The order you tell people

Why order

Marketing is not a skills problem so much as an order problem. Spend on ads before the product stands up and the money evaporates. Make content before you've decided how people find you and it reaches no one. Ship without a price and there's no reason to promote anything. The twelve steps below are arranged so that each one only makes sense once the previous one is done.

Source: rules from restaurants and retail spaces, learned from space planner Gongjion's Offline Ground channel, translated by us into software terms. We've carried over the substance rather than the wording, and where it's been adapted for apps that reading is ours. Restaurants get graded on being wrong by closing, so what survives there has been tested hard.

The item is half the outcome. Restaurant people say "the revenue is decided the day it's born." Change it later and it costs several times over.

Pick something you like from among the things people like. The other way round shrinks the market to your taste.

The sharper your taste, the fewer people share it, and that number is your ceiling. Build something you'd never use, though, and you'll quit at the hundredth revision. Choose from the overlap.

• Written out breadth, frequency and durability separately
• Judged whether I'd still use it on the 200th day
• Chose from what people want, not just what I want

The same product faces a completely different opponent depending on whether it sits in the everyday zone or the occasion zone. And people behave in opposite ways across that line.

Where they spend little, people are more conservative. Where they spend a lot, they're more adventurous.

A meal you're just getting out of the way, you don't want to get wrong, so you pick the brand you know. For an anniversary you're bored of the usual place, so you look for something new. Which means the everyday zone belongs to the big brands. Individuals who meet them head-on there usually lose.

There is exactly one exception for individuals in the everyday zone: offering one thing that is certainly guaranteed. Feature count isn't what wins here. It's one promise, kept so reliably that the promise itself becomes the safe choice.

• Written in one sentence which zone this sits in
• If everyday: named the one guaranteed thing that beats a big brand
• If occasion: named the reason someone makes the trip

An empty market looks like a gift. Usually it's the spot people retreat to when they don't fancy the fight.

Taking money someone already meant to spend, and making them spend money they didn't have to, are different jobs.

An empty market means creating the demand from scratch. And if you succeed, whoever has more capital walks in and takes it. Successes get written up; failures disappear quietly, so from outside all you see is the successes.

You're usually better off entering a proven market and twisting one axis inside it. Keep the category familiar; change exactly one of viewpoint, rule, or audience. And if difference is your only advantage, it stops being an advantage the moment you're copied.

• Two or three live competitors exist here (demand is proven)
• Named the specific competitor I believe I can beat
• The twist is one thing: viewpoint, rule, or audience
• Written down what remains if I'm copied

In food, people say 70% of a thriving business is location. Taste is necessary, not sufficient. In software, location means how people arrive.

A lease can't be changed once signed. For an app, the name, subtitle, category and first keywords are that lease.

Heavy footfall that isn't your customer is a dead location. A keyword with clear intent beats a bigger one, for the same reason purpose beats place name. "Restaurant for meeting the in-laws" converts better than "restaurants downtown".

The real cost of a bad location is misdiagnosis. With no arrivals you can't tell why things aren't working, so you keep fixing the wrong thing and fixating on single reviews. A good location gives you fast feedback.

• Ten target keywords written down and fixed
• Checked review counts and last-update dates of the top three
• Chose keywords by intent, not volume
• Treated name, subtitle and category as unchangeable

The name is a hard gate you clear before starting. Later is too late, and too late is expensive.

Search it and only you should come back. The closer it is to a common noun, the deeper it's buried.

Claiming a name in a store is not a trademark. The store name belongs to whoever registered first, but a trademark holder can make you take it down. Get the order wrong and you lose it after you've grown it.

• Web and store search return essentially nothing
• Checked registrability, not just availability (generic terms get refused for lack of distinctiveness)
• Domain and social handles are available
• Trademark application filed
• It sounds natural inside a sentence ("let's use ___")

Price isn't a number you attach at launch. It's part of the design. With no price there's no reason to promote and no case for spending on ads.

Pricing is decided by the person who built it. One of the few jobs you can't hand over.

Cost starts after launch. Building isn't the expense; keeping it alive is. It's why "30% margin" is mostly fiction in food. Take out commission, tax and hidden labour and far less remains.

Ship something with metered costs for free and the losses grow automatically as it succeeds. Don't build a structure where success is the same as bleeding.

Two things when you price. First, every category has a psychological ceiling, usually set by the market leader. Second, make the customer feel they're winning the game. "They charge that for this?" is the reaction you want.

• Calculated what's actually left after commission, metered spend and maintenance
• If there's metered spend, a price or a usage cap exists
• Researched the category's reference price
• Wrote the trigger for raising it in advance ("can I still hold this quality?")

From here on you're telling people. One check remains first.

Uncomfortable, but true: in the short run a weaker product that's well told beats a strong one nobody hears about. So the answer is both, and the order is product first. Wanting the money sooner is exactly what makes people spend on ads too early.

