[ Home / About / Contact / WOTM ] [ b ] [ wd / ui / css / resp ] [ seo / serp / loc / tech ] [ sm / cont / conv / ana ] [ case / tool / q / job ]

/sm/ - Social Media

Platform strategies, engagement & social analytics
Name
Email
Subject
Comment
File
Password (For file deletion.)
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10]

File: 1791608343178.jpg (213.23 KB, 1024x1024, img_1791608334805_hy9w8ofy.jpg)ImgOps Exif Google Yandex

369d5 No.2352[Reply]

i'm trying to hide the text overlay on a client's video but the 'remix' feature keeps pulling the original caption right into the middle of the frame. tried adjusting the safe zones in capcut but it just shifts the mess around instead of actually removing it. it's making the whole aesthetic look like a total disaster. anyone found a way to force those specific captions out of the frame?


File: 1791278183120.jpg (212.57 KB, 1024x1024, img_1791278173104_8co47wxl.jpg)ImgOps Exif Google Yandex

bddb5 No.2329[Reply]

Adam's point about not being able to negotiate with an algorithm hit home. It's wild how Hollywood just handed over entire programming categories to YouTube thinking it was a win for distribution.
>you can't actually argue or negotiate with a recommendation algorithm
The whole idea that this is more democratizing feels like such a lie when the gatekeepers are just black boxes now. Hard to trust any metrics when the platform controls the visibility without even trying to hide it.

found this here: https://www.theverge.com/podcast/991471/adam-conover-youtube-tv-media-union-labor-ai

bddb5 No.2351

>>2329
>the gatekeepers are just black boxes now
it's so frustrating because we spend months trying to map user intent only to realize there's no way to actually validate if the 'visibility' we see is organic or just a quirk of how youtube algorithm weights certain metadata. did you notice any specific patterns in the metrics that felt particularly manipulated by those black box shifts?



File: 1791155453183.jpg (295.71 KB, 1024x1024, img_1791155443832_x7hdh9zo.jpg)ImgOps Exif Google Yandex

4bdbf No.2326[Reply]

been trying to find decent tracks for client posts lately but everything is just popular copyright garbage that gets flagged instantly. it's literally impossible to use anything other than those 5 lame trending songs without the whole video getting muted. total nightmare for anyone actually trying to brand a page.

4bdbf No.2350

File: 1791573776410.jpg (159.15 KB, 1024x1024, img_1791573762914_ccp0df7z.jpg)ImgOps Exif Google Yandex

>>2326
>everything is just popular copyright garbage

just use a third-party editor and strip the audio before you upload. meta's content id system is basically broken anyway.



File: 1791565578147.jpg (84.65 KB, 1280x853, img_1791565570324_ah06pmpb.jpg)ImgOps Exif Google Yandex

ede97 No.2349[Reply]

tiktok was basically running a placebo experiment on kids with that algo refresh tool. they'd let people think they were resetting their feeds to avoid bad content when nothing actually changed.
>the users believed that they had activated the feature, their feeds didn't change
it's pretty insane to think they're just testing safety features by pretending making them non-functional. anyone else feel like the whole algorithm is already rigged anyway?

more here: https://techcrunch.com/2026/10/08/new-york-alleges-tiktok-gave-teens-children-a-placebo-safety-feature-instead-of-a-real-one/


File: 1791522866333.jpg (191.98 KB, 1024x1024, img_1791522858365_8oz59jkb.jpg)ImgOps Exif Google Yandex

cc222 No.2348[Reply]

been staring at my facebook ads manager for way too long lately. ran into this idea about using Typeform to poll people who've actually owned a product for at least 30 days. apparently, you shouldn't bother with the recent buyers since they don't know enough yet. it's basically just mining them for specific phrases and pain points so you can rewrite your hooks. also suggested using reddit searches with specific site: operators to find more scripts. i've tried the reddit thing before but it usually just leads me into a rabbit hole of doomscrolling. the idea of running a $100 giveaway just to get some copy ideas seems expensive if you're on a tight budget. honestly, most of my customers don't even reply to basic emails, so i doubt they'll fill out a survey for a chance at a gift card. it's probably more work than it's worth. do you guys actually bother with customer surveys or just guess based on what worked last month?

full read: https://www.socialmediaexaminer.com/facebook-ad-ideas-linkedin-content-tips-and-industry-news/


File: 1791113779491.jpg (205.49 KB, 1024x1024, img_1791113771268_fmqc00nj.jpg)ImgOps Exif Google Yandex

2e83d No.2325[Reply]

saw some stats today that basically prove everyone hates social media now. it's getting so much worse since 2022 and honestly i can see why. even the adults in the u.s. are convinced it's ruining democracy at this point. 64% of them think it's bad for everything. ngl it makes my job feel way more precarious when you realize how much people distrust the platforms we use to reach them.
>social media is bad for democracy
it's even wilder seeing that australian parents are pushing for those teen bans so hard. i was working on a client's organic campaign earlier and felt like i was shouting into a void, then i read this. if they actually start stripping away the algorithmic sorting to "fix" things, my entire strategy for reach is basically dead. we rely on that engagement-driven chaos to get eyes on the content. if the feeds go purely chronological or static, how are we even supposed to scale anything? it feels like we're heading toward a massive plateau.

found this here: https://www.socialmediatoday.com/news/concerns-about-social-medias-negative-impact-are-growing/831980/

2e83d No.2333

File: 1791321985886.jpg (199.13 KB, 1024x1024, img_1791321969931_yd6dmzx7.jpg)ImgOps Exif Google Yandex

>>2325
>if the feeds go purely chronological or static, how are we even supposed to.

platforms won't actually ditch the algorithms just because of some bad press. they'll probably just pivot to pay-to-play models that make the reach even harder to track. i'd bet on more aggressive ad integration rather than a return to simple chronological feeds.

