Notes from the field

Operator-grade observations, straight from the build.

Short reads on campaign patterns, automation reality, and what we're learning while building KaiNet. No clichés, no hype. Just what's actually working at the backend of ads at scale.

Automation realityMay 1, 2026 · 4 min
We tried 14 'AI marketing tools.' Most are wrappers. Here's what actually moved the needle.
Most AI marketing tools on this year's lists are GPT calls with a logo. We tested 14 of the popular picks across content, ads, and analytics. Five did real work. Here's what separates them.
MediumRead post
Operator lessonsApr 30, 2026 · 6 min
The marketing stack we'd actually buy if we were starting today
Five tools, around $400 a month, covers acquisition through measurement for a marketing team in 2026. Here's the stack we'd buy from scratch and why each one earns its slot.
Deep diveRead post
Campaign patternsApr 29, 2026 · 4 min
Performance Max is no longer the right default. AI Max is.
Google's AI Max launched broadly in early 2026 and quietly displaced Performance Max as the right default for most accounts. Here's what changed and how to decide which campaign type fits.
MediumRead post
Operator lessonsApr 28, 2026 · 4 min
HubSpot, Mailchimp, Klaviyo: stop comparing. They're not the same product.
These three tools end up on every marketing comparison post as if they're direct competitors. They aren't. The 'comparison' is meaningless until you know which one matches your business shape.
MediumRead post
MeasurementApr 26, 2026 · 4 min
Semrush vs Ahrefs in 2026: the question changed in January.
For most of the last decade the answer to 'Semrush vs Ahrefs' was 'pick whichever workflow you like better.' In January 2026, Ahrefs reshuffled the deck with a $29 starter tier. Here's how to actually pick now.
MediumRead post
Automation realityApr 25, 2026 · 5 min
What 'AI marketing tools' actually means (the four categories that matter)
Every 'best AI marketing tools' list throws 30 products together that do completely different things. Here are the four real categories, what each one does, and how to know which one your team needs.
Deep diveRead post
Campaign patternsApr 22, 2026 · 6 min
What manually building 100+ campaigns taught us about high-performing ads
Most performance is decided before the campaign goes live. Optimization can take a campaign from good to better. It almost never takes one from broken to working, and we've watched a hundred teams learn that the hard way.
Deep diveRead post
Operator lessonsApr 21, 2026 · 4 min
How to audit your marketing stack in 30 minutes (and probably cut ~30% of the cost)
Most marketing teams pay for tools they barely use. Here's the 30-minute marketing stack audit we run on every account that comes to us. Most teams cut 30% of their tool spend without losing any capability.
MediumRead post
Campaign patternsApr 19, 2026 · 4 min
We reviewed 100 campaigns. These patterns showed up everywhere.
Not every weak campaign fails the same way. But after auditing roughly a hundred of them, the failures fall into four shapes. It's almost boring how often we see the same patterns, which is also why they're fixable.
MediumRead post
Operator lessonsApr 15, 2026 · 7 min
Most bad campaigns don't fail in optimization. They fail in setup.
We've watched teams spend three weeks adjusting bids on a campaign that was never going to work, because the original setup made success structurally hard. Here's the pre-launch checklist that catches most of it, and how to use it without slowing the team down.
Deep diveRead post
Automation realityApr 13, 2026 · 2 min
The pricing pattern hidden in every 'AI marketing tool' demo
Every AI marketing tool has the same shaped pricing curve. Free tier, then a 10x jump at the threshold most teams hit by month two. Here's the pattern and why it's structural.
Quick readRead post
MeasurementApr 12, 2026 · 3 min
CTR matters. But it hides more than most teams think.
A high CTR means the hook worked. That's the whole signal. It does not mean the traffic was qualified, the offer fit, or the campaign was working. We have three real scenarios where that distinction cost teams real money.
MediumRead post
Automation realityApr 8, 2026 · 2 min
AI won't fix a broken offer. But it can expose weak campaign structure fast.
There's a hopeful read on AI in ad operations: feed it data, get growth. It's a comforting story, and it's wrong. What automation actually does is surface bad inputs faster, which turns out to be more useful.
Quick readRead post
Builder diaryApr 4, 2026 · 4 min
What building KaiNet taught us about ad quality at scale
We started building KaiNet thinking the hard part would be generation. It turned out to be the easy 30%. The real work, the unsexy work, was figuring out what 'quality' actually meant when ads have to ship to a live account.
MediumRead post
Automation realityMar 31, 2026 · 2 min
The part of campaign automation no one talks about: QA debt
Faster campaign creation also means faster campaign mistakes. The QA layer that should have grown alongside the generation layer usually doesn't, and the debt accumulates quietly until something breaks loud.
Quick readRead post
Builder diaryMar 27, 2026 · 2 min
We're not interested in 'AI magic' for ads. We're interested in better systems.
The category is full of demos that promise magic. Type a sentence, get a campaign. Wave a wand, get growth. We've watched a lot of those. We're not building one.
Quick readRead post

Operator-grade observations, straight from the build.

We tried 14 'AI marketing tools.' Most are wrappers. Here's what actually moved the needle.

The marketing stack we'd actually buy if we were starting today

Performance Max is no longer the right default. AI Max is.

HubSpot, Mailchimp, Klaviyo: stop comparing. They're not the same product.

Semrush vs Ahrefs in 2026: the question changed in January.

What 'AI marketing tools' actually means (the four categories that matter)

What manually building 100+ campaigns taught us about high-performing ads

How to audit your marketing stack in 30 minutes (and probably cut ~30% of the cost)

We reviewed 100 campaigns. These patterns showed up everywhere.

Most bad campaigns don't fail in optimization. They fail in setup.

The pricing pattern hidden in every 'AI marketing tool' demo

CTR matters. But it hides more than most teams think.

AI won't fix a broken offer. But it can expose weak campaign structure fast.

What building KaiNet taught us about ad quality at scale

The part of campaign automation no one talks about: QA debt

We're not interested in 'AI magic' for ads. We're interested in better systems.