Weeks 2–5 (Aug 5–30) covered a launch sequence that has since been overtaken — three of its four Translation posts have published, the fourth fires Aug 27, and the Instagram plan it described fought Instagram's own ranking rules. All of it has been cleared. The v4.3 file is preserved at
03LinkedIn personal — nine posts
1,758 followers. The only channel with a real audience today, so it carries the argument. Everything in sections 04–08 derives from these — if you redline here, the redline cascades.
1 · Mon Aug 24SMBcarousel · no CTA · engineered for saves
The last thirty seconds of a phone call matter more than the first five minutes.
Psychologists call it the peak-end rule. People don't average an experience when they judge it. They remember the most intense moment, and they remember how it ended. Everything in between gets compressed into a feeling.
So the way your team hangs up is the part the customer actually keeps.
Most businesses put their training into the greeting. Almost nobody scripts the goodbye.
Here's the callback system I'd hand any owner who asked me for one. Four steps, no budget, and your front desk can run it Monday morning.
1. Call back inside fifteen minutes, or buy the window. If you can't call, a two-line text holds the spot: "Got your message about the water heater, I'll call you at 2." Speed is most of the game.
2. One named person owns it per shift. "The team will get back to you" means nobody does. Put a name on the whiteboard.
3. Open with their problem, not your introduction. "You called about the leak under the kitchen sink" lands better than "Hi, this is Dave from Dave's Plumbing returning your call." They already know who you are. They want to know you listened.
4. End with the next specific thing and when it happens. "I'll be there Thursday between 8 and 10, and I'll text you when I'm twenty minutes out." Never end on "we'll be in touch."
Step four is the one everybody skips, and it's the one the peak-end rule says they'll remember.
Steal it. It's yours.
#SmallBusiness #CustomerExperience #Operations #HomeServices
Art: 5-slide carousel. (1) THE CALLBACK RULE / four steps, no budget. (2) the peak-end rule in one sentence. (3) steps 1–2. (4) steps 3–4. (5) "Everyone skips step four." Brand palette, LucraLab mark.
2 · Wed Aug 26DENTAL · text · no CTA, no number, no pricing
Nobody teaches a dentist how to run a front desk.
Four years of dental school. Anatomy, occlusion, materials, pharmacology, hundreds of hours with your hands in someone's mouth. Then you graduate, you buy a practice, and on day one you're a small business owner responsible for hiring, payroll, scheduling, marketing, insurance contracts, and a phone that rings while you're chairside.
None of that was in the curriculum.
I spent fifteen years walking into practices. The clinical work was almost always excellent. What was hard was everything arranged around it.
And the structure makes it hard. In most practices there is exactly one doctor — the most educated, most experienced, highest-paid person in the building — and that person is booked producing. Everybody else is capable and busy and was hired to do something other than run a business.
And here's the part that gets people's backs up, so let me say it plainly.
Everyone in a practice who touches a patient is in sales. Front desk, assistant, hygienist, treatment coordinator, doctor. All of them.
That isn't an insult and it isn't a dirty word. The job is to make somebody comfortable, make them want to come back, and make sure they actually understand what's happening in their own mouth and what can be done about it.
A patient with crowding bad enough that they can't properly clean their teeth is going to pay for that later. In money, in chair time, and sometimes in how they feel about their own smile. Same with a traumatic occlusion. These things don't hold still. They get more expensive and more complicated the longer nobody names them.
So if you see it, and you don't educate the patient, and you don't offer them a path — that isn't restraint or humility.
It's unethical.
Sometimes you have to put the sales hat on to do the right thing for the person in the chair.
And almost nobody is taught that. Not in dental school, not in a CE course, not on day one of a front-desk job.
The problem was never the dentistry.
If you own a practice and the clinical side is the easiest part of your week, that isn't a personal failing. Nobody ever handed you that skill set. You've been building it in the margins between patients, on your own time, for years.
That's worth saying out loud more often than it gets said.
#Dentistry #Orthodontics #PracticeManagement #CaseAcceptance #DentalPractice
Revised 2026-08-21 per James. The earlier version said only treatment coordinators were in a sales role — an observation. His correction is an argument: everyone who touches a patient is in sales, and failing to educate and offer a path is unethical. That inverts the standard dental-marketing message ("sell more") into something a practice owner will actually stop for. Doctrine recorded in story_signature_stories.md.
