: turn the pattern into a gap",[43,1631,1632,1633,1636],{},"The deliverable is not “this is what the competition does”. It is ",[61,1634,1635],{},"“this is what we are going to do”",". Three questions close the analysis:",[94,1638,1639,1645,1651],{},[58,1640,1641,1644],{},[61,1642,1643],{},"Which territory is saturated?"," If all twelve accounts do the same thing, that space no longer differentiates.",[58,1646,1647,1650],{},[61,1648,1649],{},"Which territory is free and fits us?"," A gap that does not suit you is not an opportunity.",[58,1652,1653],{},[61,1654,1655],{},"Which format performs in the category and we are not using?",[43,1657,1658],{},"With those three answers you have a content decision you can defend in a meeting without falling back on “I think”.",[50,1660,1662],{"id":1661},"how-often-to-repeat-it","How often to repeat it",[43,1664,1665],{},"A full benchmark each quarter and a light review each month is usually the right balance. More often confuses noise with signal; less often and you arrive late to format changes, which move fast on Instagram.",[43,1667,1668,1669,272],{},"If doing it by hand every quarter costs more than it gives you, that is exactly the work Trendia automates: you can ",[147,1670,1671],{"href":1395},"analyse your whole category with a seven-day free trial",{"title":20,"searchDepth":21,"depth":21,"links":1673},[1674,1675,1676,1677,1678,1679],{"id":1517,"depth":21,"text":1518},{"id":1531,"depth":21,"text":1532},{"id":1545,"depth":21,"text":1546},{"id":1600,"depth":21,"text":1601},{"id":1628,"depth":21,"text":1629},{"id":1661,"depth":21,"text":1662},"Strategy","2026-08-07","A five-step method to turn an Instagram benchmark into content decisions, and which metrics you cannot compare no matter what a tool promises.","\u002Fblog\u002Fhow-to-analyze-instagram-competitors.png",{},"\u002Frecursos\u002Fhow-to-analyze-instagram-competitors",{"title":1509,"description":1682},"Five-step method to analyse your Instagram competitors: which metrics to compare, which ones are invented, and how to turn the benchmark into decisions.","How to analyse Instagram competitors | Trendia","recursos\u002Fhow-to-analyze-instagram-competitors","Five steps from a folder of screenshots to a content decision you can defend.",[1692,1504,1693],"Competitors","Benchmark","f3VP1ic_PN6JQOxShqZ3Ql_qnTiZkYNkeXWIBDb8W9g",{"id":1696,"title":1697,"author":1698,"authorRole":1699,"body":1700,"category":1898,"date":1899,"description":1900,"draft":6,"extension":24,"featured":6,"image":1901,"meta":1902,"navigation":10,"path":1903,"seo":1904,"seoDescription":1905,"seoImage":1901,"seoTitle":1906,"service":31,"stem":1907,"summary":1908,"tags":1909,"__hash__":1912},"resources_en\u002Frecursos\u002Ftypes-of-hooks-for-reels.md","Types of hooks for Reels: six families and how to know if they work","Pol Guasch","Co-founder & CTO",{"type":17,"value":1701,"toc":1885},[1702,1705,1712,1716,1720,1723,1734,1738,1741,1749,1753,1756,1764,1768,1771,1779,1783,1786,1794,1798,1801,1809,1813,1816,1830,1833,1837,1840,1846,1852,1858,1865,1869,1876,1879],[43,1703,1704],{},"The hook decides whether your video exists. That is not a metaphor: if nobody gets past the first seconds, the rest of the work — the script, the edit, the location — reaches no one. That is why Trendia specifically analyses that opening window across tens of thousands of videos.",[43,1706,1707,1708,1711],{},"What follows is not a list of formulas to copy. These are the six families that show up again and again, and above all how to know which one works ",[61,1709,1710],{},"in your category",", which is the only thing that