How to Run Creative & Landing Page Tests That Actually Improve Profit
Share
Your new landing page converts at 2.4% instead of 2.0%. Twenty percent better. Ship it?
Not yet. If the new page also nudges people toward a cheaper product, you might be selling more orders and making less money. Most testing goes wrong at this point: the test is run properly, but it is judged on the wrong number.
Here is a simple way to run creative and landing page tests that improve profit, not just clicks.
1. Decide the winning metric before you start
Pick one number that decides the test, and write it down before you launch. Changing the metric after you see the results is how losing tests become "wins".
| What you test | Judge it on | Not on |
|---|---|---|
| Ad creative (hook, format, angle) | Cost per purchase, ideally cost per new customer | CTR, likes, CPM alone |
| Landing page or product page | Revenue per visitor | Conversion rate alone |
| Offer (bundle, gift, discount) | Profit per visitor, after the cost of the offer | Orders |
Click-through rate and conversion rate are useful clues. They are not the result.
2. Revenue per visitor: the landing page number
Revenue per visitor = conversion rate × average order value.
Back to the opening example. Each page got 10,000 visitors.
| Page A | Page B | |
|---|---|---|
| Conversion rate | 2.0% | 2.4% |
| Orders | 200 | 240 |
| Average order value | $60 | $48 |
| Revenue | $12,000 | $11,520 |
| Revenue per visitor | $1.20 | $1.15 |
Page B wins on conversion rate and loses on revenue. It also means 40 more orders to pick, pack and ship, each with its own shipping cost and payment fee. Page A is the better page.
If the two pages sell different products or offers, go one step further and compare profit per visitor: revenue per visitor minus product cost, shipping and fees per visitor.
3. Test one thing at a time
If you change the headline, the hero image and the offer at once, you'll know which page won but not why, and you can't reuse the lesson.
Good single-variable tests:
- Creative: the hook (first 3 seconds), the format (UGC vs studio vs static), the angle (problem, result, social proof).
- Landing page: headline, first image, where reviews sit, the offer above the fold.
- Offer: 15% off vs free gift vs bundle. (For the margin math behind offers, see the Pricing & Promo Impact Simulator.)
4. Wait for enough data
Small samples lie. With 30 orders per variant, a few lucky purchases can flip the result.
Rules of thumb that work for most small and mid-size stores:
- Run at least a full week, so weekdays and weekends are both in the test.
- Aim for roughly 100 purchases per variant before calling a winner on conversion or revenue. If that would take months, test bigger changes, because small tweaks need large samples to show up.
- Keep budget and audience the same on both sides of a creative test.
- If the difference is under about 10% and the sample is small, call it a tie and keep the cheaper option.
These are working rules, not statistics. If your platform has a built-in significance check, use it too.
5. Keep a test log
A one-line record per test is enough. Over time it becomes the most valuable marketing document you own.
| Date | What changed | Metric | Result | Decision |
|---|---|---|---|---|
| Oct 7–21 | Landing page: reviews moved above the fold | Revenue per visitor | $1.20 → $1.31 (+9%) | Keep |
| Oct 14–28 | Creative: UGC unboxing vs studio video | Cost per purchase | $41 vs $38 | Tie, keep studio |
Review the log once a month. Patterns appear fast: which hooks keep working, which page elements matter, which offers bring buyers rather than bargain hunters.
6. Scale winners, then check the whole store
A creative that wins on cost per purchase gets more budget. Then look one level up: as you scale, does your store-wide MER (total revenue ÷ total ad spend) hold? Winners often look less impressive at three times the spend. The MER-Based Ad Spend Planner helps you set how far to scale before efficiency drops below your target, and the paid ads tracking guide explains the weekly check.
The takeaway
Decide the metric before you test, judge pages on revenue per visitor and creatives on cost per purchase, wait for enough data, and write every result down. A test that raises conversion but lowers revenue is a loss, however good the chart looks.
👉 Track what your tests do to the bottom line with the free Profit Tracker.