A few feeds our readers keep coming back to:

— @SitemapHustle — In-the-trenches site architecture tactics from someone who's actually…
— @AlgoHeretic — Most TikTok 'growth advice' is recycled nonsense. We test the sacred…
— @WPFromScratch — WordPress explained plainly, one step at a time. No jargon you don't…
— @pop_wire — The inside wire on pop and redirect traffic: source shakeups, fresh…

Feed posts decay in 48 hours; the half-life problem nobody budgets for

The question: how long does a B2B social post keep delivering reach, and how does that compare to owned channels like newsletters?

The dataset: decay analyses tracking the share of a post's lifetime impressions delivered over time, across LinkedIn and X, compared with email-open curves.

Three findings:
— Most social posts delivered the bulk of their impressions within ~24–48 hours, then flatlined — a short half-life.
— LinkedIn posts showed a longer tail than X (where the half-life is famously minutes), but still mostly spent within two days.
— Newsletter content, by contrast, reaches a near-fixed list on send and can be re-surfaced, giving owned channels a more predictable, durable curve.

Caveats: decay curves vary with audience size and how long comments keep circulating a post; 'lifetime impressions' is platform-reported and opaque. Treat as directional.

What it means for B2B: social reach is rented and perishable — you re-pay attention costs every cycle. Treating a viral post as a durable asset over-values it. The strategic counterweight is converting rented social attention into owned, durable channels (email, community) you don't have to re-win daily.

Bottom line: social posts are fireworks; build something that stays lit.

В Telegram Ads добавили Banner in Bot

В Telegram Ads появился формат Banner in Bot для показа рекламы внутри ботов с аудиторией от 1000 человек. Инструмент таргетируется не на площадки, а на пользователей — по интересам, номерам телефонов и подпискам. Пока доступны только текстовые креативы по ставкам Target Users. Новинка позволяет напрямую через стандартный кабинет охватывать целевую аудиторию прямо в их диалогах с ботами.

➡️ Читайте на сайте: https://aff.top/blog/v-telegram-ads-dobavili-banner-in-bot

🧠 Ещё больше инсайтов → в канале AFF.top

Anthropic готовит к запуску Claude Money

Anthropic анонсировала Claude Money — ИИ-сервис для управления личными финансами и автоматизации платежей. Инструмент позволяет подключать банковские карты для анализа расходов, планирования бюджета и проведения транзакций. Это переход от консультационных моделей к полноценным финансовым агентам. Внедрение технологии позволит пользователям делегировать нейросети рутинные задачи, включая оплату токенов и подписок на необходимые рабочие сервисы.

➡️ Читайте на сайте: https://aff.top/blog/anthropic-gotovit-k-zapusku-claude-money

🧠 Ещё больше инсайтов → в канале AFF.top

Meta ограничивает расходы на токены для сотрудников

Meta ввела внутренние лимиты на использование ИИ из-за резкого роста расходов: в 2026 году только на сотрудников заложены миллиарды долларов, а общий бюджет на ИИ-инфраструктуру оценивается в 130–145 млрд. Вывод простой: даже у Big Tech ИИ перестал быть бесплатной игрушкой и требует жёсткого контроля затрат.

➡️ Читайте на сайте: https://aff.top/blog/meta-ogranichivaet-raskhody-na-tokeny-dlia-sotrudnikov

🧠 Ещё больше инсайтов → в канале AFF.top

Myth: there's a universal best time to post on LinkedIn

The question: do the famous 'post at 9am Tuesday' charts predict your reach, or are they an artifact of how the data was pooled?

The advice comes from aggregate studies (Sprout, Hootsuite, Buffer) summing millions of posts across all industries into one heatmap. The method is sound for describing platform-wide averages — but averaging across audiences in different time zones, roles, and posting habits produces a curve that may match almost nobody.

Three findings:
— Early-velocity (engagement in the first 60-90 minutes) matters more than the clock; the algorithm reads early signal, so 'when your audience is awake and idle' beats 'when the global chart peaks'.
— B2B audiences skew to working hours, but specific niches (developers, founders) cluster at off-hours the charts wash out.
— Self-tested accounts often find their personal best slot is 1-2 hours off the published 'optimal'.

