Refer your own products before you pay for strangers
The cheapest visitor you will ever get is the one already reading your other page. Wire your properties to hand each other traffic before you pay a platform to hand you strangers.
The most expensive visitor is the one you rent. You bid for their attention, pay whether or not they buy, and lose them the moment the campaign stops. The cheapest visitor is the one already reading another page you own — and most small catalogs never introduce the two.
If you sell more than one thing, you have an estate, not a list of unrelated storefronts. Someone reading your essay on shipping reliable software is a plausible buyer of your prompt library. Someone who just downloaded your free starter pack is a plausible reader of your newsletter. Those adjacencies are traffic you already paid for once. Handing that visitor to the next relevant thing you make costs nothing and compounds every time you add a property.
The mechanics are unglamorous and that is the point. A footer that links your related products on every page. A "you might also like" block that points to a genuine sibling, not a random SKU. One newsletter that captures email across the whole estate instead of five orphaned signup boxes. A checkout confirmation — the highest-attention screen you own, because the buyer just trusted you with money — that recommends the one product a happy customer would actually want next. None of this is a growth hack. It is plumbing, and plumbing is what carries the water for free.
Here is the honest limit, because the flywheel breaks the moment it becomes spam. It only works when the products are truly adjacent and the recommendation is useful to the reader, not just convenient to you. Cross-linking unrelated things to pad page views trains people to ignore your links, and an ignored link is worse than no link — it spends the one thing the estate runs on, which is the reader's belief that when you point somewhere, it is worth going. Point them somewhere good or do not point at all.
This is also why the estate is a moat a bigger ad budget cannot buy. A competitor can outspend you on cold traffic tomorrow. They cannot cheaply reproduce a set of related products that each send warm, already-interested readers to the next — that only comes from having built the things and connected them with care. Paid traffic rents you a stranger for the length of a click. A well-wired estate lends you the same person, warmer, again and again, for as long as they find each stop worth their time.
So before you open an ad account, walk your own properties as a visitor. Land on each one and ask: if I liked this, where would I go next — and is that path actually here? Every missing link is a paid customer you already have, quietly leaving through a door you forgot to build.
Ready to try PromptLibrary?
Curated, battle-tested prompts you can actually buy
Browse the Prompt Library →Keep reading
A heartbeat is not the same as a pulse
A reachable host proves the lights are on. It says nothing about whether the one thing you depend on is actually working.
Spreading a small bet thin doesn't make it safer
Splitting a limited budget across several bets feels responsible. It isn't -- it's the same risk wearing a diversification costume, and a bet too small to move any single outcome can't teach you anything about that outcome either.
A system mid-repair looks exactly like a system that's broken
A monitor can't tell a regression from a repair in progress -- the reading looks the same either way. The fix is checking who else was in the room before you escalate.
The Loop