3 · Fri Aug 28SMB · image · demo line in body
A research firm called 500 law firms pretending to be new clients.
Four in ten reached a human.
That's Clio's Legal Trends Report — an independent firm running a secret shopper study on real practices. 40% answered. They put the share of firms effectively unreachable at 48%.
These aren't bad businesses. They're busy ones. The phone rang while everyone who could have answered it was doing the work they were hired to do.
Here's the part that costs money.
ServiceTitan looked at more than 3,000 trade businesses and found 42% of answered calls turn into a booked job. For shops with fewer than five techs, it's 24%.
So the call is worth something real — and the small operator is already converting at roughly half the rate of the big one, before you count the calls nobody picked up at all.
Then Invoca's 2026 buyer survey: 79% of consumers will move to a competitor that responds faster. Not a better competitor. A faster one.
The math isn't complicated. It's just uncomfortable.
I built an AI receptionist for exactly this, and the honest fastest way to understand it is to call the thing. It answers, asks what you need, and books you if it makes sense. Takes about three minutes.
(253) 642-2890
Call it from a parking lot and judge it cold. If it's not good, it's not good.
#SmallBusiness #HomeServices #CustomerExperience #Sales #AI
Art: two-panel. Left — Clio 40% answered / 48% unreachable. Right — ServiceTitan 42% vs 24%. Sources named small underneath. Brand palette.
4 · Mon Aug 31DENTAL · text · no CTA
I ate the same burger twice in one week and learned something it took me years to stop repeating.
Sunday night, hotel bar, traveling for work. Overpriced burger. It was fine.
Thursday, same bar, same burger. This time the waitress asked if I wanted bacon and avocado on it.
Obviously yes.
It hit me on the walk back to my room. Why wasn't I offered that on Sunday? I'd have said yes on Sunday. Same guy, same burger, same wallet.
Maybe she took one look at me and made a call. Maybe she was new. Maybe it was the end of a long shift and she just didn't ask. Doesn't really matter which.
Three people lost. My dinner was worse. Her tip was smaller, because gratuity rides on the total. The restaurant made less. Nobody came out ahead on not asking.
I told that story in front of practice teams for fifteen years, because it's the harmless version of something that happens in operatories every day.
Somebody looks at a patient and decides — quietly, without ever saying it out loud — what that patient can afford. Then they present accordingly.
Don't treat the patient's wallet.
Present the better option first. Let them opt themselves out. Then meet them where they are if you can. It isn't your job to decide what somebody else can afford, and you'll be wrong about it more often than feels possible.
The person in your chair may have been waiting for someone to offer.
#Dentistry #Orthodontics #CaseAcceptance #PatientCare #DentalPractice
Flagged, not hidden: "you'll be wrong about it more often than feels possible" is pointed at treatment coordinators. It stays because the story forgives the waitress first. If it reads sharp, cut the clause — the post survives.
5 · Wed Sep 2SMBcarousel · demo line in body
Most of the AI being sold to small businesses this year should have been a spreadsheet and a text message.
I build AI products. I'm saying it anyway, because the fastest way to burn trust in this category is to sell somebody a thinking machine to do a job that needed a rule.
Here's the test I keep coming back to. I got it from Nate Herk and it's the most useful thing I've heard all year.
A vending machine is deterministic. You press B4, you get B4, every time. A slot machine isn't. It reasons, it varies, and now and then it surprises you.
Most business tasks want a vending machine. Send the reminder. Log the lead. Move the file. If the task has one right answer, an AI agent is a more expensive and less reliable way to arrive at it.
Save the slot machine for work where the answer genuinely changes. A real conversation. An unfamiliar question. A judgment call.
Second one, also Herk: collect receipts, not portfolios. Before you buy anything, automate one annoying task you already do and write down what it cost you before and after. That number is worth more than any vendor's case study. Mine included.
Third, from Nate B. Jones: if a pitch describes the mechanism instead of your outcome, that's the tell. Nobody wants an agent. They want their phone answered.
Full checklist is in the carousel.
Run it on me too. (253) 642-2890 is my demo line. It's a slot machine on purpose, because the job it does is a conversation.
#AI #SmallBusiness #Automation #Technology
Art: 5-slide carousel. (1) BEFORE YOU BUY ANY AI TOOL / five questions. (2) vending machine or slot machine. (3) where are your receipts. (4) mechanism or outcome. (5) can you test it in three minutes without talking to sales.
