GarthBox

I was in a strategy session yesterday, getting cheerfully interrogated about positioning, when I heard myself tell a story I hadn’t reached for in years. A man named Johnny Earle opened two bakeries. No cupcakes. You walk in, the air is warm and smells like vanilla and sugar, you pull open an oven, and inside there are t-shirts. He calls them fakeries. People came in for cake. They left with a shirt and a story they could not stop telling at parties.

The safe read on this is “be quirky, be memorable.” You should ignore that read. It is the read that takes a strange, useful idea and flattens it into a fridge magnet.

The shirts did not sell because the shop was cute. They sold because of the half-second of vertigo between expecting cake and getting cotton. That little gap, that small honest betrayal of what you promised, is the actual product. The shirt is just the thing you carry the feeling out in.

The gap has a name nobody likes to say out loud in a pitch meeting. Tension.

We have spent a decade training the tension out of our work and calling it craft. Smooth onboarding. Frictionless checkout. A homepage so balanced it reads like a hostage statement. And it works, in the sense that nothing breaks and nobody complains.

It also fails, in the sense that nobody remembers you the next morning. Comfort is what people say they want. Friction is what they actually buy, because friction is the only thing that leaves a mark, and memory is just the catalog of marks.

I watched this play out in real numbers this year. The companies pulling away in crowded categories were not the ones with the cleanest value prop. They were the ones whose salespeople opened with some version of “here’s what I think you’re getting wrong,” and only then, once there was a little heat in the room, got around to what they sell.

A CMO I trust told me the thing she loves most in a meeting is being challenged, because at least then something is at stake. Everyone else was sending her case studies. She was deleting them on her walk from her office to the kitchen.

Here is the inconvenient part… Tension cannot be A/B tested into existence.

You can optimize a button color. You cannot optimize courage.

A test will always, given enough rounds, walk you toward the blandest option that nobody hated, because “nobody hated it” is exactly what an average is good at finding.

Tension requires a human being willing to leave a bruise on the page and defend it on Monday. It is a decision, not a result.

I know this because I got caught not doing it.

In that same session I admitted, out loud, that I did not recognize the headline on one of my own landing pages. We had wired up a slick little loop that kept rewriting the copy toward whatever read cleanest, and it did its job beautifully. Every pass sanded a hair more grip off the sentence. Correct, then polished, then smooth, then gone. I had automated a brand voice into a pleasant hum I could no longer pick out of a lineup. The machine was not malfunctioning. Smoothness is its native gravity. Anything left running long enough without a thumb on the scale drifts toward the absence of friction, because friction is the one thing it cannot generate and does not miss.

That is the quiet danger in all of this, and it has very little to do with robots.

The safest sentence in your marketing is usually the one doing the most damage, precisely because it is too agreeable to argue with. It survives every review. It offends no committee. It dies on contact with a real human’s attention and nobody attends the funeral.

We’d be better off building more sentences (ovens?) without the cake.

Stay Positive & “Smooth” Is Simply Slow Disappearance

A director of product marketing told me recently what actually fires her up at work. Not the launch, not the dashboard turning green, not the win itself.

Being challenged.

Somebody walking into the room and leaning on the way she thinks until it bends. She said we should always be challenging our way of thinking, and she said it the way a person names the one meal they’d eat for the rest of their life.

That stuck with me, because the entire business world is selling the opposite. Frictionless onboarding. Seamless handoffs. Zero resistance, zero drag, zero reason to feel anything.

We have built an economy around the removal of tension and then act surprised when so much of the work feels like wet cardboard.

Tension is not a problem to be solved. It is the thing that tells you you’re still alive in the work.

I sent a SaaS founder a point of view on his AI agents, and the part I knew would sting was about their names. Naming is tender ground. You name a thing and you’ve decided what it is, so telling someone the name is off is telling them they saw the thing wrong. I could have padded it. I sent the tension instead, because a point of view with the tension taken out is a greeting card. Pretty, agreeable… forgotten by lunch.

Tension shows up in a lot of costumes. The idea that might not work. The person who presses on how you think. The funny one stuck in a serious room, or the serious one stranded in a funny one. The gap between what we swear is possible and what we quietly do to make it feel possible.

They don’t all serve you the same.