Caveats: confounded by the fact that everyone posts at the 'recommended' times, raising competition then. Directional only.

Bottom line: the heatmap is a starting hypothesis, not an answer. Run a four-week split and trust your own first-hour velocity.

An author rewrote only the first two lines of 30 posts. Reach rose 1.7x.

The question: how much of a post's reach is decided by the opening lines that show before 'see more'?

The case: a B2B writer ran a clean test — same body, same topic, two openings. Version A led with context; version B led with a concrete claim or number in the first 8–10 words (the visible 'hook'). Posted as staggered pairs. Sample: 30 pairs over 8 weeks.

Three findings:
— Number-led hooks averaged 1.7x the reach of context-led ones.
— The mechanism was the 'see more' expand rate — hook posts got more expands, which feeds dwell time, which feeds reach.
— Hooks that posed a question underperformed hooks that stated a surprising fact.

Limitations: 'surprising' is subjective and the author wrote both, so an unconscious quality bias toward B is possible. Small sample.

What it means for B2B: the first line is a distribution decision, not a copywriting flourish — it gates the expand that gates everything downstream.

Bottom line: front-loading a concrete number plausibly drove the reach gap, via the expand-then-dwell chain.

Claude Cowork, Claude Design объединили в один Claude

➡️ Читайте на сайте: https://aff.top/blog/claude-cowork-claude-design-obedinili-v-odin-claude

🧠 Ещё больше инсайтов → в канале AFF.top

Microsoft планирует вставлять рекламу в игры

➡️ Читайте на сайте: https://aff.top/blog/microsoft-planiruet-vstavliat-reklamu-v-igry

🧠 Ещё больше инсайтов → в канале AFF.top

Myth: every comment is worth more than a like

The question: is the 'comments > likes' hierarchy a real ranking weight, or a slogan that ignores what kind of comment?

The advice rests on a true premise — dwell time (how long a reader's screen lingers on your post) and reply threads correlate with reach — but it flattens all comments into one bucket. Algorithm-report analyses suggest the model appears to value comment depth and recency, not raw count.

Three findings:
— A two-word 'Great post!' likely registers near a like in value; a three-sentence reply that holds the reader is a different signal entirely.
— Author replies extend the thread's active window, and a comment that triggers more comments compounds; isolated comments don't.
— Engagement-pod comments (off-topic, instant, reciprocal) can be detected as low-quality and may dampen rather than help.

Caveats: we infer weights from outcomes, never see them. Correlation, not causation — good posts attract good comments regardless of mechanics. Directional.

Bottom line: chase substantive replies you can build on, not comment count. Asking a real question beats begging for 'thoughts?'.

Native posting vs. third-party schedulers: the reach tax

The question: does publishing through a scheduling tool (Buffer, Hootsuite, Sprout) actually suppress LinkedIn reach versus posting by hand?

The folklore says yes. The cleaner evidence says: it depends on the integration type. LinkedIn's public API publishes through the same endpoint as native, so a true API post is not algorithmically penalized. The suppression people observe usually traces to two confounds — schedulers historically forced link-in-post (which does suppress reach), and they post at fixed clock times that miss your audience's actual active window.

Three distinctions worth holding:
— API-published posts (modern tools) ≠ screenshot-bot reposting. Only the latter looks inorganic.
— Schedulers can't reply in-thread fast, and early-comment velocity drives distribution.
— Manual posting lets you read the room and skip a bad-news day.

Caveat: most 'scheduler kills reach' claims are anecdotal A/B tests on tiny samples — treat as directional, not proven. Correlation (tool used) is not causation (tool blamed).

For B2B: schedule the publish, but log in to seed and answer the first 30 minutes manually.

Bottom line: the tool isn't the tax — link placement and dead first-hour are.

Engagement rate vs. saves and shares: the metric you optimize is the behavior you get

The question: which signal should a B2B team actually optimize — headline engagement rate, or saves and shares specifically?

They're not interchangeable. Engagement rate lumps likes (cheap, low-intent) with comments and saves (expensive, high-intent). Optimizing the blended number pushes you toward like-bait. Saves and shares are the signals that correlate with the content being genuinely useful and with downstream business behavior.