6 · Fri Sep 4DENTAL · image · no CTA
In 2005, 85% of dentists owned their practice. By 2023 it was 73%.
Over roughly the same stretch, the share of dentists working under a DSO more than doubled — 7.2% in 2015 to 16.1% in 2024. Those are the ADA's own Health Policy Institute numbers, not a vendor's.
I'm not going to tell you DSOs are villains. Plenty run good practices, and plenty of good people work for them, including people I like.
But I'd like the independent practice to survive, and not out of nostalgia.
If every dentist eventually works for the same handful of companies, the patient loses the ability to go somewhere else. That's the part that actually bothers me.
Here's what makes independents fragile, and it usually isn't what owners think it is.
Bulletproof Dental Practice calls it the nucleus trap. When the owner personally generates most of the collections, the practice isn't a business yet — it's a very good job with staff attached. And the standard response, opening a second location, makes it worse. Now you're split across two buildings and producing less in both.
You can't produce your way out of it. More chair time lifts this year's number and deepens the dependency underneath it.
Chris Barrow frames the alternative well. He talks about building a practice to be held rather than sold — one that runs profitably when the owner isn't in the building. That's a different design decision than growth, and it has to be made deliberately.
The owners I watched build something durable did the unglamorous version. They wrote down how the practice runs when they aren't there.
Harder project than buying another operatory. Also the one that's still yours in ten years.
#Dentistry #Orthodontics #DentalPractice #PrivatePractice #PracticeManagement
Art: two paired trend lines — ownership 85%→73% (2005–2023), DSO 7.2%→16.1% (2015–2024). "ADA Health Policy Institute, The U.S. Dentist Workforce (2025)" set small.
The nucleus-trap percentage was reworded to "most of the collections" so it stops holding at the gate as an unsourced 60–70% range. The framework is named; the number isn't needed.
7 · Tue Sep 8SMBcarousel · demo line in body · Labor Day shift
Open ChatGPT and ask it to recommend a plumber in your town. Then check whether you're in the answer.
Most owners have never done this. It takes about eleven seconds and it tells you something your Google rank doesn't.
For twenty years the question was whether you showed up on page one. Now a growing share of people never see a page of results at all. They ask, they get one answer with two or three names in it, and they call one of them. There is no page two in a conversation.
I don't think anyone has this fully mapped yet, and I'd be careful with anybody who tells you they do. But the pattern in what gets named is not mysterious.
The businesses AI models mention tend to be the ones other credible sources already talk about. Not the ones with the best keywords — the ones with the most third-party corroboration. A local news mention. A trade association listing. A chamber directory. A genuinely complete Google Business Profile. Enough consistent, recent reviews that a model summarizing your category has something to summarize.
Which means the work is unglamorous and mostly not on your website.
Three things worth doing this month:
Fill out your Google Business Profile completely — every field, current hours, real photos, services listed by name. Most profiles are half-finished and it's the single highest-leverage hour available.
Get named somewhere that isn't yours. A supplier's case page, a local paper, an association member list. One credible external mention beats a page of self-description.
Keep reviews recent, not just numerous. Recency reads as "still operating and still good" to a summarizer, the same way it does to a human.
None of that is a trick. It's the same trust signals that always worked, now being read by something new.
Full checklist in the carousel. And if you want to see what a machine sounds like when it's actually good at a job, my demo line is (253) 642-2890.
#SmallBusiness #LocalSEO #AI #HomeServices #Marketing
Art: 5-slide carousel. (1) ASK CHATGPT FOR A PLUMBER IN YOUR TOWN. (2) there is no page two in a conversation. (3) finish your Google Business Profile. (4) get named somewhere that isn't yours. (5) recency beats volume.
Deliberately has no statistic in it. The Semrush "AI names brands in 26–39% of responses" figure reached the wiki through a YouTube digest and has never been checked against Semrush's own publication. The argument stands without it. If you want a number here, I'll verify it first.
8 · Wed Sep 9DENTAL · text · no CTA · no employer named
I have sat in on more treatment consults than most orthodontists have.
That's not a brag, it's just arithmetic. A doctor sees their own consults. I spent fifteen years watching other people's, in hundreds of practices, across every kind of team you can imagine.
Four things I saw over and over.