The job is to look at the ones in front of you and pick the one worth the discomfort, the way you choose which hill you’re actually going to run up.

And when none of them show up, you build one.

My wife and I have planted fairies in our daughters’ fairy garden after they fell asleep. Out on the lawn near eleven, pressing little wings into the dirt so the morning would hold proof that something magic had come through in the night. The girls believe. We stage the belief. That isn’t a lie, it’s labor.

Tension works the same way.

If the friction you need won’t arrive on its own, you go out in the dark and you make it, on purpose, with your own hands, for the version of you that has to wake up and keep going.

The harder kind is the tension you didn’t stage. The challenge you never asked for, the POV that lands on you instead of coming from you. That one is more uncomfortable and almost always more useful, and the only real move is to let it in instead of smoothing it over.

Seamless feels like a compliment. Mostly it’s a warning.

Stay Positive & Pick Your Tension (Or Go Out On The Lawn And Plant One)

I got a tattoo yesterday, and somewhere around an hour and a half I realized I was being cared for by an expert, and that none of it was on purpose.

She took her time. She checked in. She geeked out about manga research and tracing she’d done that I never asked for and would never have known about. When a younger artist asked her for feedback, she gave him all of it. When a different one didn’t ask, she leaned over anyway. “Oh, bamboo? I can fuck with bamboo.” The guy lit up like she’d handed him something. She told me early, before I could even think to worry, that she sometimes forgets to knock the deposit off the final, so hold her to it. She pointed me to the water like someone who’d genuinely prefer I not faint on her table.

And the small one, the one that stuck. I passed her on the way to the bathroom and caught her checking her teeth in the mirror. “Gotta make sure I don’t have anything in there.”

None of that was a move. That’s the whole thing.

You cannot fake the signal layer. Care isn’t a posture you decide to assume at the start of a job. It’s a leak. It escapes you in a hundred small involuntary tells, and that leak is exactly why we trust it. The tells are the most honest data in the room because nobody chose to send them.

Brands know the tells are valuable, so they try to manufacture them. The “seamless experience.” The handwritten font that was never near a hand. The bot that chirps “happy to help!” with a little heart on the end. We clock the forgery before we can explain why, because manufactured care has a tell of its own, and the tell is that it’s pointed at us. Real care isn’t pointed anywhere.

It just seeps.

I felt it from the other chair last week. I built a mock pitch to a CISO and a CRO for a company that does not exist, and I had an embarrassing amount of fun inventing the thing. Researched the fake company like it owed me money. Burned a whole day (and my claude tokens) building prototypes so I’d have something live to show instead of something to describe. And the one that actually hurt: I recorded myself and listened back, which is its own quiet horror. They felt it on the call. Not because I told them I cared. Because it leaked out of me the same way it leaked out of my tattoo artist.

The exercise here is not a gentle one.

Take a piece of work, one you’re proud of or one you’re avoiding, and write out the signals.

Not the deliverable.

The tells.

The research nobody asked for. The third draft no one will ever see. The water you pointed to.

The count of signals you land on is the most accurate read you will ever get on how much you actually cared, and it does not negotiate. You can’t talk your way to a higher count. Your gut won’t let you.

Stay Positive & Did You Check Your Teeth Today?

Every list we make does two things to us at once.

It hands us a weight. Twelve things, and now we owe twelve things, and somewhere around item nine we’ve already started rehearsing the apology we’ll give ourselves tonight for the ones we didn’t touch. That’s the accountability talking. Its quieter cousin is guilt, and guilt shows up whether or not we earned it.

But the same list does something kinder in the same breath.

It tells us what matters. It puts the four real things near the top and lets the other eight sit there like furniture. The relief of that is hard to name. It’s the grace of not having to do all of it, handed to us by the very page that’s also trying to make us feel bad.

One object, two weather systems.

For a long time we think the guilt means we’re making the list wrong. Too ambitious. Bad at estimating our own day. So we try shorter lists, honest lists, lists we can actually finish. And we finish them and feel nothing, because a list you complete to the letter is a list that never asked you to choose. No tension = No bliss.

That changes the moment you watch a sales team work, of all things.

