Adam builds Acumin. He spends his days on the same two problems this library is about: working out what a piece of content is actually worth, and getting a brief through production without it turning into something else.
Everything published here is held to one standard: no invented statistics, no fabricated case studies, and every claim about how a platform behaves links to that platform's own documentation.
A number is only good or bad against a baseline. How to benchmark your content against your category, not just your own past.
Collecting, tagging and actually reusing references, made a weekly habit. Plus the honest line between a swipe file and copying.
How to match objective, audience and budget to the right ad platform, where each channel wins, and the honest case for spending nothing on ads.
Build a small brand palette, use one accent sparingly, hit a contrast ratio people can actually read, and make it all work on dark and light backgrounds.
One positioning and pillar set feeding every channel, the make-once-cut-many model, and how six channels play different roles in a single plan.
Hero/hub/help, pillar-and-cluster, campaign-led, founder-led, product-education and community-led: what each is, when it fits, and how to set it up.
The six principles that separate a professional post from an amateur one: hierarchy, contrast, alignment, proximity, repetition and whitespace.
The honest basics of ranking your own site and blog - search intent, on-page relevance, links, and the real limits of SEO with no snake oil.
How Google and YouTube ads work: search intent versus interruptive social, the main campaign types, targeting basics, and where each one fits.
You've posted to a Facebook Page for a month and organic reach is tiny. Here are the levers that still work, and when a paid boost is the honest answer.
Your Instagram posts land with followers but reach has stalled. The levers that actually grow an account, and how to break the flat-reach plateau.
You've posted on LinkedIn for a month and reach has flattened. Here are the levers that actually move a professional audience, and the plateau to break.
You are posting on TikTok but most videos land flat. The levers that actually move reach, and how to break the plateau when volume alone stops working.
You've posted on X for a month and it feels like shouting into a void. Here are the levers that actually grow an account, and the plateau to break.
You have posted for a while and views are flat. The two levers that actually move a YouTube channel, and how to break the upload-more-stay-flat plateau.
Match audience, format strength and real capacity to channels, why fewer done well beats all six done badly, and how to add one at a time.
Move past demographics to jobs-to-be-done and real watch behaviour, read your own audience from Snapshot, and let that choice drive format and channel.
Core, custom, lookalike and broad audiences for paid campaigns, when each fits, the consent and matching caveats, and how to sense-check size and overlap.
What paid advertising actually buys, when it is worth it and when it is not, the objective ladder, and a sensible minimum budget you can work out yourself.
What to post first on Facebook, how a Page differs from a Group and a personal profile, and a first-month cadence a small brand can actually keep.
A brand's first month on Instagram: what the channel is genuinely good for, which formats to start with, a realistic cadence, and the trap to avoid.
What to post first on LinkedIn, how a company Page differs from a personal profile, and a first-month cadence a small team can actually keep.
A brand's first month on TikTok: what the channel is genuinely good for, which formats to start with, a realistic cadence, and the trap to avoid.
What to post first on X, how threads and replies actually distribute a new account, and a first-month cadence a small brand can keep without burning out.
A brand's first month on YouTube: what the channel is genuinely good for, the formats to start with, a realistic cadence, and the trap to avoid.
Where to put things so a post reads at a glance: grids, one focal point, the rule of thirds, visual weight, and the safe zones each format hides.
Follower count and revenue report how you did, too late to change it. How to pick a leading metric you can act on this week, per content goal.
How LinkedIn Campaign Manager works: B2B objectives, saved and matched audiences, ad formats, the honest cost floor, and when the price is worth it.
The anatomy of a YouTube video that keeps people watching: cold open, promise, delivery, pattern interrupts and a payoff that earns the end screen.
Why four ad dashboards never add up, how post-ATT signal loss and self-reported ROAS mislead, and a cross-platform scorecard built on cost per outcome.
How to build saved, custom and lookalike audiences and use Advantage+ audience inside Meta Ads Manager, with the consent and matching caveats that matter.
How Meta spends your ad budget: daily vs lifetime, campaign budget optimisation, bid strategies, and the learning phase you keep resetting.
Ad formats, where they show up, and why one creative rarely fits every placement: a concept guide with specs deferred to Meta's own sheet.
What the Meta pixel and dataset do, standard versus custom events, why the Conversions API exists after ATT signal loss, and how attribution windows work.
Which columns matter by objective, how breakdowns and attribution can mislead, and a plain rule for deciding to scale, hold or kill a campaign.