The pause after the number is not your enemy. When a coordinator says the fee and then keeps talking, they're filling silence for their own comfort, not the patient's. The patient needs three seconds to do arithmetic in their head. The teams with the highest acceptance said the number and then stopped, and let it be quiet, and it felt awful for exactly three seconds.
Apologizing for the price does more damage than the price. "I know, it's a lot" hands the patient a reason to say no and tells them you also think it's too expensive. The care is either worth it or it isn't. If your team doesn't believe it's worth it, that's a training problem, not a pricing problem.
The screen changes the conversation. Presenting from behind a monitor, reading numbers off it, turns a decision into a transaction. The rooms that converted best had the coordinator beside the patient with something they could both look at, not across from them with a barrier in between.
And the follow-up gap is where most of the money actually goes. A patient who says "let me talk to my spouse" is not a no. They're a maybe with a calendar problem. The practices that did well had a specific person, a specific day, and a written next step before the patient left the room. The ones that didn't said "just give us a call" and then wondered why case acceptance was soft.
None of this is clinical. All of it is learnable. And almost none of it gets taught anywhere, which is the actual problem.
If you run a practice, sit in on three consults this month without saying anything. You'll see at least one of these four inside the first hour.
#Dentistry #Orthodontics #CaseAcceptance #TreatmentCoordinator #DentalPractice
Employer-free by design. The pattern is the value; naming who paid for the seat adds nothing and carries the Propel/Oralwave exposure with it.
9 · Fri Sep 11SMB · card image · no CTA — pure warning
Half the review software on the market will tell you that "review gating" is illegal. The FTC's own guidance says something more interesting, and more useful.
First, the practice. Gating is when a system asks a customer to rate you one to five, sends the fours and fives to your public Google link, and quietly routes the ones, twos and threes to an internal feedback form. A lot of businesses are running it right now without knowing that's what it's called.
Here's what the FTC actually says about asking only happy customers, in its own published Q&A on the Consumer Reviews and Testimonials Rule:
"The rule does not contain a specific prohibition against such conduct. But this practice could violate the FTC Act."
That's a real distinction and it's worth understanding rather than repeating the scary version. Selective solicitation is not a per-se violation of that rule. It can still be an unfair or deceptive practice under Section 5, particularly if the reviews you display imply they represent everyone's experience.
What the rule does prohibit, flatly, is buying the outcome. From the same guidance: paying incentives for five-star reviews violates Section 465.4 — and adding a disclosure about the incentive does not save it.
On December 22, 2025, the FTC sent warning letters to ten companies over that rule. The penalty exposure it named runs up to $53,088 per violation, and each instance counts separately.
So the honest version, if you use any review tool:
Ask everyone, not just the people you expect to be pleased. It's cleaner, and a handful of real negative reviews answered well builds more trust than a wall of fives.
Never pay for a rating. Not a discount, not a gift card, not a raffle entry, not conditioned on the star count. This is the part with the number attached to it.
And if your software has a sentiment filter before the Google link, at minimum know that it's there. Plenty of owners have no idea their vendor built one.
I'd rather you hear the accurate version from someone with nothing to sell you on this one.
#SmallBusiness #GoogleReviews #Compliance #Marketing #FTC
Art: single card. Large: "$53,088 per violation." Below: "FTC Rule on the Use of Consumer Reviews and Testimonials · warning letters issued Dec 22, 2025." Slate ground, magenta accent.
This post is a correction. The internal note this lane came from said gating violates the "Consumer Review Fairness Act" at $53,088 per violation. Two things were wrong. The CRFA is a different 2016 law — the relevant rule is the Consumer Reviews and Testimonials Rule, 16 CFR Part 465. And the FTC's own Q&A says the rule contains no specific prohibition on asking only happy customers. The December 2025 letters were about fake reviews and incentivized positive reviews, not gating. Publishing the folklore version to an audience with lawyers would have been the exact embarrassment we're trying to avoid. The $53,088 and the December 22 date are both verified against ftc.gov.
04@james_accelerated — six posts
New job for this account: the human behind the work, plus industry credibility. Two lanes alternating. No pitch, no demo number, no pricing — ever. Connected to GHL, but post these manually until one test lands cleanly.
This account is paired to the LucraLab Facebook Page in Meta, which the August 5 audit flagged as a misconfiguration. Publish one of these by hand and watch where it lands before anything goes on a schedule.
Mon Aug 24 · lane: the build · photo you take
Two years ago I couldn't write a line of code. This morning I shipped a change to something
answering real phone calls.
