Spoke
Comparing channel economics
A switch, a building, and something in between. The comparison that matters isn't cost per lead today. It's what each one is worth after you stop paying for it.
A way of thinking about it. Deliberately contains no benchmark figures, because no cost-per-lead benchmark with a traceable source and a stated method exists.
The usual comparison puts cost per lead side by side and picks the lowest number.
That's wrong twice over. It compares different units, and it ignores what each channel leaves behind when you stop.
Which is better value, paid search or SEO?
They're different instruments, and the honest answer depends on your time horizon. Paid delivers now, you control it, and it produces nothing the moment you stop. Organic takes quarters to build, can't be switched on, and keeps working — fading slowly rather than stopping dead.
A firm that needs cases this month buys paid. A firm building an asset invests in organic. Most need both, which makes the real question the ratio, not the choice.
Three different shapes
| Paid | Organic / AI visibility | Pay-per-lead | |
|---|---|---|---|
| First result | Days | Quarters | Weeks, after screening |
| You're billed for | Clicks | Nothing per unit | Leads |
| When you stop | Stops that day | Fades over quarters | Stops that day |
| Main risk | Paying for clicks that bounce | Building something nobody finds | Paying for junk leads |
| Leaves behind | Nothing | An asset | Reviews and reputation |
That last row is the one missing from most comparisons, and it's where the actual difference lives.
Build it from your own numbers
Four inputs, from your records rather than from an article:
- Your conversion rates by channel, from contact to qualified to signed.
- What a signed matter is worth, and how long until you see the money.
- Your real channel costs, including the internal time nobody counts.
- Your capacity. A lead you can't service isn't revenue.
With those, you can compute cost per qualified outcome per channel. That's the only figure that compares across different billing models.
Why there are no benchmark numbers on this page
Because none with a traceable source exists.
The cost-per-lead tables going around have no stated method and vary enormously by category, geography, season and competition. A number with no method behind it is worse than no number, because it anchors your decision while looking like evidence.
If a vendor quotes one, ask where it came from and watch what happens next.
The part everyone leaves out
Spend on paid for a year and you have the leads it produced. Spend the same on content and entity work and you have the leads and a library that keeps producing. That's a balance-sheet item, not a marketing expense.
The honest counterweight: that asset can be destroyed. A penalty removes it, which is why black-hat risk prices banned tactics against everything you've built rather than against the monthly saving.
An asset built on bought links is an asset with a liability stapled to it.
The ratio that usually works
Enough paid to fill your current capacity, so the business isn't waiting on something that compounds slowly. Steady investment in the compounding one, so the paid budget isn't load-bearing forever.
The pattern to avoid: running the switch alone and never building anything. It's comfortable and it measures well, and it leaves you renting your own demand indefinitely.
More in Buying services
Red flags
A short list of things that end the conversation, and why. Most of them aren't dishonesty. They're somebody describing a product that can't exist.
Black-hat risk
Cheap links and generated pages aren't a cost saving. They're a bet against everything your domain has earned, and you should price them that way before you buy them.
Contract terms
Everybody's pitch sounds the same. The contract is where firms differ, and these are the clauses that decide what happens when results are slow — which is the situation you'll actually be in.
Recovery playbook
What to do when visibility collapses. The first job is working out which of two very different things happened, because the response to one is a process and the response to the other is a diagnosis.