matters.",[50,1713,1715],{"id":1714},"the-six-families","The six families",[1333,1717,1719],{"id":1718},"_1-question","1. Question",[43,1721,1722],{},"Opens with a direct question to the viewer. It works because the brain answers before it decides whether it wants to keep watching.",[43,1724,1725,1729,1730,1733],{},[1726,1727,1728],"em",{},"When to use it:"," when your audience has a recognisable problem. ",[1726,1731,1732],{},"Risk:"," if the question is too obvious, it reads as advertising and gets discarded.",[1333,1735,1737],{"id":1736},"_2-counter-intuitive-claim","2. Counter-intuitive claim",[43,1739,1740],{},"Says something that contradicts what the viewer takes for granted. It creates a tension that only resolves by watching the rest.",[43,1742,1743,1745,1746,1748],{},[1726,1744,1728],{}," in categories full of clichés. ",[1726,1747,1732],{}," if you do not back the claim afterwards, you burn trust.",[1333,1750,1752],{"id":1751},"_3-data","3. Data",[43,1754,1755],{},"Opens with a concrete figure. It works because it is specific and checkable.",[43,1757,1758,1760,1761,1763],{},[1726,1759,1728],{}," professional audiences, B2B content. ",[1726,1762,1732],{}," a number without a source is noise dressed as data.",[1333,1765,1767],{"id":1766},"_4-promise","4. Promise",[43,1769,1770],{},"States explicitly what the person who stays will take away. The most direct of all.",[43,1772,1773,1775,1776,1778],{},[1726,1774,1728],{}," educational or process content. ",[1726,1777,1732],{}," if the promise is generic, it does not stand out from a thousand other videos.",[1333,1780,1782],{"id":1781},"_5-scene-in-motion","5. Scene in motion",[43,1784,1785],{},"Explains nothing: it starts with the action already happening. The viewer stays to understand what is going on.",[43,1787,1788,1790,1791,1793],{},[1726,1789,1728],{}," visual content, behind the scenes, product in use. ",[1726,1792,1732],{}," if the scene is not visually clear in one second, you lose them.",[1333,1795,1797],{"id":1796},"_6-identification","6. Identification",[43,1799,1800],{},"Names who it is talking to. “If you run client accounts, this will sound familiar.”",[43,1802,1803,1805,1806,1808],{},[1726,1804,1728],{}," defined niches. ",[1726,1807,1732],{}," it trades reach for depth. Sometimes that pays off, sometimes it does not.",[50,1810,1812],{"id":1811},"structure-tone-and-emotion","Structure, tone and emotion",[43,1814,1815],{},"The hook family is only one of three variables. The other two weigh the same:",[55,1817,1818,1824],{},[58,1819,1820,1823],{},[61,1821,1822],{},"Tone"," — the same hook changes completely if it is said in a close register or an institutional one.",[58,1825,1826,1829],{},[61,1827,1828],{},"Emotion"," — what it activates in the first seconds: curiosity, urgency, recognition, discomfort.",[43,1831,1832],{},"When we analyse openings we measure all three together, because a question in a corporate tone and the same question in a conversational tone do not perform remotely the same.",[50,1834,1836],{"id":1835},"how-to-know-if-your-hook-works","How to know if your hook works",[43,1838,1839],{},"This is the part almost nobody does. Three ways, from least to most reliable:",[43,1841,1842,1845],{},[61,1843,1844],{},"1. Retention in the first seconds."," In your own Instagram metrics you have the playback curve. If it drops off a cliff before second three, the problem is the hook, not the content.",[43,1847,1848,1851],{},[61,1849,1850],{},"2. Comparison against your own average."," A video is not judged in absolute terms, but against what your account usually does. Ten thousand views are excellent or mediocre