The Cost of Switching Between Two Brands in One Week

Essay header graphic: The Cost of Switching Between Two Brands in One Week

Written by

in

I write in two completely different voices before lunch most weeks, and for a long time I thought that was a scheduling problem.


One of my two learning projects is a systems-and-frameworks brand, faceless, editorial, patient. Every piece opens by answering the reader’s question directly and only then explains why. The other is a product-verdict brand, confident, plain-spoken, built to hand someone a decision in the first two lines. Same person, same week, opposite instincts about what a reader wants first.

The naive fix is discipline: switch hats, remember the rules, do not let one voice bleed into the other. That works for about a week. Then a framework post starts opening with a verdict, or a review starts opening with a system, and the drift is not a discipline failure. It is what happens when two things that need to stay distinct only exist in someone’s head.

Inside this piece: a summary of what it covers

Rules you remember are rules you eventually break

This is a smaller version of a mistake I have watched whole marketing teams make. A brand voice guide gets written, everyone nods, and six months later every piece of content sounds slightly like whoever wrote it that day, because the guide lives in a document nobody re-opens and not in anything that actually checks the work before it ships.

What actually holds a voice in place is not the rule, it is a structure that makes the wrong move harder than the right one. A template that only has room for a verdict in the first line. A checklist that gets read before anything ships, not trusted to memory. The rule was never the fix. The rule was a description of what the structure should do, and the structure is the part that has to exist for the rule to survive contact with a Tuesday when you are tired and switching projects for the third time that day.

What actually keeps the two voices apart

Three things, in order of how much I underestimated each one before I needed it.

A locked template, not a remembered style. The systems brand’s template makes it structurally awkward to open with a verdict, because the first section is built to hold a direct answer to a question, not a scored recommendation. The verdict brand’s template does the opposite. Neither voice has to be remembered, because the shape of the page is already arguing for it before a word gets written.

A gate that runs before publishing, not a guideline that runs in your head. Every piece for either project gets checked against a written list before it goes live: right tone markers present, wrong ones absent, sourced where it needs to be sourced. The check does not care how confident I felt writing it. It only cares whether the finished piece actually matches the list, which is a different and much more reliable question.

Physically separate visual language. Different colours, different type, different rhythm on the page. This sounds cosmetic and is not. Switching into a different visual frame resets something in how a piece gets written that switching projects in the same visual frame never quite does. I did not expect the biggest discipline aid to be a font, but it is.

The part that generalises past writing

None of this is really about voice. It is about the difference between a rule and a system, and I keep relearning it in smaller and smaller rooms. A rule is a sentence someone has to remember and apply correctly every time, under whatever state they happen to be in that day. A system is an arrangement of the environment so that the correct behaviour is the easy one and the wrong behaviour is the awkward one. Rules degrade under fatigue. Systems mostly do not, because they do not depend on anyone’s state.

Running two projects with opposite voices in the same week turned out to be a useful, low-stakes place to notice this, because the failure mode is fast and visible: a wrong-voice sentence reads wrong within a paragraph, not within a quarter. Most systems problems take longer to show their cracks. This one shows them by lunchtime, which has made it a genuinely good teacher.


If you’re building something where consistency keeps slipping and you suspect it is a systems problem rather than a discipline problem, I would be glad to compare notes. Get in touch, or see the Work page for more on both projects.