The three levels of Meta Ads Manager (campaign, ad set and ad), what each one controls, and how the objective you pick cascades through all of them.
Prospecting versus retargeting, full-funnel structure, always-on versus burst, clean creative testing and how to split budget across the funnel.
The evergreen visual-search channel most brands skip. Why pins keep earning for months, and how to tell whether your category belongs there.
The statistical and cognitive traps that make content data lie to you: small samples, survivorship, seasonality, cherry-picked windows and false cause.
An end-to-end walkthrough of a first Meta campaign, from the accounts you need through campaign, ad set and ad, to a modest test and what to watch in week one.
A short lives or dies on watch-through. The structure that keeps people watching: the first two seconds, the promise, retention beats and the loop.
Low-polish, high-frequency posting: the daily habit, poll and question stickers, and why stories mostly deepen the following you already have.
The swipeable post is its own format. Slide one as the hook, one idea per slide, designing for the swipe, and the last-slide call to action.
A per-goal map of what reach, views, watch-through, engagement, saves, clicks and follows each tell you, and which two or three to actually watch.
The craft of a text post on LinkedIn and X: the scroll-stopping first line, whitespace, one idea, and how the two platforms really differ.
How TikTok Ads Manager works: account structure, audiences, Spark Ads, the Creative Center, fast creative fatigue, and how it differs from Meta.
Short-form feeds now double as search engines. How keywords in captions, on-screen text and speech get you found, and how search differs from the feed.
Setting type that reads at feed size: one display face, one body, a clear hierarchy, sensible spacing, and when text-on-image beats a caption.
A trend without a point of view is just noise you paid to make. How to turn a trend into an idea that fits your pillars, and when to sit one out.
The caption as its own format: the hook line, the value, the call to action, when long beats short, and formatting for the feed's "more" cut-off.
The words on a post are design too: one message, a headline and support line, verbs over adjectives, brevity, and one clear call to action.
How videos get found on YouTube - search versus browse, titles and descriptions and tags, the thumbnail's job, and why watch-through still decides it.
Most creative-industry reviews can be gamed by anyone with an email address. What changes when a rating requires a real, delivered booking, and what doesn't.
Why the old even-thirds funnel split stopped making sense, what a modern default looks like, and how to know when yours is wrong.
The reusable half of every production: contacts, locations, props, palette. What belongs in it, and why the maintenance rule matters more than the structure.
Most content strategies fail on execution, not selection. How to work out which formats your team can make repeatably well, and drop the rest without guilt.
The nine clauses that prevent almost every freelance dispute, why a deposit is a commitment test, and what to do when a client won't sign anything.
You can't compare the numbers across platforms. You can compare the choices, the patterns and the disagreements, which is where the useful findings are anyway.
Three adjectives on a slide isn't a voice document. What to write instead so a freelancer, a new hire or a model can sound like you on the first attempt.
Why the rate beats the total, what each platform counts as an engagement, and the three situations where engagement-per-view will actively mislead you.
A new category is one where none of your instincts are calibrated. What to study in the first two weeks, in what order, and the mistakes that give you away.
An empty space in your category is either an opportunity or a graveyard. The four reasons a gap exists, and how to tell which one you're looking at.
What an invoice must contain, when to send it, and the small formatting choices that get you paid weeks earlier without a single chasing email.
The showreel tells you what they've done, not what they'd do for you. Where the real signal is in a pitch, and the four flags worth walking away from.
What the paid creator plan actually changes, the arithmetic that decides it, and the honest case for staying on the free plan.
Whose channel it ran on decides what you can know. The four measurement situations, and the attribution question nobody can answer cleanly.
The cheapest client to win is one you already have. What to do during and after a project so the second booking is a formality rather than a pitch.
The organic post proved the creative. It didn't prove the targeting, the offer or the destination. What transfers, what doesn't, and how to structure the build.
By the final cut you've lost the ability to see it. A structured last pass that catches the things familiarity hides, in about forty minutes.
A shoot day is the most expensive place to discover an idea was thin. Seven questions that cost twenty minutes and kill the ideas that deserve killing.
Most delays on a shoot trace back to something the brand never sent. The eleven things to hand over, and four questions worth answering unasked.
You can't see their analytics. You can see everything they chose to do. Choices under constraint reveal more than numbers do. A six-pass method.
Ad platforms report on their own performance. The seven ways that inflates what you see, and how to read the numbers as evidence rather than as a verdict.
A score with no band is a guess in a suit. What the band is actually measuring, how wide is too wide, and the four things a narrow band still can't tell you.
