Insights
Notes on software that fits, from someone who runs the businesses it runs.
Buy versus build, what an agent can carry before it needs an app, and what happened when I tried it on my own restaurant first.
Start here
Fall in Love with the Problem, Not the Fix
Restaurant owners are natural fixers. But jumping to a solution before you understand the real problem is how good operators waste time and money.
Read →Test Your Ideas Before You Bet on Them
The best restaurant operators don't just go with their gut. They test small, track what happens, and let the results tell them what to do next.
Read →Measure Twice, Cut Once
The biggest mistakes in a restaurant usually happen when you make a big decision too fast with too little information. Here's how to make cheaper mistakes.
Read →Say What the Guest Gets First
Most restaurant and hotel marketing lists what the business has. Guests respond to what they get. A short exercise turns features into posts people act on.
Read →Who Actually Decides Where to Eat?
The person eating the meal and the person who picked the restaurant are often two different people. Understanding both changes how you market your place.
Read →Stop Watching Your Competition
That new restaurant down the street is doing well. So what? Your guests will tell you more about how to grow your business than your competitors ever will.
Read →Everything, newest first
- The Smallest Fix That WorksFixing a broken process rarely starts with a platform or an app. It starts with a prompt, then a workflow, and only sometimes ends in something built to fit.
- You Don't Need a Marketing Department to Stay on BrandMessageKit keeps a two-restaurant brand consistent on social, email, and text without a marketing hire. It's in early access and free to start.
- Four Steps to a Branded PostMost social media tools are either so loose posts look inconsistent or so rigid nothing spontaneous gets out. A four-step wizard finds the middle.
- One Row in a Spreadsheet. Three Social Platforms. Zero Human Touches.Startup social media is death by a thousand small tasks. I automated the copy, images, tone, and scheduling with n8n, Claude, and Buffer's MCP.
- Building Winote: A Tasting Notes App for Working SommeliersSommeliers taste 30 to 40 wines at a sitting, and every wine app is built for collectors. Here's how I thought through a tasting notes app for professionals.
- AI Won't Save You From Building the Wrong ThingAI-powered lean teams are the new flex in Silicon Valley. But the real advantage isn't fewer people — it's faster learning.
- Fall in Love with the Problem, Not the FixRestaurant owners are natural fixers. But jumping to a solution before you understand the real problem is how good operators waste time and money.
- Stop Watching Your CompetitionThat new restaurant down the street is doing well. So what? Your guests will tell you more about how to grow your business than your competitors ever will.
- How Your Guests Decide What to Pay ForWhy will someone pay $18 for a burger at one restaurant and think $14 is too much at another? It comes down to how they feel, not just what's on the plate.
- How to Think About Menu PricingPricing your menu is one of the hardest parts of running a restaurant. Here's a simple way to think about it that goes beyond just food cost.
- Who Actually Decides Where to Eat?The person eating the meal and the person who picked the restaurant are often two different people. Understanding both changes how you market your place.
- Your Regulars Are Your Best AdvisorsThe best feedback about your restaurant isn't on Yelp. It's in the heads of the people who come back every week.
- Test Your Ideas Before You Bet on ThemThe best restaurant operators don't just go with their gut. They test small, track what happens, and let the results tell them what to do next.
- Good People Can't Fix a Missing ProcessGreat people and a great product don't add up to a great process. Five tests show whether a checklist or routine will hold up on a busy Saturday.
- Too Many Choices Will Cost You CustomersWhen people can't decide, they don't decide. That's true at the grocery store, and it's true at your restaurant.
- Measure Twice, Cut OnceThe biggest mistakes in a restaurant usually happen when you make a big decision too fast with too little information. Here's how to make cheaper mistakes.
- Listen Like Your Business Depends on ItThe tech world calls it 'human-centered design.' In restaurants, it's simpler than that: pay attention to what your guests are telling you, and act on it.
- Say What the Guest Gets FirstMost restaurant and hotel marketing lists what the business has. Guests respond to what they get. A short exercise turns features into posts people act on.
- Three Numbers Your Booking Page Should Report Every WeekYour booking page is the front door for most guests. Three plain numbers show whether that door is open, stuck, or locked, and what to ask your vendor.
- The Psychology Behind Your MenuYour guests don't make rational decisions about what to order. They follow patterns. Understanding those patterns can help you design a menu that works.
- Simple Document CollaborationThe everyday problem nobody's solved well: simple, shared, threaded document collaboration that works for people who live in email.
- You Can't Fake What Makes Your Restaurant SpecialThe best marketing for your restaurant isn't something you can buy. It's the real stories your guests tell when they leave.
- User Feedback and Product DesignHow much user feedback should shape product design — and when to seek it. Ask early, ask often, vary your audience.
- Design CorrosionSoftware often gets worse as it gets bigger. The slow corrosion of design-by-committee, and how to keep your guideposts visible.
- The Paradox of EasyMaking things easy on one side of the counter makes them harder on the other. Tip the scale toward the guest, then use better tools to help your team.
- Handle It OnceAn owner's inbox never stops: vendors, invoices, catering requests, reviews, staff texts. Touch each item once and send it to one of three places.
Reading is free. So is the conversation.
Twenty minutes on the one thing your software won't do.