Thoughts
The same shapes turn up again and again.
Twenty years in, you stop seeing problems and start seeing patterns. A working library of the ones that repeat across high-growth consumer brands, on eCommerce, on AI, and on leadership.
Every brand believes its situation is unique.
Most of the time it’s a familiar shape wearing a new logo. The replatform that solves the wrong layer. The board that mistakes activity for progress. The agency relationship nobody can honestly grade. Naming the pattern is the first half of fixing it.
Why your £50k a month agency retainer feels like it’s doing nothing
Most agencies aren’t dishonest. But they are human. How to audit the relationship and know whether to stay, renegotiate or leave.
Why replatforming projects drift, and how governance fixes it
Replatforming projects don’t fail because of bad technology. They fail because nobody says no to scope creep.
Why wholesale brands struggle going direct
Same product, same presentation, new Shopify store. Same ingredients, different dish.
Why discovery phases fail scale-ups
Six figures for a discovery phase that tells you what you already know. This is the Learning Tax.
Why your portfolio brand’s digital operation is stuck
Growth stalled post-acquisition? It’s usually infrastructure, not strategy. The three hidden risks in PE portfolio brands.
Your agency doesn’t work for you. They work for your brief.
The quality of agency output is capped by the quality of your brief, and most brands have nobody whose job it is to question it.
Your AI pilot didn’t fail. Your brief did.
Most AI pilots don’t produce bad output. They produce generic output. That is not a technology problem, and buying better technology won’t fix it.
When an agent does the shopping, what makes it choose you?
Agentic commerce is arriving faster than most boards think. The brands that get chosen won’t be the loudest. They’ll be the clearest.
The board wants an AI update. That’s the wrong request.
“What are we doing about AI?” produces a list of activity. The better question produces a decision.
Nobody knows what to spend on eCommerce. That’s the problem.
Ask five advisers what an eCommerce operation should cost and you’ll get five confident answers to five different questions.
The D2C playbook worked. Copying it doesn’t.
Direct-to-consumer looked like a business model. For a while it was mostly a set of conditions, and the conditions have gone.
What happens when everyone’s tech stack looks the same?
Shopify, Klaviyo, Meta, the same dozen apps. When the stack stops differentiating, the difference has to come from somewhere else.
The expensive silence between strategy and execution
The plan is agreed. The work is underway. The costly part is the conversation that never happens between the two.
The phone number that wasn’t
Customers kept asking for one thing. The answer was something else. What feature requests actually tell you.
Turn it off for a fortnight
Half of an eCommerce roadmap is optimising things nobody has ever tested for necessity. Subtraction is the cheapest experiment there is.
Nothing in the library matches that directly. It might still be a question worth asking.
The librarian
A diagnosis you’ve seen before is a diagnosis you can make faster. The library is the method getting sharper.
Where it’s heading
Every pattern here is a page of the book.
This library is Look Again being written in public, one shape at a time. Follow along here, or join the launch list and get the finished method, first chapter ahead of everyone.