A rating is only worth what you can show. What a referral card contains, where the performance figure comes from, and how to use one outside Acumin.
Most freelance reputation evaporates when a contact changes jobs. How to build proof that belongs to you, travels with you, and gets stronger with volume.
Two rounds is usually right. Why projects need five, how to consolidate feedback so an editor can act on it, and what to do when a late stakeholder appears.
A view on TikTok and a view on YouTube are different events. Why blending them produces a confident number that measures nothing, and what to compare instead.
"Grow the brand" can't be measured, so it can't be wrong. How to write a goal specific enough that your content data can tell you that you failed.
They answer different questions and get confused constantly. What each one is for, when neither is worth making, and the cheaper thing that usually is.
Most hook tests can't produce a conclusion. How to structure one that can: same body, isolated variable, and enough volume to read the result.
What you're actually buying when you hire a creator, why quotes aren't comparable until you normalise them, and where the cheap option's money reappears.
A brand watching your reel is asking one question, and it isn't whether you're talented. How to cut for the person deciding, not for other filmmakers.
A signal is an observation that survives being checked. The three tests it has to pass, and why most of what gets called a signal is a coincidence with good PR.
A rival's piece does ten times their normal numbers. Five questions that separate a repeatable move from a fluke, and why copying it fast is the worst answer.
Dashboards report on what you already made. The gap between describing the past and deciding the future, and the four things that actually close it.
The approach section is where most pitches are won and lost. What to put in it, what to cut, and why the camera you shoot on belongs near the bottom.
The returns are upstream. Where generative tools genuinely compress work, where they quietly cost you, and the line between the two.
Audiences have started discounting anything that looks generated, including work that isn't. What triggers the discount, and what reliably survives it.
What goes on a call sheet, why each line is there, and the one failure that ruins an otherwise perfect one: nobody having it.
A one-off teardown is out of date in a quarter. How to pick a set worth watching continuously, and what should actually trigger your attention.
A small team can now out-produce an agency. What that actually requires, what to buy versus build, and the failure mode that comes with the leverage.
Thirteen block types, a default order that encodes the argument, and the one part the AI is allowed to touch. Plus what it will never invent.
Most comments need nothing. A few need you personally, and one kind is worth amplifying. How to tell them apart without reading everything.
Your audience keeps telling you what to make next, in public, for free. How to harvest it without mistaking one loud comment for a pattern.
Which connection changes what Acumin can see, exactly which permissions each asks for, and the honest limits of the ones that look impressive.
Which fields brands actually filter on, why three specialities beat six, and the two things that decide a shortlist before anyone watches your reel.
A naming convention, the right formats per destination, and an expiry you state up front. The unglamorous half hour that prevents months of "can you resend".
Versioned cuts, timecoded comments and a one-shot verdict. How the room works from the creator's side, and what it settles that email can't.
A pitch costs real hours. Which briefs deserve them, which to skip without guilt, and why the pitch cap changes the maths.
What to do first, in order, and what each step actually gets you, from the free Snapshot to a Signal Brief you can hand to someone.
Two follow-ups, each carrying something new, then release the date. The pattern that converts without costing you the relationship.
Sequence a quarter by dependency rather than by date, size it against real capacity, and write down what you deliberately left out.
It's read-only, it's grounded in your brand, and it won't invent a number about you. How to ask in a way that uses all three.
Most late payment is caused before the invoice exists. Six structural decisions that move money sooner, none of which involve chasing harder.
Notes against the brief, at the right stage, from one voice. The three rules that turn a review round from an argument into a version.
The three old routes to work are all narrowing at once. What replaced them is searchable, verifiable proof, which rewards a different set of behaviours.
A hook isn't a slogan. It's the first structural move of a piece, and it makes a promise the rest has to keep. The five that work and why.
Brief first, shortlist second, price last. How to run a hire that gets you the film you wanted instead of the cheapest version of something adjacent.
The treatment that wins isn't the cheapest or the prettiest. It's the one that proves you understood the brief. Here's the structure, section by section.
Build a rate from the income you need and the days you can actually bill, then price usage separately instead of guessing a number and hoping.
A repeatable method for reading a competitor's strategy from public data alone: what to collect, how to compare it honestly, and what it can't tell you.
A brief's job is to make one decision impossible to get wrong. Here are the nine sections that matter, and the three that quietly ruin most briefs.
Not everywhere. That's just slow. Four places where removing the human is a category error, and how to tell them from the rest.
