A working model of how content gets found for a $500k business
Search catches demand that exists; social creates it. Each piece serves one stage of a buyer's readiness, volume matters with diminishing returns, and one clear topic per account wins. The model, the evidence for each part, what it did for a business stuck at $500k a year, and a diagnostic to run on your own accounts.
2 steps in order, lit one after another by a travelling dot: The Magnet: social media creates intent, Exploration; The Pumpjack: SEO extracts existing intent, Exploitation. A return line closes the loop: Returns fund more exploration.
Contents
The answer
4 steps in order, lit one after another by a travelling dot: TOFU: discovery; MOFU: trust; BOFU: conversion; Search: capture existing demand. A return line closes the loop: Returns fund more exploration.
Content gets found through two different machines, and most businesses run only one of them.
- Search catches demand that already exists. Social creates demand where there was none. Social explores for what people want; search then captures the winners.
- Every piece of content serves one stage of a buyer's readiness: discovering you, trusting you, or deciding to buy. Asking one piece to do all three is the most common failure.
- Volume matters, with diminishing returns. In Buffer's study of over 2 million Instagram posts, 3 to 5 posts a week was the sweet spot for reach and follower growth.[1]
- One clear topic per account wins. Since 10 December 2025, Instagram users can see the topics the algorithm thinks they like and remove them in one tap.[2] An account that mixes topics is easier to remove.
- The first seconds carry the scent. People forage for information the way animals forage for food, and follow cues that promise the most value for the least effort.[3]
What it did for a business stuck at $500k
Four steps light in turn: a book, a video course, live group coaching in cohorts of eight at about $8,000 each, and the $50,000 package, the last in green.
Marley Baird Media sold one offer: a $50,000 done-for-you YouTube package. It had flatlined at $500,000 a year. The founders thought they needed better ad management. The ads were running and the traffic was arriving. Nothing converted, because cold traffic was meeting a $50,000 offer with nothing in between.
We built the missing middle as a value ladder: a book, a video course, live group coaching for cohorts of eight at about $8,000 each, then the $50,000 package at the top. The lower rungs didn't turn directly into $50,000 clients. They paid for the advertising, and by the time a serious buyer arrived, Marley's work was everywhere they looked. The $50,000 bookings quadrupled, and at its height the business was running more than $100,000 a month in ad spend, profitably.[4]
The model's funnel stages come from running campaigns at scale before that: at Willowbrook Chrysler, discovery spots on the Rogers Arena jumbotron during Canucks games, mid-funnel display, and bottom-funnel ads for people who already knew they wanted a Ram.[4]
The five parts and the evidence for each
5 steps in order, lit one after another by a travelling dot: Explore vs. Exploit; TOFU / MOFU / BOFU; Batch Combinatorics; Foraging and Visual Resets; Thing for People in Place.
1. Explore, then exploit. Treat social as exploration, a search for what people respond to, and the website as exploitation, catching the demand that's been found. This is the explore-exploit trade-off from decision theory, applied to marketing effort. A business that only explores goes viral and stays broke; one that only exploits ranks well and slowly loses relevance.
2. One stage per piece. Discovery content answers "who is this?", trust content answers "can they help?", and decision content answers "should I buy?". Our evidence is the Marley Baird ladder above: the missing trust stage was the whole problem.
3. Volume, made affordable. Volume helps, with diminishing returns.[1] The way to afford it is to record content as interchangeable parts: openings, main points and calls to action. Ten openings, five main points and five calls to action make 250 possible videos from one recording session, and every combination is a test of which part did the work.
4. One topic per account. Describe each offer as a service for a specific group in a specific place: "website design for solo law practices in British Columbia", not "web design". Since December 2025, users can purge topics from their own feed,[2] so mixed signals now cost followers as well as reach. Instagram's Trial Reels, launched in December 2024, let you show a new angle to non-followers first.[5]
5. Scent in the first seconds. The opening has to promise the value the viewer is foraging for.[3] Frequent visual changes hold attention; the right pace is a claim to test on your own audience.
A deep dive on each part:
- Strategic deployment with Thing for People in Place
A framework for thinking about hyper-specific targeting versus broad appeal, and how the December 2025 algorithm updates raised the weight of niche purity in content distribution.
- The Explore vs. Exploit Framework
Notes on why a website and social media can be read as serving different purposes, using the explore/exploit problem from machine learning as a lens.
- The TOFU/MOFU/BOFU Content Framework
Research notes on why short-form video struggles to serve every audience at once, and a three-tier model for segmenting content by the viewer's stage of readiness.
- The neurology of foraging and visual resets
A framework linking information foraging theory to the neural circuitry proposed to govern attention, discovery, and the decision to keep watching, with the sources behind each step.
- The production engine is batch combinatorics
Research notes on a framework that treats videos as engineered assemblies of interchangeable modules (hook, meat, CTA) rather than organic creative works, and on the combinatorial arithmetic that follows.
Not yet proven
A panel labelled the model holds five blocks on a shelf, each with a green check. Its door has a keyhole labelled primary source. On a bench outside sit three cards with empty checkboxes: slower speech, 3x comments; account-based, 33x return; and a brain reason, not established. One at a time each card slides to the door, the keyhole flashes because it has no key, and the card goes back to wait on the bench.
These came from practice and secondary sources, and haven't been checked against a primary source:
- Slowing speech to about 0.8 seconds per word while keeping visuals fast tripled comments.
- Hyper-targeted account-based campaigns returned 33 times their spend.
- A specific brain mechanism makes viewers swipe after visual stagnation. The foraging model[3] predicts the behaviour; the brain-level account is not established.
What this says about how we work
Two blocks side by side. The romanticized, artistic view: Content as creation. The empirical view: A model of how human cognition and platform algorithms interact. Whether it holds a lasting advantage is an open question; platforms change their rules.
We build models from campaigns we've run, then check each part against the best evidence we can find, and say plainly which parts are proven and which are still bets. That's the same work we'd do on your distribution before you spend another month's budget.
Run the diagnostic on your own accounts
A list of 4 rows: Audit the current balance; Build a production engine; Refine the signal; Deploy strategically. A green tick fills each box in turn.
- Explore or exploit? All social and no website: you're creating demand nobody captures. All website and no social: you're living off demand that's slowly shrinking.
- Is there a middle? List what someone can buy or join between hearing of you and your main offer. If nothing, build one rung.
- Can you reach 3 to 5 posts a week? Record one session of parts: 10 openings, 5 main points, 5 calls to action.
- One topic per account? Write each account's focus as "[service] for [people] in [place]". If you can't, split it.
- Test before you commit. Use Trial Reels to try a new angle on non-followers first.
If your content is getting views and no buyers, I'd like to hear about it: mackenzie@mackenziebowes.com or @_mackenziebowes.
Sources
- Buffer (2025), How often should you post on Instagram? Data from over 2 million posts. buffer.
com/ resources/ how- often- to- post- on- instagram/ - Search Engine Land (December 2025), Instagram's "Your Algorithm" tool lets users see and edit the topics in their Reels feed. searchengineland.
com/ instagram- your- algorithm- 465987 - Pirolli and Card (1999), Information foraging, Psychological Review 106(4). doi.
org/ 10. 1037/ 0033- 295X. 106. 4. 643 - How to quadruple revenue in six months: the Marley Baird Media case, on this site. www.
mackenziebowes. com/ stories/ quadruple- revenue- six- months - 9to5Mac (10 December 2024), Instagram's Trial Reels let creators test content with non-followers. 9to5mac.
com/ 2024/ 12/ 10/ instagrams- new- trial- reels- feature- lets- creators- test- content- with- non- followers