No good sales org builds a pipeline equal to its number. They build three or four times the number. If you need to close a million, you carry three or four million in deals you know will not all land. Nobody walks the floor at quarter’s end pointing at the deals that didn’t close and calling them failures. The extra pipeline isn’t debt. It’s the only way to know which deals deserve the phone call. You over-set the top so the bottom can sort itself.

A list is personal pipeline.

We write twelve because writing twelve is how we find the four. The eight we don’t get to aren’t a tab we owe. They’re the cost of knowing. We paid in ink for the privilege of prioritizing, and prioritizing only happens when there’s more on the page than there is in the day.

Which means the guilt has been reading the wrong document. Guilt treats the list like a contract, a thing we signed and then breached. It was never a contract. It’s a forecast. A forecast of a possible self on a generous day. You don’t breach a forecast. You compare against it, you learn the shape of your real capacity, and you forecast again tomorrow with slightly better data.

The trouble is we hand this grace to spreadsheets and almost never to ourselves. We let a sales team carry four times its quota and call it discipline. Then we carry a twelve-item Tuesday and call ourselves behind. Same math, different mercy.

If we’d coach a team to build a pipeline bigger than they can close, we should coach ourselves the same way, and then coach the people who work for us that way too, because an unfinished list is not a character flaw. It’s evidence we aimed.

Stay Positive & Aim Past The Number

I had it half-written in my head. Storyboard a short film a five and three-year-old could shoot. Six scenes. Teach framing and motion along the way. Twenty seconds of typing and I’d have had a clean little plan to hand my daughters, complete with shot angles and a suggested color palette.

My hand was already drifting toward the keyboard. Then it stopped, the way you stop at the top of the stairs when you’ve forgotten why you climbed them.

So I ripped the back cover off a coloring book (because we didn’t have any scrap paper). Markers that had mostly given up. A ruler none of us had patience to find, so the frames came out crooked. By the time we finished the first scene, one of six, I felt a strange low hum of guilt that I had even considered the other version.

I want to be precise about the guilt, because it surprised me. And quite frankly, this is therapeutic.

It wasn’t that I’d almost robbed them of something. They’d have been fine either way. It was that I’d almost robbed myself of something I didn’t know I needed until it was already happening. The arguing about what object Emree should have in her frame. The bad ideas that made us all laugh. The slow, inefficient, irreplaceable feeling of a few people making a thing that did not exist before they sat down.

Now here’s where I’ll defend the machine, because I’m not a purist and I’ve seen it earn its keep.

I ran win/loss software that leaned on AI to generate interview questions and talk to prospects and customers. And it worked. It pulled data we could actually act on, at a volume and consistency I could never have hit by hand. That was the right call. That was the boring, necessary, high-volume eighty percent of the work, and handing it to a system was not a betrayal of anything. It was good judgment.

But the line I carried into every meeting after that did not come from the software.

It came from an IT director sitting across the virtual table from me, talking about a tool his company had finally adopted. He said, “It was finally something that didn’t make me want to vomit.”

I have quoted that man in many-a-meeting. It resonates so thoroughly every time.

No survey produced it. No model would ever generate it. Ask an AI to write customer praise and you get “intuitive,” “seamless,” “a game-changer.”

You get the center of the bell curve, sanded smooth. You do not get nausea reframed as a compliment.

That sentence was too specific, too a little bit gross, and too true to have been predicted by anything optimizing toward the average.

That’s the whole thing, and it’s a marketing point as much as a parenting one. A model’s job is to find the mean and return it to you, fast and clean. But the mean has never sold anything memorable.

The line that closes the deal, the detail that makes someone trust you, the moment that sticks, those all live in the outliers. The weird specific real ones. And outliers are not something you retrieve. They’re something you make, usually by hand, usually inefficiently, usually in the part of the work you were most tempted to skip. (It makes me think of a sales leader I heard the other day say “You don’t want your prospect-facing reps prompting, you want them selling!” Which, what I heard is you want them being the human outlier in the sales process.)

The messy parts and the important parts keep turning out to be the same parts.

Some people will tell you the mess is the whole reason any of it is worth doing. I’m starting to think they’re right.

Stay Positive & Crooked Boxes Might Just Contain The Best Ideas

A motorcycle track coach taught me the most useful thing I know about business, and he did it by making me afraid on purpose.