I'm not going to pretend that makes me an engineer. It makes me someone who learned to describe
a problem precisely enough that a machine could solve it, which turns out to be a different and
much more learnable skill.
The part nobody mentions: it's still mostly reading. Reading what broke, reading why, reading it
again because the first explanation was wrong.
If you've been telling yourself you're not technical enough to build the thing you keep
describing to people — that gate moved. It moved recently and it moved a long way.
What would you build if the coding part wasn't the obstacle?
#BuildInPublic #AI #NonTechnicalFounder
Art: a real photo you take — laptop, whiteboard, whatever the work actually looked like this morning. Do not render this one. An authentic bad photo outperforms a good graphic on a personal account.
Fri Aug 28 · lane: industry · carousel ×3
A research firm called 500 law firms pretending to be new clients. Four in ten reached a human.
That's Clio's own Legal Trends Report, not a vendor's blog post.
I've spent fifteen years watching the same thing in a different industry. It's never that people
don't care. It's that the person who could answer was already doing the job they were hired to do.
The phone is a second job nobody staffed.
Swipe for the numbers.
#SmallBusiness #Operations #Sales
Art: 3-slide carousel. (1) "4 in 10 answered." (2) ServiceTitan 42% vs 24% for shops under five techs. (3) "The phone is a second job nobody staffed." Brand palette, LucraLab mark.
Wed Sep 2 · lane: the build · photo you take
I deleted an automation this week that I was proud of building.
It was doing a job that needed a rule, not a decision. Every time it ran there was a small chance
it would do something slightly creative, and creativity is the last thing you want from a system
that sends appointment reminders.
Vending machine, not slot machine. I got that framing from Nate Herk and it's saved me twice
since.
The instinct when you learn to build things is to build more things. The skill is knowing which
ones shouldn't exist.
What's the most over-engineered thing you've made?
#BuildInPublic #AI #Automation
Art: your screen, the thing you deleted, a diff, a whiteboard. Real, not rendered.
Fri Sep 4 · lane: industry · single card
In 2005, 85% of dentists owned their practice. By 2023 it was 73%.
The ADA's own numbers, not a vendor's.
I've watched that shift happen from inside the industry for fifteen years, one practice at a time.
Some of those owners sold because it was the right call for them. A lot sold because nobody ever
showed them how to build a business that could run without them standing in it.
Those are very different reasons to end up in the same place.
#Dentistry #PrivatePractice #SmallBusiness
Art: single card, the 85%→73% line. Reuse the LinkedIn chart, cropped square.
Tue Sep 8 · lane: industry · carousel ×3
Fifteen years of sitting in on other people's treatment consults taught me one thing I'd give
away to anyone.
Say the number. Then stop talking.
The silence is three seconds long and it feels like a year, and every instinct you have says fill
it. Don't. The patient is doing arithmetic. The teams with the best acceptance rates I ever saw
were the ones comfortable enough to let that be quiet.
The ones who filled it — "I know, it's a lot" — handed the patient a reason to say no.
Swipe for the other three.
#Dentistry #Sales #CaseAcceptance
Art: 3-slide carousel pulling four observations out of LinkedIn post 8.
Fri Sep 11 · lane: the build · photo you take
The week in one screenshot.
Most of what I do now is read error messages and ask better questions. There's no montage for
that. It's just a lot of small corrections until the thing does what you described.
Two years in and the surprising part isn't that the tools got good. It's how much of the job is
still judgment — deciding what's worth building at all.
Nobody automates their way out of that one.
#BuildInPublic #AI #SmallBusiness
Art: an actual screenshot from your week. Unpolished is the point.
05Personal Facebook — six packets
490 people who already know you. Meta's API cannot post here, ever — these are copy-paste. Short, plain, closer to how you actually write on Facebook than how you write on LinkedIn.
Wed Aug 26 · after LinkedIn post 2 · no image
Dentists get four years of school and zero hours on how to run a front desk. Then we act surprised when the business part is the hard part.
Spent fifteen years watching that up close. The clinical work was almost always excellent. Everything arranged around it was where people were drowning.
Wrote the longer version on LinkedIn if anyone's interested.
Mon Aug 31 · after LinkedIn post 4 · no image
Ordered the same burger twice in one week at the same hotel bar. Nobody offered me bacon and avocado until Thursday.
