When to optimize your campaign — The 20% rule

I'm gonna teach you a golden rule to follow when it comes to knowing when to optimize you campaigns. I use it all the time and it's a great rule for both advanced data analysts and newcomers alike.

Previously I've written about historical benchmarking as well as competitive benchmarking. One of the most asked question I get is when should I take action if I see my campaign is not going well?

The answer to that is the 20% rule.

Learning to use benchmarks

Whether you use your own benchmarks or a competitive benchmarks, you should always have a benchmark that you can compare yourself too. That's the basis of this 20% rule.

You need something to compare your previous campaign too, or in some cases your competitors benchmarks.

But I've only ran a couple campaigns before...

Don't worry if you haven't had a lot of campaigns running to gather a solid enough data set. Even one campaign can be a benchmark to use this rule to know when to take action on your campaign.

What exactly is the 20% rule?

The 20% rule on when to optimize your campaign is what I call the golden rule of campaign optimization.

This rule gives you the answer on when you should take action.

Essentially, the rule says that if your marketing outcomes are above or below 20% your benchmark, you should take action and investigate further.

Here is an example to make it clearer:

Let's say you're running a shopping campaign and you know from previous campaigns with the same budget that you've made 1000 sales after 30 days on average.

This time however, you've only made 650 sales after 30 days. That falls below our 20% mark and it tells you that it's time to investigate and optimize

On the other end of the scale, if your campaign generate 1250 sales after 30 days, the campaign will have done over 20% better that usual. This is also a time to look further into what you did right.

Some people might just pat them self on the back after a very successful marketing campaign and do nothin more, but when it falls over 20% from your benchmark, you should definitely take a look and learn and try to figure out what you did correct.

In some cases, the reason it went below or higher than 20% might be out of our control.

Examples

Take a look at the table below. We have 4 different sources with benchmarks on each KPI we're tracking. I've done the calculations to show you what are the numbers that falls below or higher than the 20% for each benchmark we're tracking.

Source

Metric

Lower than 20%

Benchmark

Higher than 20%

Facebook post

Avg. likes

18 or lower

22

26 or higher

Facebook ad

Clicks

488 or lower

610

732 or higher

Website

Conversion rate

1.4%

1.8%

2.2%

Email

Open rate

28%

35%

42

How to calculate a 20% decrease of increase

Calculating a 20% decrease or increase on your benchmark is not that hard. You just have to split the calculations into two parts.

Find 20% decrease

a) Benchmark x 0.2
b) Benchmark - Result of a

For example I want to calculate a benchmark of 42:
42 x 0.2 = 8.4
42 - 8.4 = 33.6
Round it up and we get 34. That means 34 is 20% lower than 42.

Find 20% increase

a) Benchmark x 0.2
b) Benchmark + Result of a

For example we have a conversion rate of 1.3:
1.3 x 0.2 = 0.26
1.3 + 0.26 = 1.56
Round it up and we get 1.6 which means 1.6 is 20% higher than 1.3.

How to use this rule properly to optimize your marketing campaigns

Here are some things to keep in mind before applying this rule to optimize your marketing campaigns.

Calibrate based on channel

There might be times where you're working with channel or sources that has an extreme volatility. That means that you might hit both 20% over and under your benchmark quite often.

One example of that might be organic social media posts. Depending on the day you post, your post might get a large variety of the number of likes. You might have a benchmark average likes of 50, but one post might generate 89 and another one just 23 likes. In this example we've fallen way below and way higher than 20% of 50.

So it's also worth taking the volatility in the channel into consideration before  optimizing your marketing activities. 

You can have a 40% rule on more volatile channels while keeping the 20% rule on other marketing channels. Be aware not to set it too high, or else you might not need to take any action at all.

Same goes for setting the percentage too low. You will then need to take action too often and that just isn't right either.

Don't use data that is too old

Another thing to keep in mind is not using way too old data when doing historical benchmarks. Six months or a year is a good timeframe to aim at. Digital marketing changes fast so data that is more than 2 year old is probably past due date and is not reliable as benchmarking statistics.

One example is cost per click which almost always goes up year-on-year. Using CPC from 3 years ago will simple be a wrong benchmark in todays advertising world.

We think you will also like...