Practitioner · Lesson 5 of 5
Landing pages that convert paid traffic
About 11 minutes · Yogender Kumar, Managing Partner
By the end of this lesson you'll be able to
- Check a landing page against what a paid visitor needs
- Calculate how a conversion rate change moves cost per order
- Plan an A/B test that gives a trustworthy answer
Why this matters
You can pay for the right visitor and still lose them on the page. A landing page that converts a little better makes every pound of media work harder.
The core idea
A paid visitor arrives with one question: is this what the ad promised? A landing page that answers it quickly does five things:
- Matches the ad. The same offer, product and words as the ad they clicked.
- Loads fast on a phone. Much paid social traffic is on phones, and slow pages lose people before they see anything.
- Asks for one action. One main button, not a menu of choices.
- Keeps forms short. Every extra field is a reason to stop.
- Shows proof. Reviews, delivery and returns terms, and who you are.
The money side is simple. If the cost per click stays the same, cost per order falls as conversion rate rises.
Worked example
Ashby Outdoor is the imaginary shop from Lesson 3. The figures are illustrative. It sends paid social traffic to its jacket range. It pays £1 a click, and 2% of visitors buy, so each order costs £50.
Its conversion analyst walks the journey on a phone:
- The ad shows one waterproof jacket. The page shows 40 jackets.
- The page takes six seconds to load on a mobile connection.
- Delivery costs only appear at checkout.
Ashby builds a page for that one jacket, with its reviews and delivery costs near the top, and makes it load faster.
It runs an A/B test. Visitors are split at random, so half see the old page and half the new one, at the same time. After four weeks:
- Old page: 6,000 visitors, 120 orders: 2%.
- New page: 6,000 visitors, 180 orders: 3%.
At £1 a click, cost per order falls from £50 to £1 ÷ 0.03 = £33.33.
This tests the whole new page against the old one, so it can't say which change did the work. Later tests can change one thing at a time.
Ashby decided the stopping point before it started, and waited for at least 100 orders on each page. My rule: never call a test on a few dozen orders, because chance alone can swing them.
Cost per order on each landing page, at £1 a click
In practice
- Open every landing page on a phone, on a mobile connection, before ads go live.
- Match the headline and image to the ad that sends the traffic.
- Show delivery costs, returns terms and reviews near the top.
- Cut the form to the fields you really need.
- Test one change at a time, both versions running at once.
- Wait for enough orders in each version before deciding.
Common mistakes
- Sending paid traffic to the homepage. It makes visitors search for what the ad promised. Send them to the matching page.
- Calling a test too early. A few days and a few dozen orders aren't enough. Set the stopping point first.
- Testing before and after. Comparing this month with last month mixes in the season and other changes. Run both versions at once.
Try it yourself
You pay 80p a click. Your landing page converts 2.5% of visitors. A new page converts 3.2%. What's the cost per order before and after, and how much would you save on 400 orders a month?
Show answer
Before: £0.80 ÷ 0.025 = £32 an order. After: £0.80 ÷ 0.032 = £25 an order. On 400 orders the saving is 400 × £7 = £2,800 a month.
Quick check
Not scored. Pick an answer to see the explanation.
What's the most reliable way to test a new landing page?
You pay £1.50 a click. Conversion rises from 3% to 5%. What's the new cost per order?
Key takeaways
- A paid visitor wants to see what the ad promised, straight away.
- Match the ad, load fast, ask for one action, keep forms short and show proof.
- Cost per order is cost per click divided by conversion rate.
- Test both versions at once, and wait for enough orders to decide.