More traffic or better conversion — which is worth more?
A conversion fix is bought once and works forever. Traffic has to be bought again every month. So if any step in your funnel is below benchmark, fix that first — and only buy traffic once there's nothing left to fix.
A conversion improvement is bought once and works on every visit afterwards — including the traffic you buy later. It costs nothing extra per visitor.
A step already at benchmark cannot be improved by copy. More visitors is the only lever left, so that is where the budget goes.
That is the whole asymmetry. Traffic is consumed on arrival; a better conversion rate keeps working. Both can be priced against contribution per order — which makes this arithmetic rather than a matter of preference.
The comparison, stated properly
Most answers to this are opinion. This one doesn't have to be — both options end in the same number, so both can be priced and put side by side.
Price both against contribution per order. Traffic is worth extra sessions × your current conversion rate × contribution per order, less what the sessions cost. Conversion work is worth the percentage-point gap × the volume at that step × the downstream step rates × contribution per order, less the cost of the work. Compare the two:
One full year, from the P&L tracker. Each slice is a real cost line, not a category — and all fourteen add up to exactly 100% of gross revenue.
Gross revenue
€4.620.040
Before the product is even made
Making it
Getting it to the customer
Running the business
Basis. Four quarters, €4.620.040 gross revenue, 15.200 orders. Referral fee at 15% of gross and payment fees at 1,9% plus €0,25 an order are the two largest fee lines and both scale with revenue, not with volume. Slices sum to 100,00% with no residual. Two figures in the source are unresolved: units sold is lower than order count in every quarter, and the product-level revenue block does not reconcile — neither affects the slices above, but both affect any per-order figure derived from them.
The order matters more than the choice
There is a sequencing point here that is worth more than the comparison itself.
Buying traffic into a funnel with a known leak spends the acquisition budget at the current conversion rate rather than the achievable one. Every visitor bought before the leak is closed converts at the worse rate, and that difference is unrecoverable. Close the priced gap first, then buy the traffic that will convert at the better rate.
This is the practical reason the answer is usually "conversion first" even when the two figures look comparable: doing conversion work first makes the traffic more valuable, while doing traffic first does nothing for the conversion work.
Six months, the same budget, two ways to spend it. The difference is not which produces more orders this month — it is what you own in month six.
Six payments, converting at the rate you already had. Month seven starts at zero.
One payment, then every visit converts at the better rate — including the traffic you buy afterwards.
Buying traffic into a funnel with a known leak spends the budget at the current rate rather than the achievable one, and that difference is unrecoverable. Close the gap first, then buy the traffic that will convert at the better rate. Once every step sits at benchmark, traffic is the only lever left — and then it is the right one.
When traffic is the right answer
When should you buy paid traffic? Only in three scenarios:
Your funnel already works. If every step meets or exceeds category benchmarks, copy and design tweaks won't move the needle anymore. More visitors is your only growth lever.
You have a capacity constraint, not a conversion problem. You have the conversion rate down—you just need the scale to fill production or operations.
You need data for a new product or market. You can't optimize a funnel with 40 visitors a month—that’s just statistical noise. You need baseline traffic before a conversion rate even means anything.
Does conversion work run out?
Yes. Conversion work closes the distance between a step's current rate and its benchmark; it does not push past the benchmark indefinitely. Once a step reaches its category benchmark the remaining headroom is small and expensive, and further effort produces less than the same budget spent on traffic. The ceiling is the benchmark, not perfection.
What a shop with no data should do
Measure before choosing. A shop that has never benchmarked its funnel per step cannot price either option, and the choice becomes a guess. Recording each step's current rate and volume against a category benchmark takes roughly an afternoon and is the only thing that makes the comparison answerable rather than rhetorical.
What this does not tell you
The comparison assumes the gap is closable by content. If a product page leaks because the delivery promise is uncompetitive or the price is wrong for the market, the gap is real and the money is real, but copy will not close it. Check which kind of gap it is before committing the budget.
It also assumes traffic can be bought at a stable cost. In practice cost per session rises as volume scales, so the traffic side of the comparison degrades at exactly the moment it is being scaled — which pushes the answer further toward conversion work than a single-point calculation suggests.
