Top 5 tips to make testing more profitable

When does testing go wrong? The answer is simple – when it’s not planned properly. If you try to test too much, too quickly or without focus, it’s doomed to fail.

So if you want to make the most of testing and maximise your ROI, make sure you’ve got the basics right. Your testing should be well thought-out and integrated into all campaign plans. That way, you’ll gain the knowledge you need to improve your marketing and increase profits.

So here they are – our top 5 testing tips…

1. Planning is key

A good test starts with smart thinking – build it into the campaign strategy and allow the appropriate time to keep it on track. Set very clear objectives, understand why you’re testing and what you want to learn. Work out what questions do you want answered and what will give you the greatest return.

2. Test one thing at a time

It could be anything from the format to the creative, the timing to the offer – but limit yourself to one. Testing more than one element at a time will make it impossible to identify what’s caused the difference in response.

3. Choose the right sample

Sounds obvious, but ensure your sample size is statistically valid and that you have enough volume to give reliable results. Fortunately there is a handy formula for working this out:

The Institute of Direct Marketing recommends the following approach:

The criteria for determining your minimum sample size are:

• The confidence level required (A) (usually 95%)
• The % variance above or below your observed test result which you will tolerate – your acceptable margin of error (B)
• The approximate expected response rate (C)

Because you want to be A% certain that C% of your universe, plus or minus B%, will respond in the same way. So, where:

S = sample size per cell
C = expected response rate
B = acceptable margin of error

The 3.8416 figure relates to the confidence level you require (in this case 95%) and changes depending on your chosen confidence level.

Worked examples:

1. Expected response rate of 0.1 and acceptable margin of error of 0.03.

2. Confidence level of 95%, expected response rate of 0.25 and acceptable margin of error of 0.03.

Once you’ve done your test you can check the ACTUAL margin of error that you ended up with – to see how this compares with your required margin of error.

4. Analyse the results

Take the time to study your results properly and document them in writing. Remember, averages can hide a lot of interesting facts – so ensure the numbers are thoroughly interrogated. Always aim to identify the ‘why’ – knowing what works is good but knowing why it has worked is much more powerful.

5. A test done twice is a test done right

You should run every test at least twice, especially if you discover a significant change in results. After all, it could just be a fluke.

 

« Back to the news page

• ‹ Previous article
• Next article ›