Most marketing systems produce. Few compound.

A producing system makes more output every month and knows nothing new. A compounding system carries what worked into what it does next. Both look identical in a demo; only one is an asset a year later. Growth Engineering is the practice of building the second kind, across seven disciplines: Growth (own a number; without it, agents run and nothing is accountable to anything), Direction (command the machine without approving every step; without it, surrender or a second job), Distinctiveness (stay tell-apart-able from the machine mean and the category mean; without it, correct work nobody can tell from anyone else's), Memory (keep what worked in a form the next run can read; without it, month four repeats month one), Experiments (hold beliefs that can lose; without it, nothing is ever wrong so nothing is ever learned), Production (make the asset end to end at channel standard), and Eval (test yourself before you change yourself; it runs through the other six). Distinctiveness Engineering and Direction are two of its disciplines, each with its own home.

Every marketer is now a growth engineer. AI made the production side of marketing cheap; what is scarce is the person who can direct the machine, keep the work tell-apart-able, and prove it worked. That is not a new job title. It is the job. The seven disciplines are learnable: Stack Score grades your own stack at its weakest discipline, and The Marketing Machine is a six-week programme where marketers build all seven on their own company. Everything written on the practice is collected in the writing archive.

The research behind the practice: Ivanooo runs a standalone research program on how professionals lose command of AI without noticing. The chain it studies ends at the brand: people hand their thinking to the machine, work produced that way converges toward the machine's average, and a converged brand becomes one AI cannot tell apart — a brand AI never recommends. The middle links are proven science; the full chain is what the program studies, and it is why the practice begins with a human directing a working machine, measured. The research publishes at research grade or not at all; marketing is where it is proven commercially, and that order never reverses.