A Scientific Method Beats a Best-Practices List, Every Time
Every industry has its folklore — a list of best practices passed hand to hand like a family recipe. Post at this time of day. Use this subject line. Put the button here, make it this colour, price it just below the round number. Some of it works. Most of it worked for a business that isn’t yours, in a moment that has already passed. And all of it is comfortable, because a borrowed answer requires no thinking. That comfort is exactly its weakness.
Why the list feels safe and usually isn’t
A best-practices list is attractive for an honest reason: it removes the burden of deciding. Someone smarter, or at least louder, has already done the work, and you get to skip to the answer. No hypothesis, no risk, no exposure. Just do the thing everyone agrees on.
But look closely at where those practices came from. Almost all of them are a single business’s result, generalised into a law. “Emails sent Tuesday at 10am get opened more” is a real finding — for the company that measured it, with their list, their audience, their industry, in the year they measured it. Somewhere along the way the caveats fell off and it became a rule for everyone. By the time it reaches you, it is a photograph of someone else’s Tuesday, presented as a map of yours.
The deeper problem is that best practices converge. If everyone follows the same list, the list stops being an edge and becomes the baseline — the thing you have to do just to look normal, not the thing that makes you win. Advantage never lives in what everyone already knows.
The method that actually travels
There is an alternative, and it is not a better list. It is a way of working, borrowed from science and applied to a business you actually run: see a real signal in your own data, form a hypothesis grounded in that data, test the smallest version of the change, keep what wins, and discard what doesn’t. Then do it again.
Run that loop long enough and you learn a lesson the list can never teach you. The method is portable. The answer never is. Ten years of running an honest scientific method on real marketing — thousands of small tests, most of them wrong — does not leave you with a giant book of correct answers. It leaves you with something far more valuable: the reflex to distrust every borrowed answer until your own data has voted on it.
What worked for one business’s homepage headline tells you nothing reliable about yours. What your own data says, tested on your own site, tells you everything. A best-practices list is a shortcut around thinking. A method is a way of thinking that happens to produce an answer — your answer, to your question, about your business.
What it means for a business owner
This is not an argument for ignoring what others have learned. Best practices make a fine source of hypotheses — a list of guesses worth testing. It is when they get treated as conclusions that they cost you. “Move the button” is a great thing to try. “Move the button because a blog said to” is a great way to change your site with no way of knowing whether it helped.
The shift is small and total at once. Instead of asking “what’s the best practice here?”, you start asking “what does my own data suggest, and what is the smallest test that would tell me if I’m right?” The first question outsources your judgment. The second builds it. Over time, the business that tests owns knowledge nobody can copy, because it was never published anywhere — it was earned, on their own site, one small experiment at a time.
What to do about it
Treat every “best practice” as a hypothesis, never a rule. When you read that something works, do not implement it. Write it down as a guess and put it in line to be tested against your own audience. If it survives, keep it and now you know. If it doesn’t, you have just avoided a change that would have quietly cost you.
Test the smallest version first. You do not need statistical machinery or enterprise traffic to learn something. Change one thing, watch whether the number responds the way your guess predicted, and let the result — not the reputation of the source — decide.
Write down what would prove you wrong. A hypothesis you cannot disprove is folklore in disguise. Before you run a test, name the outcome that would make you abandon the idea. If you cannot name one, you were never testing. You were decorating.
Keep a record of what your own data has already settled. This is the quiet compounding asset. Every test that survives becomes a fact about your business that no competitor’s best-practices list contains. That growing pile of earned answers is the real moat.
The list runs out. The method doesn’t.
Borrowed answers are a loan you keep paying interest on — they were never yours, they go stale, and everyone else has the same ones. A method is something you own outright. It works on a headline today and a pricing page next month and a channel you haven’t even adopted yet, because it was never about the answer. It was about how to find one.
The next time someone hands you a best practice, take it gladly — as a question to test, not a verdict to obey.
That loop — signal, hypothesis, test, keep what wins — is the engine behind conversion intelligence, where it gets pointed at the specific pages and funnels that decide whether a visitor becomes a customer. The guide there shows how to run it without enterprise-scale traffic. And when you want the signals surfaced for you so you can spend your energy on the tests, Auditry reads your own data and shows you where the real questions are.