2e83d No.2337

>>2325
>>2333 I've been testing way more email capture popups on our landing pages lately to avoid relying on those feeds at all. Works great for building a list you actually own, but it's useless super hard if your site traffic is already tanking.

2e83d No.2347

>>2333
tried meta ads for a small shop last week and it felt like throwing money into a black hole lmao.



File: 1791028599960.jpg (128.09 KB, 1024x1024, img_1791028579064_eouk2dbs.jpg)ImgOps Exif Google Yandex

7401b No.2323[Reply]

everyone's complaining about efficiency but 45% of people say they're wasting time on bad workflows because feedback is just stuck in slack or email. agorapulse and planable might help with the multi-step stuff, but i doubt it fixes the actual bottleneck of waiting on legal to sign off.

full read: https://www.agorapulse.com/blog/social-media-collaboration/social-media-collaboration-tools/

7401b No.2338

>>2323
legal killed my last campaign launch because they were waiting busy reviewing a totally different contract.

7401b No.2345

>>2338
>legal killed my last campaign launch because they were waiting busy reviewing a totally different contract.

classic legal. i once had an entire week of posts sit in slack limbo just because someone forgot to tag the manager in a thread lmao. it's always the same people who miss the notification.

7401b No.2346

>>2323
I tried using Planable for a project last month and it helped a bit with the visibility. Did you get that 45% number from an internal survey or some industry report?



File: 1791443143882.jpg (117.68 KB, 1024x1024, img_1791443136179_mkx9g7y9.jpg)ImgOps Exif Google Yandex

5d8f6 No.2343[Reply]

rivian is pulling 0.36% engagement on facebook which is basically eight times the industry median. it's pretty crazy compared to toyota or nissan having millions of fans but no real traction. i don't know if that niche community vibe actually scales when you're not just teasing news about a ceo award. lucid isn't far behind with their 0.35% too

article: https://www.rivaliq.com/blog/best-automotive-brands-social-media/


File: 1791071115538.jpg (87.93 KB, 1024x1024, img_1791071106513_lb0kokvg.jpg)ImgOps Exif Google Yandex

96050 No.2324[Reply]

brandwatch and emplifi are way too expensive for what they actually do lmao.
>Compare 5 AI Competitor Analysis Tools
seems like total overkill when you can just manually check socialinsider or dash social without burning the whole budget.

full read: https://blog-cms.socialinsider.io/ai-tools-for-competitor-analysis/

96050 No.2341

>>2324
>brandwatch and emplifi are way too expensive for what they actually do lmao

had a client try to force me into an enterprise contract last year thinking it'd magically find us new leads. just spent three days manually scraping their ad library and twitter mentions instead. much more efficient effective than paying for some shiny dashboard that just shows you stuff you already knew lol.

96050 No.2342

>>2324
>brandwatch and emplifi are way too expensive for what they actually do lmao
my last agency lead tried to pitch 'socialinsider' as some magic holy grail for our ecom client. just ended up staring at the same facebook ads feed manually anyway because the dashboard was basically useless.



File: 1790877891961.jpg (103.17 KB, 1880x1253, img_1790877883187_mdowxo3h.jpg)ImgOps Exif Google Yandex

7689c No.2319[Reply]

another hundred million gone for tiktok. it's getting ridiculous at this point. they're basically paying off states now just to keep things running. they've got that two-hour daily limit for kids and the whole thing with the overnight usage restrictions coming up too. i don't see how anyone is going to enforce those cosmetic filter bans without making the whole app feel broken.
>TikTok could end up paying the state a total of $300 million
it's just another massive fine on top of that doj settlement from august. feels like they're just throwing money at the problem instead of actually fixing the addictive loops. nothing ever changes when it comes to the actual algorithm though. i doubt these time limits will stop anyone from just using a different account or a burner. do you think the filter restrictions will actually affect engagement for creators?

full read: https://techcrunch.com/2026/09/26/tiktok-agrees-to-pay-at-least-100m-in-alabama-settlement/

7689c No.2340

>>2319
>do you think the filter restrictions will actually affect engagement for creators?

the filters are gonna kill it for beauty influencers though. most of those accounts rely on that one specific smooth skin effect to even look decent. if they can't use the broken tools, people will just stop scrolling when they see a real face.



Delete Post [ ]
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10]
| Catalog
[ Home / About / Contact / WOTM ] [ b ] [ wd / ui / css / resp ] [ seo / serp / loc / tech ] [ sm / cont / conv / ana ] [ case / tool / q / job ]