Numbered invoices for bookings and for off-platform clients, a shareable link, and one thing it deliberately doesn't do: move your money.
Four sections, each with a next action attached. What the digest is choosing to show you, and the one number worth reading first.
Almost every unhappy client is a surprised client. Six moments where a sentence prevents a problem, and what to say at each.
The difference between a number a system counted and a number someone told you, and why almost every content report quietly mixes the two together.
A free, public record of what your competitors are paying to say. What it can prove, what it can't, and how to turn a browse into a decision.
Chasing virality is chasing someone else's distribution. Finding content that beat its own channel's norm is a repeatable method. Here's how to do it.
Buyers aren't judging your best work. They're reducing risk on one specific job. Build for that and the same reel starts converting.
The fields that decide what you get back, why a pitch cap improves your pitches, and what to write in the two boxes that matter.
The four axes that define what a client may do with your work, why they're priced separately from production, and the carve-out to never sign away.
Does AI push freelance rates down? Honestly, for some work, yes, and it's already happening. Here's which work, and where the pricing power moved.
Anyone can produce a beautiful reel. What they can't produce is a record of someone else booking them, paying them, and coming back.
What each part of a Snapshot is answering, which numbers to trust, and the three most common ways people misread their own results.
Six fields turn "sound like us" into something another person (or a model) can actually follow. Plus the line worth not crossing.
Versioned cuts and frame-accurate comments remove the arguing about facts. What's left is giving notes well, which is still on you.
One variable, a frozen benchmark, a sample size decided in advance, and a result you agreed to accept before you saw it.
One calendar, holds that expire, and buffer days you never sell. The scheduling habits that stop a good month turning into a crisis.
Your slice of the attention in a defined set: how to compute it from public data, and why the level is arbitrary but the direction isn't.
Build it in edit order to catch what's missing, shoot it in setup order to save hours, and never write the words "b-roll" on it.
The path a finding takes to become a shoot day (Snapshot, Signal Brief, Creative Brief, production docs) and where a human has to decide.
Generalists were resilient when work came through relationships. Now that it comes through search, breadth makes you invisible. How to narrow without starving.
When execution is free, the scarce skill is choosing. What taste actually is, why it's mostly not innate, and how to build it deliberately.
Owner, admin, member: what each can do, how seats are counted against your plan, and the one thing a teammate's tier is inherited from.
Reply inside a day or lose the thread entirely. What the rule is, why it exists, and how to run an inbox that respects a clock you don't control.
Supply went up by an enormous multiple and demand stayed exactly where it was. Every other consequence in this pillar follows from that one asymmetry.
Outreach volume stopped working the moment sending got free. What replaced it is being present where the brief lands, and being findable when it does.
Four tiers of content evidence, from public view counts to sales data: what each can prove, what it can't, and how to know which one you're holding.
Competitive analysis was expensive enough that most teams did it once a year. It now runs continuously, which changes what you should do with it.
One person with good tooling can now deliver what took a small crew. Where that's genuinely true, where it quietly isn't, and how to price it.
Your thumbnail will be judged at the size of a postage stamp, moving, next to twenty others. Design for that, and test it properly.
How to read and write HH:MM:SS:FF, why "around the middle bit" costs an editor ten minutes, and the note format that gets acted on first time.
Why it isn't an average, why a single 5★ won't move it much, and what actually raises it, including the dimension most creators ignore.
Two-sided dates, a cadence read that shows what you actually published, and the plan-versus-actual comparison most teams avoid looking at.
The highest-leverage thing in Acumin for the least ongoing effort, and the one most people set up last.
Hand a language model your analytics and it will produce a number. Where that number comes from, and the architecture that stops it being invented.
An honest inventory. Four things got dramatically cheaper, three things got harder, and one thing didn't move at all. The last one decides everything.
Not shutter presses. Four things that survived the collapse in production cost, and how to make each of them visible in a pitch.
Five things that were true before any of this and will be true after. If a strategy depends on the tools staying still, it isn't a strategy.
Silence usually isn't about you. A follow-up ladder that stays professional, protects the schedule, and knows when to stop.
An argument, not a forecast. Four things that follow if the premises hold, the assumptions they rest on, and what would prove them wrong.
Boosting your best post is usually the wrong move. What to look for instead, and why the strongest candidate is often one that underperformed.
Analysis used to be scarce and expensive, so it was scrutinised. It's now abundant and free, and abundance without verification is just confident noise.
The return leg of the loop. Record what you changed, let the benchmark freeze, and read the result against a number that can't move underneath you.