depending on where you come from.",[43,1853,1854,1857],{},[61,1855,1856],{},"3. Pattern by type."," If you group your last thirty videos by hook family and look at the average retention of each group, you see which family works with your audience. Almost always one clearly outperforms, and almost always it is not the one the team believed.",[43,1859,1860,1861,1864],{},"That third analysis is tedious by hand and is exactly what ",[147,1862,1863],{"href":1395},"Trendia does automatically",", both on your content and on your category.",[50,1866,1868],{"id":1867},"a-note-on-copying","A note on copying",[43,1870,1871,1872,1875],{},"Identifying that a hook works does not mean reproducing it word for word. What transfers is ",[61,1873,1874],{},"the pattern",": the structure and the emotion it activates. The execution has to be yours, or the viewer notices the graft.",[43,1877,1878],{},"Copying a viral is not a strategy. Understanding why it worked and applying it to what you can actually tell, is.",[43,1880,1881,1882,272],{},"If you want to see which openings are working in your specific category, ",[147,1883,1884],{"href":1395},"try it with your accounts for seven days",{"title":20,"searchDepth":21,"depth":21,"links":1886},[1887,1895,1896,1897],{"id":1714,"depth":21,"text":1715,"children":1888},[1889,1890,1891,1892,1893,1894],{"id":1718,"depth":1486,"text":1719},{"id":1736,"depth":1486,"text":1737},{"id":1751,"depth":1486,"text":1752},{"id":1766,"depth":1486,"text":1767},{"id":1781,"depth":1486,"text":1782},{"id":1796,"depth":1486,"text":1797},{"id":1811,"depth":21,"text":1812},{"id":1835,"depth":21,"text":1836},{"id":1867,"depth":21,"text":1868},"Content","2026-08-06","The first five seconds decide whether anyone watches the rest. Six hook families, when to use each, and how to measure if they actually work.","\u002Fblog\u002Ftypes-of-hooks-for-reels.png",{},"\u002Frecursos\u002Ftypes-of-hooks-for-reels",{"title":1697,"description":1900},"Six hook families for Instagram Reels, when to use each, and how to measure if they work with data instead of instinct.","Types of hooks for Reels and how to measure them | Trendia","recursos\u002Ftypes-of-hooks-for-reels","Six hook families and the metric that tells you if yours is working.",[1910,1911,1898],"Hooks","Reels","dqtxIs4yI4ksmkEUk6MSPcHxGbOQDSNFvuobSnuWgzw",{"left":4,"top":4,"width":5,"height":5,"rotate":4,"vFlip":6,"hFlip":6,"body":1914},"\u003Cg fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\">\u003Cpath d=\"M10 12h4m-4-4h4m0 13v-3a2 2 0 0 0-4 0v3\"\u002F>\u003Cpath d=\"M6 10H4a2 2 0 0 0-2 2v7a2 2 0 0 0 2 2h16a2 2 0 0 0 2-2V9a2 2 0 0 0-2-2h-2\"\u002F>\u003Cpath d=\"M6 21V5a2 2 0 0 1 2-2h8a2 2 0 0 1 2 2v16\"\u002F>\u003C\u002Fg>",{"left":4,"top":4,"width":5,"height":5,"rotate":4,"vFlip":6,"hFlip":6,"body":1916},"\u003Cg fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\">\u003Cpath d=\"M21.42 10.922a1 1 0 0 0-.019-1.838L12.83 5.18a2 2 0 0 0-1.66 0L2.6 9.08a1 1 0 0 0 0 1.832l8.57 3.908a2 2 0 0 0 1.66 0zM22 10v6\"\u002F>\u003Cpath d=\"M6 12.5V16a6 3 0 0 0 12 0v-3.5\"\u002F>\u003C\u002Fg>",{"left":4,"top":4,"width":5,"height":5,"rotate":4,"vFlip":6,"hFlip":6,"body":1918},"\u003Cpath fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\" d=\"M10 18v-7m1.119-8.795a2 2 0 0 1 1.762 0l7.84 3.846A.5.5 0 0 1 20.5 7h-17a.5.5 0 0 1-.22-.949zM14 18v-7m4 7v-7M3 22h18M6 18v-7\"\u002F>",{"left":4,"top":4,"width":5,"height":5,"rotate":4,"vFlip":6,"hFlip":6,"body":1920},"\u003Cpath fill=\"none\" stroke=\"currentColor\" stroke-linecap=\"round\" stroke-linejoin=\"round\" stroke-width=\"2\" d=\"M7 7h10v10M7 17L17 7\"\u002F>",1791902722248]