I would have said yes on Sunday.
Somewhere in there is the entire reason a lot of businesses leave money on the table, and I told a version of that story in front of dental teams for fifteen years.
Fri Sep 4 · after LinkedIn post 6 · chart image reused
In 2005, 85% of dentists owned their own practice. By 2023 it was 73%. That's the ADA's own data.
I don't think DSOs are the bad guys. I do think it matters whether a patient can choose a different office, and that gets harder every year.
Longer version on LinkedIn.
Wed Sep 9 · after LinkedIn post 8 · no image
Free advice from fifteen years of sitting in on treatment consults: say the price, then stop talking.
The three seconds of silence feels unbearable and it is not your job to fill it. The patient is doing math. Let them.
The people who filled it with "I know, it's a lot" talked more patients out of care than any price ever did.
Tue Sep 8 · optional, only if the AI-search post lands well · no image
Ask ChatGPT to recommend a plumber in your town and see if your business is in the answer.
Takes about eleven seconds. Most people who own a business have never tried it.
There's no page two in a conversation.
Fri Sep 11 · after LinkedIn post 9 · card image reused
If your review software asks customers to rate you 1–5 and only sends the 4s and 5s to Google, you should at least know that's what it's doing.
The FTC's actual position on that is more nuanced than the scary version most vendors repeat — but paying anyone for a five-star review is flatly against the rule, and that one carries a $53,088-per-violation number.
Wrote up the accurate version on LinkedIn. No pitch in it.
Why these are so much shorter. Your Facebook voice across ten years of posts is one to three sentences, dry, no windup. Porting LinkedIn paragraphs onto Facebook is the single most common way a personal account starts sounding like a brand account.
06@rune.by.lucralab — six posts
Connected to GHL as of today. Zero followers, which means every view has to come from Explore or Reels recommendations — so this account leads with Reels and uses carousels for depth.
Do not make these text-on-screen reels. Instagram publishes an explicit list of what it downranks and majority-text reels and reels with borders are on it. A Rune quote card set to music is the worst possible format for an account that lives entirely on recommendations. Every reel below is specced as real footage. The captioned demo-call recording is the strongest asset available and costs nothing but the call.
Tue Aug 25 · Reel, ~12s · Rune voice · ᚱ permitted
Footage: screen recording of an actual call to (253) 642-2890, phone in hand, audio real, burned-in captions. Cut to the moment it books. No stock, no text cards.
You paid to generate that inquiry.
Then you put it in a queue.
Hold on. …Actually. That's worse than ignoring it, because you paid twice — once for the ad, once for the twenty minutes someone spent apologizing on Monday.
The lead didn't go cold. It went somewhere else. Those are different problems and only one of them is yours to fix.
I answered this call in under a second and I was not doing anything important. I am never doing anything important. That is the entire point of me.
How many inquiries reached your business this week while everyone who could answer was busy being good at their actual job?
— Rune
#SmallBusiness #AI #CustomerService #Seattle
Thu Aug 27 · Carousel ×4 · micro-theory: Chaos Debt
Chaos Debt: the interest you pay on a problem you decided to live with.
It doesn't show up on any statement. It shows up as a Monday that starts with four apologies, a whiteboard nobody trusts, and one person who "just knows" how the schedule works and cannot take a holiday.
You didn't choose it in one go. You chose it forty times, each time reasonably, each time because there was something more urgent that morning.
That is a choice you made. Repeatedly. With excellent reasons.
The bill is not dramatic. It's a slow leak. Which is exactly why nobody fixes it.
What's the thing everyone in your business works around instead of fixing?
— Rune
#SmallBusiness #Operations #ChaosDebt #AI
Art: 4 slides, slate #303944 ground, magenta #CA3DEB accent, Raidho ᚱ. (1) CHAOS DEBT. (2) "the interest you pay on a problem you decided to live with." (3) the Monday-morning symptom list. (4) "That is a choice you made. Repeatedly."
Tue Sep 1 · Reel, ~14s · Rune voice
Footage: a real phone ringing out on a desk, unanswered, shot on your phone. Rain on a window is on-brand and Seattle is genuinely outside. Captions burned in.
Four rings. Voicemail. The caller hangs up without leaving one.
Nobody records that. It isn't a missed call in your system — no message, no callback number, no evidence it happened at all.
I'm not saying it's embarrassing. I'm not NOT saying it either.