• A first-time user reaches the core value without explanation
• I'd recommend this to someone without hedging
• The difference fits in one sentence a user can repeat to a friend

Marketing spend splits into two kinds that behave nothing alike. Miss the distinction and you'll pay the same money every month forever.

Ads are a cost and they melt. Your own channel is an asset and it stays.

Ad prices rise every year while efficiency falls, and the harder the economy the fiercer it gets. So buying exposure without a weapon just melts money. Someone who arrives via your content, on the other hand, arrives already convinced. Explanation, enquiry and conversion all get faster.

Ads bought in a panic don't work. The asset you built when you had room is what works when you don't. Reverse that and you use the most expensive method at the most expensive moment.

• Picked one channel (several means all of them die)
• Split this month's spend into cost and asset
• Confirmed assets aren't zero before buying any ads

This is where two different things get confused. Content that reaches and content that sells are not the same object.

Separate reach pieces from conversion pieces, roughly ten to one.

Post only the selling kind and people leave. Nobody wants the friend who pitches every time you meet. Ten useful or enjoyable things, one that sells.

Planning is skill; view count is luck. The piece that exploded and the one that didn't usually took the same time to plan. Accepting that split is what lets you keep going.

A format that isn't yours can't be sustained even when it works. Plenty of people manufacture a hit in a style they can't stand and quit. Ask first whether you could keep doing it if it succeeded.

And without a forcing device it will stop. A private resolution that isn't a routine breaks. Fix the day and time in a calendar, and where you can, attach something with stakes: a partner, a forfeit, a paid commitment.

• Set this month's ratio of reach pieces to conversion pieces
• Confirmed I could sustain the format if it worked
• Day and time fixed in the calendar
• The material includes the story only we can tell
• Publishing the build in advance, so there's an audience on launch day

Being known doesn't arrive gradually; it arrives all at once, one day. The problem is that if you're not ready, that wave is your last one.

Reputation doesn't decay. It accumulates. One feature ends and fades; what has piled up keeps pulling more in.

When the wave lands, throttling arrivals to protect quality breaks the flow. Quality collapsing under a crowd is a lack of preparation, not a problem with the crowd. An opening that comes once has to be pushed all the way.

One more: being known in your own industry is not being known to customers. Peers recognising you feels good and doesn't pay.

• Verified it holds at ten times the users
• There's a path for handling a pile-up of enquiries
• First screen and store copy are written for a first-time reader
• Checked the route reaches customers, not peers

Pulling numbers is staff work. Deciding what to do about them is the founder's.

Write ten causes, then fix only the two or three you control.

There are always plenty of reasons it isn't working: the economy, the season, platform policy, competitors. All true, all outside you. Every hour spent on what you can't control is an hour not spent on what you can.

If the underlying records are contaminated you can never improve anything. Mix personal and business spending and you'll never know what's left. Separate the accounts and the bookkeeping first.

• Business and personal spending are separated
• I can see the numbers by arrival route
• This quarter's fixes are down to two or three
• Fixed the process, not the person

The hardest step. Most people can't close, and because they can't close they never get to try the next thing.

The one thing money can't buy is time. A product ticking along doesn't cost you a loss. It costs you every hour you'd have spent on the next attempt.

It half works, so you can't close it; you can't close it, so there's no money or time to experiment; so it keeps half working. Holding on because of what you've already put in is the biggest loss of all. Money already spent doesn't belong in the decision.

Growing has a rule too. The second one comes after the first runs without you. If one person leaving collapses it, it isn't a system yet.

• Judged whether this is eating the time for new attempts
• Removed sunk cost from the decision
• If closing: listed the shutdown steps (billing, servers, data, notice)
• If growing: confirmed the first one ran a month without me

Outside the sequence, all the time

The crisis arrives when things are going well

Collapse follows the record month. When it's working you start trusting yourself, your people and your customers, so you let small cracks pass as "close enough", and those add up into an incident. The warning sign is your ego growing. By the time you've stopped hearing customers it's already late. Which is why this page is one to re-read when things are good.

Benchmarking means taking the journey apart

Go somewhere that works and ask one question: why does this work? The moment you say "there's nothing to it", there's nothing left to learn. If there really is nothing to it and it still works, the answer is in how they tell people, so learn that.

The method is splitting the journey: discovery → first screen → first use → first value → the ask → leaving. Write down whether each point has a moment worth remarking on. The empty cells are your opening. And don't hunt for a grand secret. Steal something small and apply it the next day.

Decide what you won't do

• No ad spend before the product stands up
• No costs while assets are still zero
• No imitating a format that isn't yours
• No mistaking a trend you keep seeing for a trend you spotted
• No starting the next thing before this one runs without you

The whole thing, in one line

What to sell → where to fight → what to twist → how they arrive → what to call it → what to charge. That's all before you tell anyone. Then: product → assets → content → ship → measure → close or grow.

Keeping the order alone removes most of the waste. We're working through it ourselves, and we'll revise this page as we learn. If something here is wrong, tell us and we'll fix it.