What is the difference between a hobby and a business? Money. A business exists when something is exchanged for money, and validation is the habit of checking whether that exchange is really happening, before you build, while you spend and every time somebody shows you a number.
There is no opinion as important as whether or not someone will actually give you money in exchange for what you offer. Without validation, you're guessing, and you are probably wasting resources on activity that feels like work and moves nothing. Validation is critical thinking applied to marketing: first principles, scepticism about data, and tests cheap enough to run before the expensive decisions get made.
Start with the exchange
The test of the exchange is simple: did money change hands. Interest is a step toward it, and so is a booking, an enquiry or an order, but none of them is the exchange itself.
My first ever BookPrint customer paid something like a 20% deposit and never paid the rest. The order was real, and most of the money never arrived. An order is not the exchange. The money is.
Traffic + Offer = Results, and in the equation the result that counts is somebody paying, not a line on a dashboard. If the offer is solid and nobody is seeing it, you have a traffic problem; if people are seeing it and nobody is buying, you have an offer problem. The diagnostic sorts out which one you have, and validation is how you find out whether the fix worked.
Three checkpoints before you build
The site's one instruction is the cheapest version of all this: before you build anything, put a minimum of traffic in front of the offer and let real people tell you whether the business works. The 3-checkpoint validation plan sets it out step by step. Interest from strangers first, then a sale to a real person, then the maths on whether a customer is worth more than they cost to win.
The maths is the checkpoint I have the oldest paperwork for. In 2009, planning a house washing business called Ultimate Wash, I wrote down that 15 jobs a week at three bedroom single storey houses meant $3,750 turnover weekly, which was 780 jobs a year. The same plan priced the flyers at $1,300 for each drop of 7,000, three drops, and 500 fridge magnets at $265. A to-do list from the next year set a target of 100 customers from 25,000 flyers. None of that was proof of anything. It was the arithmetic that told me what the flyers had to do before any money went out the door, so that the results could be checked against something.
A test also has to run. Plenty of campaigns that never work were simply never started, because the owner got nervous about how much to spend or where, and a test that never runs validates nothing. A small amount of real traffic answers questions that another month of planning can't.
Every number, for the life of the business
Validation doesn't stop once the business is trading. The question to ask of every metric, for as long as the business runs, is whether it connects to actual business results.
If you can't draw a clear line from a metric to a business outcome, that metric is not evidence of anything. Page views, followers, impressions and engagement can all move without a single extra dollar arriving, and a report full of them can make a quiet month look busy. A content calendar that is full every week is not a strategy either, and neither is a follower count that grows while sales stay flat. The outcome at the end of the line is profit, not just revenue, because good marketing makes the business better at making a profit rather than scaling revenue for smaller and smaller gains from more and more spend.
Society follows fashion, and data is fashionable today. That doesn't make every number true. Creativity and data are both fashions, and each is wrong on its own, so an idea nobody tested and a number nobody questioned need the same treatment.
Platform numbers are the ones to question first
The numbers most worth checking are the ones a platform reports about its own performance, because the platform marking its own homework has an obvious interest in the mark.
A public case shows how far apart the reported number and the real one can be. In a 2014 working paper, Consumer Heterogeneity and Paid Search Effectiveness, economists Tom Blake, Chris Nosko and Steven Tadelis reported a series of large field experiments run at eBay to measure what paid search ads actually caused. Returns from paid search turned out to be a fraction of the usual non-experimental estimates, because the people who click on search ads are often people who were going to buy anyway. Ads on eBay's own brand name had no measurable short-term benefit at all. For other keywords, new and infrequent users were influenced by the ads, but frequent users who would have bought regardless accounted for most of the spend, and average returns came out negative. The paper was later published in Econometrica in 2015.
The method is the lesson. They didn't argue with the attribution report; they switched ads off for some people and on for others and measured the difference in sales. That is validation in its plainest form, and the same idea works at any size: change one thing, hold the rest, and watch the money.
Scepticism cuts both ways
The numbers can miss what works as well as flatter what doesn't. Plenty of what makes the phone ring never shows up in a report, and it doesn't need to, because the phone ringing is the evidence.
So the question is never whether the dashboard looks good or bad. It is whether money arrived, and what you can honestly say caused it.
Validation keeps going
A channel that worked last year may not work this year, and an advantage that came from a platform's cheap years can look like a strategy until the window closes. Every new channel, price or audience gets the same treatment as the first offer did: a small test, real money, and a line drawn from the number to the bank.
On 4 Quadrant AI, Human Validation is the quality gate in the operating framework, the same check applied to operations rather than marketing.
Once a test says the offer holds up, traffic and the offer go deeper on each lever, and persuasion covers how to describe the offer in a way that still works when the customer can see what you are doing. The traps covers what happens when businesses scale before any of this has been checked.
Put the offer in front of real people, count the money rather than the attention, and ask of every number you are shown whether it connects to money in the bank.
Marketing Curious: Working the Noise is the longer telling.