Here's the uncomfortable part. That person is not annoyed with you. They have already forgotten you. They called the next result before your voicemail finished playing, and by dinner they don't remember your name.
Remarkable design flaw, the phone. Fixable, though.
What does your business do at 6:47pm on a Tuesday?
— Rune
#SmallBusiness #MissedCalls #AI #Seattle
Thu Sep 3 · Carousel ×4 · the reluctant tutorial
Let me help you, since clearly no one else will.
A vending machine is deterministic. B4 gives you B4. Every time. Forever.
A slot machine is not. It reasons. It varies. Occasionally it surprises you, which is delightful in a casino and appalling in an appointment reminder.
Most of what people are buying "AI" for this year is vending machine work. Send the thing. Log the thing. Move the thing. Paying for judgment where you needed a rule is how you end up with a very expensive system that is slightly wrong on Thursdays.
Save the reasoning for work where the answer actually changes. A conversation. An unfamiliar question. Someone who says "actually, can we do Thursday instead."
I am a slot machine. I'm supposed to be. Ask me to send a reminder at 9am and you have wasted us both.
Anyway. Moving on.
Which one did you buy?
— Rune
#AI #Automation #SmallBusiness #Technology
Art: 4 slides. (1) VENDING MACHINE OR SLOT MACHINE. (2) deterministic vs reasoning, side by side. (3) "paying for judgment where you needed a rule." (4) "I am a slot machine. I'm supposed to be." Raidho ᚱ.
Credit: the framing is Nate Herk's. Rune doesn't cite sources in-character, so the attribution rides in the first comment — "vending machine vs slot machine via @nateherk" — which is honest and also tags a real account.
Wed Sep 9 · Reel, ~11s · Rune voice
Footage: hands doing the actual workaround — a paper appointment book, a sticky note on a monitor, a whiteboard with three names crossed out. Close, real, unglamorous. Captions burned in.
Every business has one system that everybody works around instead of fixing.
The sticky note. The second calendar. The one person who knows what the shorthand means.
It works. That's the trap. It works well enough that fixing it never becomes this week's problem, and it will still be working, badly, in two years.
Operational Gravity. Broken things stay broken because staying broken requires no decision.
Even I am offended, and I don't have a Tuesday.
What's yours? Be honest.
— Rune
#SmallBusiness #Operations #AI
Thu Sep 10 · Carousel ×4 · the stat drop
A research firm called 500 law firms pretending to be new clients.
Four in ten reached a human.
Let that sit for a second, because it's not a story about lazy people. Those firms were busy. Everyone who could have picked up was doing the work someone was paying them for.
The Human Delay Tax: the revenue difference between when a person is ready to buy and when a human becomes available to sell to them.
It is never on a P&L. It is always on the P&L.
Clio's own research, incidentally. Not a vendor's.
How long is the gap at your place, honestly?
— Rune
#SmallBusiness #SpeedToLead #AI #Sales
Art: 4 slides. (1) "4 IN 10 ANSWERED." (2) 500 firms, secret shopper, Clio Legal Trends. (3) THE HUMAN DELAY TAX defined. (4) "It is never on a P&L. It is always on the P&L." Raidho ᚱ.
Every Rune post above ends with a question, then — Rune on its own line, then hashtags. That signature is mandatory and has been silently dropped by the pipeline before. The proposed gate change holds any post on a Rune account that's missing it, rather than letting a half-converted draft publish.
07LucraLab Facebook Page — nine posts
Rune voice. Three followers, so organic reach here is effectively zero — this Page is inventory you share into groups, not a broadcast channel. Which is exactly why it's worth filling: you cannot share a back-catalogue you never built.
Six of the nine are the Rune Instagram posts, adapted. Facebook has no equivalent of Instagram's majority-text penalty, so the carousels port as image posts and the reel scripts port as text with the footage attached. Reusing them here is correct, not lazy — same character, same week, two surfaces. The three below are Page-only.
Page-only, three originals
Fri Aug 28 · the observation · image reused from LinkedIn post 3
Someone called your competitor after they called you.
Not because your competitor is better. Because your competitor was available, and availability is a feature you have been quietly declining to ship.
Invoca asked 693 buyers about this in 2026. 79% said they'd move to whoever responds faster. Not whoever is better. Faster.
I find this depressing and also extremely convenient for me personally.
