When your metrics start sinking you need to find the ‘leak’ and quickly before your campaign is scuppered.
It’s unlikely that everything is going wrong with your campaign at once (unless you’re having a really bad day). It’s probably one particular element that’s bringing your metrics down.
Perhaps your targeting is off. Or your creative content isn’t hitting quite the right notes. Maybe your CTAs aren’t enticing enough, or your subject lines too easy to ignore.
How can you sift through all of these factors and pinpoint which one, precisely, is causing the problem?
Well, there’s no definitive answer. But following this guide will help:
Work from the big picture inwards
Usually, pinpointing poorly performing elements is a case of narrowing down and narrowing down until you identify a small set of problem factors to test.
So, start with your overall strategy, and work your way down from there:
- Go back to your goals. Re-establishing what you want to achieve with your campaign can be a great help when working out where to start your investigation. For example, if your goal is to increase ROAS, focusing on the factors contributing to your CPC metric would be a good idea. If you’re aiming to increase engagement, zone in on things like opens, clicks, and online interactions.
- Identify problem segments. Looking at goal-focused metrics is a great start, but there are often a number of factors which feed into your goals. How can you narrow down the selection? Well, try identifying your worst-performing segments. Different segments will like different things – so you can learn a lot about your campaign’s issues from the segments you’re having the least success with. Isolate problem segments and look at what they have in common (for example, device, geographical area, age etc). If you find, for example, that your lowest-performing segments are all accessing your content on mobile, you can be reasonably sure that your problem has something to do with mobile optimisation.
- Find the worst-performing ad-groups. Within these segments, extract the ad groups which are performing the worst. Then, as you did earlier when analysing your poorest performing segments, look for similarities within these ad groups. Do they have the same CTA? If email, are they using the same or similar subject lines? Do creative elements like the tone, colour scheme, or imagery have things in common.
- Test. By now, you should have identified and narrowed down a group of common factors in your worst performing areas. One or some of these will undoubtedly be at the root of your campaign performance issues. If you can’t see from your metrics which of these factors is the most problematic, it’s time to do some testing.
Testing your problem factors
Traditionally, A/B or ‘split’ testing is used to narrow down problem factors. We go into how this works in a bit more detail here but, briefly, it works like this:
You decide a factor to test. CTA, for example.
You then create two ads, each identical apart from differing CTAs. So, if you want to determine whether or not the CTA is what’s causing your problems, you’d give ad A the CTA from your poor performing ads, and ad B the CTA from one of your better performing ads.
These ads are then run, live, in front of two separate (but, ideally, similar) audiences.
If ad A continues to perform poorly under these conditions while ad B thrives, you can be reasonably certain that the CTA is, indeed, the source of your campaign’s issues.
This isn’t ideal. It requires you to run an ad that you know isn’t working, in front of live audiences which will think less and less of you the more bad ads that they see.
Is there an alternative?
There absolutely is!
Pinpointing problems using Datasine
If your problems are centred around the creative content of your ads, it can be a long and laborious process trying to find out what’s not working via A/B testing.
However, Datasine’s platform makes it quick and easy to both identify problems with creative content, and to optimise that content without the need for A/B testing.
By using your own past campaign data, our AI is able to establish the kind of creative content your audiences will be most engaged by (and which they won’t). It will analyse your ad content and give you granular insights into not only where your content is going wrong, but also into how you can improve things.
And all without the need for live A/B testing!