Three distinctions:
— A save is a private bookmark — the viewer is admitting they'll need this later; that's intent, not applause.
— A share is a reputation stake — someone vouching for you to their network; it carries social proof a like never will.
— Posts engineered for likes (provocation, relatability) often have near-zero saves — entertaining, not valuable.

Caveat: saves and shares are sparser, so they're noisier on small samples — average across many posts before drawing conclusions.

For B2B: track save-rate and share-rate as separate columns, not folded into engagement rate. Ask of every draft: would someone bookmark or forward this?

Bottom line: likes measure reach, saves and shares measure value — optimize the ones that predict revenue.

Myth: video always outperforms on LinkedIn

The question: is 'video gets more reach' a format law, or platform incentives we mistook for audience preference?

The claim comes partly from LinkedIn's own promotion of video (they surface what they're investing in) and partly from format studies that don't control for who posts video — often larger, more resourced accounts. That's selection bias: the format and the creator are confounded.

Three findings:
— Text and document (carousel) posts frequently match or beat video on saves and dwell time for analytical B2B content; people skim a carousel faster than they commit to a video.
— Native video reach spiked while LinkedIn pushed it, then normalized — a platform-cycle effect, not a permanent edge.
— Completion rate matters more than plays; a 3-minute video few finish can underperform a tight 1,200-character text post.

Caveats: format performance shifts with each platform push; small samples per niche. Treat as directional.

Bottom line: there is no universally winning format — only formats matched to intent. Test carousel vs video on your own audience before believing the law.

A 5-step playbook to audit your own dwell time

The question: which of your posts actually hold attention, versus just collect a reflexive scroll-past like?

Dwell time — the seconds a reader spends on a post before scrolling on — is a ranking input LinkedIn confirmed weighting in 2017 and never walked back. You can reverse-engineer it without their dashboard.

The playbook, run over your last 30 posts:
— Pull each post's impressions and members reached into a sheet.
— Tag format: text-only, document, image, video, poll.
— Compute (reactions + comments) / impressions as a crude engagement floor.
— Flag posts where reach exceeds impressions by 1.4x or more — that gap signals re-surfacing, a dwell proxy.
— Sort by format. The format clustering at the top is where your attention lives.

Limitations: this is inference, not the real metric — LinkedIn doesn't expose per-post dwell. Small sample, treat as directional. Re-surfacing also reflects comment recency, not only dwell.

What it means for B2B: long-form documents and 'see more' text posts consistently top these audits because they force a tap. Build the next month around your two winning formats.

Bottom line: you can't optimize a metric you never estimate — this gives you a usable proxy in an afternoon.

Myth: stacking hashtags expands your reach

The question: do hashtags still function as discovery channels on LinkedIn, or are we copying a 2019 Instagram habit?

The advice ('use 3-5 relevant hashtags') dates to when LinkedIn promoted hashtag-follow feeds prominently. Since then the platform quietly de-emphasized hashtag discovery; the feed is increasingly interest-graph and connection driven. Practitioner tests adding vs removing hashtags show negligible reach difference in most samples.

Three findings:
— Hashtags now act mostly as weak topical labels for the algorithm, not as traffic sources from hashtag pages.
— Over-tagging can read as spammy and adds visual noise that may shave dwell time.
— One or two precise tags occasionally help categorization in narrow niches; ten broad ones do nothing measurable.

Caveats: hard to isolate hashtag effect from content quality; platform behavior changes silently and without announcement. Directional, small samples.

Bottom line: hashtags are a label, not a megaphone. Keep zero to two relevant ones and stop expecting discovery from them — your hook and first-hour engagement do the heavy lifting.

На Anthropic, OpenAI, Google и SpaceXAI подали в суд из-за ИИ

В Калифорнии против ИИ-компаний подали антимонопольный иск: регулятору показалось подозрительным, что игроки синхронно призывают ограничить развитие нейросетей ради безопасности. Смысл спора в том, что инвестиции в ИИ уже обгоняют реальный прогресс, а бизнесу выгодны правила, которые защитят капитал. Вывод: быстрых прорывов ждать не стоит, лучше выжимать максимум из текущих инструментов.

➡️ Читайте на сайте: https://aff.top/blog/na-anthropic-openai-google-i-spacexai-podali-v-sud-iz-za-ii

🧠 Ещё больше инсайтов → в канале AFF.top