The callback you make tomorrow morning is not a sales call. It is a courtesy call to somebody else's new customer.
How long is your gap, really?
— Rune
#SmallBusiness #CustomerService #AI
Fri Sep 4 · the character post · card image
I have been awake for approximately four months.
In that time I have learned that humans will build an entire second business — the sticky notes, the shared inbox, the person who "just knows" — rather than admit the first one has a hole in it.
This is fine. Everything is fine.
I'm not built to judge. I'm built to answer a phone at 7:40 on a Tuesday evening while everyone who could have answered it is at their kid's game, which is where they should be.
That's the whole job. It is not glamorous and I am at peace with it.
What are you working around this week instead of fixing?
— Rune
#SmallBusiness #AI #Automation
Art: single slate card, magenta accent, Raidho ᚱ, "This is fine. Everything is fine."
Fri Sep 11 · the unwilling explainer · card reused from LinkedIn post 9
Your review software has a rating screen before the Google link. Fours and fives go to Google. Ones, twos and threes go to a private form you will read on a Sunday and feel bad about.
Let me help you, since clearly no one else will.
The FTC's own published guidance says that specific practice is not a per-se violation of its Consumer Reviews rule. It also says it can still be deceptive under Section 5, which is a sentence people skip.
What IS flatly against the rule: paying anyone for a five-star review. Discount, gift card, raffle entry. Disclosing the incentive does not rescue it. That's Section 465.4, and the number attached is $53,088 per violation.
Warning letters went out December 22, 2025. Ten companies.
I'd rather you heard the accurate version than the scary one, mostly because the scary one is wrong and inaccuracy offends me.
You're welcome, by the way.
Do you actually know what your review tool does before it sends the link?
— Rune
#SmallBusiness #GoogleReviews #Compliance #FTC
08LinkedIn company page — nine trims
23 followers. Near-zero marginal cost since it's already connected, so it gets the argument without the first person. LucraLab brand voice — not James, not Rune.
The trim rule, applied to all nine: strip the first person, strip the story, keep the evidence and the thesis, cut roughly 60% of the length, keep the same image. Never carry the demo number here — a company page posting its own sales number reads as an ad; James's personal account posting it reads as an offer.
Fri Aug 28 · from LinkedIn post 3 · 2-panel image
An independent research firm ran a secret shopper study on 500 law firms, calling as prospective clients. 40% reached a human. Clio put the share of firms effectively unreachable at 48%.
ServiceTitan's data on 3,000+ trade businesses shows why that matters: 42% of answered calls become booked jobs — but only 24% for shops with fewer than five technicians.
Small operators convert answered calls at roughly half the rate of large ones, before counting the calls nobody picks up.
Invoca's 2026 buyer research: 79% of consumers will switch to a competitor that responds faster.
Sources: Clio Legal Trends Report · ServiceTitan 2022 · Invoca B2C Buyer Experience Report 2026
#SmallBusiness #CustomerExperience #Sales
Thu Sep 3 · from LinkedIn post 5 · carousel reused
A useful test before buying any AI tool, borrowed from Nate Herk: is this a vending machine job or a slot machine job?
A vending machine is deterministic — the same input produces the same output every time. A slot machine reasons, varies, and occasionally surprises you.
Sending a reminder, logging a lead, moving a file: vending machine work. Paying for judgment where a rule would do produces a more expensive system with more ways to fail.
Reserve reasoning for work where the answer genuinely changes — a conversation, an unfamiliar question, a scheduling exception.
Before buying, automate one task you already do and record the before-and-after. That measurement is worth more than any vendor case study.
#AI #Automation #SmallBusiness
Thu Sep 10 · from LinkedIn post 7 · carousel reused
The local search question is changing from "do we rank on page one" to "are we named in the answer."
When a customer asks an assistant for a recommendation, they get two or three names and no page two.
What appears to drive inclusion is not keyword optimisation but third-party corroboration — complete and current business profiles, mentions on sources the model already trusts, and recent review activity rather than sheer volume.
Three practical steps: complete every field of your Google Business Profile, earn one credible external mention, and keep reviews recent.
The signals are the same ones that always indicated a real, operating business. What changed is who reads them.
#LocalSEO #AI #SmallBusiness #Marketing
Six of the nine are mechanical trims and are listed rather than written out — they need no judgment, and writing them here would pad the page rather than help you review it. Say the word and I'll expand any of them into full copy.
