What's the minimum budget for testing one funnel?

There's no fixed number — it depends on the geo and the cost of the target action — but the rule holds: budget should cover enough volume to get a real read on at least one audience, plus a reserve for swapping a creative or domain.

Should I test several creatives at once?

Yes — 2-3 creative variants at launch is standard practice, otherwise you can't tell whether the funnel worked or just one image did. Use a split that locks a variant to each visitor so the comparison stays fair.

Budget's gone and there are still no conversions — what's wrong?

Check the technical side first: did a reviewer or bot traffic eat your spend because cloaking wasn't set up, was the domain blocked, is traffic even reaching the PWA. Only then should you judge the funnel itself.

Can I test on a small budget if I'm targeting Tier-1?

You can, but the test will be shorter and less conclusive because of the higher cost per click. It's smarter to prove out the funnel mechanics on a cheaper tier first, then point Tier-1 budget at a validated mechanic.

How Much Budget You Need to Test a Funnel

A funnel test budget isn't one number — it's the sum of several line items: creatives, clicks to gather signal, infrastructure, and a buffer for anything that breaks mid-test. Here's how to size a PWA funnel test budget so you get real statistics in a day or two instead of just burning your deposit.

What a test budget is made of

A test budget is more than your bid per click. It covers producing 2-3 creative variants, the learning spend on a fresh ad account, the cost of a domain and PWA wrapper, and a reserve for swapping a domain or creative if the platform flags them. Skip the reserve and your test stalls on a technicality instead of ending with real numbers.

How much traffic you actually need

Judging a funnel off 5-10 clicks is guesswork — you need enough volume to see one full cohort move through click, install, and target action. Watch for the point where the funnel stops «jumping around»: conversion at each step settles into a range instead of swinging from zero to a spike. Scaling spend before that point is premature.

Splitting budget across geos and tiers

CPI and cost per click in Tier-1 usually run noticeably higher than Tier-3, and the competition for attention is fiercer. If budget is tight, it makes more sense to prove out the funnel mechanics on a cheaper tier or a low-competition geo first, then bring a proven mechanic to Tier-1. APEX's sticky split-tests lock one prelander variant to each visitor, so even a modest budget splits into comparable groups instead of a random mix.

Where test budgets usually leak

Most of the budget lost on a test doesn't go to «a bad offer» — it goes to technical holes: bot and reviewer clicks with no cloaking, a domain that gets blocked mid-test, and no end-to-end tracking, so you can't tell which creative actually worked. APEX's cloaking and antibot filters close the first hole; its own domains with one-click rotation close the second.

When to stop-loss and what to do with the result

Decide your stop-loss threshold before you start — say, spend equal to several times the expected cost per target action with zero conversions — and stick to it even when you want to push further. If the funnel turns a profit, log which creative, prelander, and geo drove it, then scale spend in steps rather than 5x in a single day.

FAQ

What's the minimum budget for testing one funnel?: There's no fixed number — it depends on the geo and the cost of the target action — but the rule holds: budget should cover enough volume to get a real read on at least one audience, plus a reserve for swapping a creative or domain.
Should I test several creatives at once?: Yes — 2-3 creative variants at launch is standard practice, otherwise you can't tell whether the funnel worked or just one image did. Use a split that locks a variant to each visitor so the comparison stays fair.
Budget's gone and there are still no conversions — what's wrong?: Check the technical side first: did a reviewer or bot traffic eat your spend because cloaking wasn't set up, was the domain blocked, is traffic even reaching the PWA. Only then should you judge the funnel itself.
Can I test on a small budget if I'm targeting Tier-1?: You can, but the test will be shorter and less conclusive because of the higher cost per click. It's smarter to prove out the funnel mechanics on a cheaper tier first, then point Tier-1 budget at a validated mechanic.

Build your PWA in APEX in minutes

Cloaking, anti-bot, push, split tests and your own domains — in one service.

Get started free