Going into a corner at speed, there is a moment where your whole body screams brake. Not a thought. A flood. The coach watched me obey it for a few laps, then handed me a strange instruction. When the fear hits, notice it. Name it. Count to three. Then turn in.

The first time he said it, I thought he wanted me to be braver. He didn’t. He wanted me to learn that the fear was early. It always was. It arrived a full three seconds before the corner actually needed anything from me. My fear was not lying about the danger. It was lying about the time.

Last week I was sitting with the founder of a SaaS company, talking positioning, and I felt the old flood come back. That nervous spike that says you are about to not be impressive, so soften it, hedge it, give him the tidy version he probably came in expecting. Brake now.

I counted. Then I said the thing I actually believed, which was that I wouldn’t touch his positioning for where the company sits today. Positioning is a hockey instinct. You don’t pass the puck to where the player is standing. You pass it to where the player is going to be. He went quiet, then leaned in. It paid off.

Now the honest part, because the clean version of this story is a lie. I did not find my braking point by being wise. I found it by running off the track.

Same corner, a few laps in a row, holding the pause a beat too long, sailing wide into the grass with the throttle full of regret. Three is not a magic number. Three is a search. You overshoot, you clip the apex late, you adjust, you try again. The seconds were never the lesson. The braking point was.

And once I had it on one corner, I could carry the same motion into the next one.

Fear is a now signal. The alarm is real. Its sense of timing is garbage.

Stay Positive & Count…Then Turn In

We had a sour on tap last spring that nobody ordered. Good beer. A Berliner Weisse, listed on the board as exactly that, “Berliner Weisse, 3.8%,” sitting there honest and ignored while it slowly went flat in the line. Two weeks of almost nothing. Then I rewrote it as “Overfruited Sour” Same keg. Same liquid. Same brewer, same recipe, same hands that made it.

You know this is coming…It sold out that weekend.

I didn’t improve the beer. I stopped describing it like a spreadsheet.

I keep running into the software version of this. There’s a whole category of products out there quietly doing extraordinary work under a name that makes them sound like a checkbox. Think about a piece of software that watches a job site all day. It notices when someone arrives. It clocks them in without being asked. It tracks where they go, trims the time card when they leave, sorts the cost back to the right project, and fixes the small human errors nobody had the patience to catch. It does this every day, in the background, for thousands of people, and it does it well.

For years a thing like that gets sold as “GPS time tracking.” A feature. A line item with a little tick next to it, shelved next to a dozen cheaper tools that do a worse version of the same chore. Buyers compare the tick marks, shrug, and pick on, who knows, probably price. The product was years ahead of its category and the category had it filed under office supplies.

Then a new word shows up in everyone’s mouth. Agent. And suddenly the exact same software, not one line of code different, is the future. The thing was already monitoring, already deciding, already acting on its own. It was an agent the whole time. It just never had permission to say so, because the word didn’t exist yet in a way a buyer would pay for.

That’s the part worth sitting with. AI didn’t upgrade the product for a lot of SaaS out there. AI upgraded the vocabulary. The capability was real for years. What changed is that the market finally had a shelf to put it on, and a shelf is the only thing a market can actually pay for. People don’t buy value. They buy value they can classify.

If they can’t file it, they can’t fund it, no matter how good it is.

This is the quiet tax almost nobody accounts for. Not a bad product. A great product wearing a cheap noun. You can be objectively the best thing in the room and still lose to a worse thing that picked a better word, because the word is what the buyer holds in their head on the drive home.

And it isn’t only products. I’ve watched the same thing happen to people. The smartest operator on a small crew is usually doing some essential job that has no name, the one everyone calls “the stuff nobody wants to deal with.” It’s invisible until somebody finally calls it a role, gives it a title, says this is yours and it matters. (Have you wondered why GTM Engineer has gotten so popular?! We’ve been here doing the work. Now there’s a title for it.) The work didn’t change. The person didn’t change. But you handed them a noun they could grow into, and they grew.

Naming isn’t decoration. It’s permission. You can do that for a beer, for a feature, and for the person sitting three thousand miles away who’s been carrying something heavy under a label that made it sound light.

The beer was always that good. I just stopped underselling.

Stay Positive & …Until The Next Cycle