5 Steps to Diagnosing Facebook Ad Performance Problems

5 Steps to Diagnosing Facebook Ad Performance Problems

Do you ever run into Facebook ad performance problems? For example:

  • Your campaigns suddenly aren’t performing as well?
  • Your costs start to go up?
  • Your new campaign doesn’t perform nearly as well as previous campaigns?

We do mentor and train agency folks like Facebook ads specialists, and one of the big skill gaps we frequently have to train for are:

  • The ability to discover the real cause of a performance problem so that you can take the right action to fix it.
  • An understanding of the most important metrics (besides your key performance indicator) and how to use them to diagnose the problem.
  • Comparing present and past performance properly to come to accurate conclusions.

Without these skills, what happens is:

  • Specialists and clients are uncertain what’s wrong.
  • They tend to grab at straws about what might be the problem.
  • They agree on what they think is the most likely cause without looking deeper into the numbers to find out for sure.
  • They end up take the wrong action, and get worse results and missing out on performance opportunities.

That’s like someone treating a disease without doing any lab or diagnostic tests- often considered malpractice.

So let’s remedy this!

Here’s how to diagnose Facebook ad performance problems, in 5 simple steps.

But first, a few basics bout the metrics and how they work:

No alt text provided for this image

In the above funnel diagram, you can see the most important metrics we need to look at, and how they are arranged as we drive a user from ad showing (impression) to click (website traffic) to conversion (result).

  • CPM = cost per 1,000 (mille) impressions, AKA the cost per 1,000 ad showings
  • CTR = clickthrough rate, a measure of how interesting your ad is, leading to clicks, AKA website traffic
  • CPC = cost per click, or cost per visit to your website
  • RR = results rate, is the % of ad viewers who convert (to a lead, sale, or whatever you’re trying to get them to do)
  • CPR = cost per result, AKA cost per conversion

Also important to note is that these metrics are all rates, not total numbers. Clickthrough rate (CTR) is more important to gauging how interesting your ad is than its sheer number of clicks, just as miles per hour is more important to gauging how fast you’re going than the # of miles you’ve gone. Total numbers can actually distract you here, and are not important for now.

Start by going into the Facebook ad dashboard and getting your columns right. Here’s what I recommend:

No alt text provided for this image

Save that column set as your Facebook ad diagnosis column set, or something like that.

STEP 1: Find a Good Date Range for Comparison

To say anything is wrong, it has to be wrong compared to something that’s right. So we need to figure out what we’re comparing, A and B, and what the date ranges for each are.

Your comparisons could be things like:

  • A: your current campaign, the most recent two weeks vs. B: the same campaign, the most recent two weeks of the previous year
  • OR
  • A: your current campaign, the most recent two weeks vs. B: a different but similar and better-performing campaign, the most recent two weeks

When it comes to choosing the right date ranges, here are some things to keep in mind:

  • When you compare different time periods (e.g. this month vs last month, or the last 14 days vs the 14 days before that), there could be seasonal variation. Things like news, economic changes, holidays, and even weather can affect performance. If you think about the Great Recession, sometimes changes can last for a long time and can even throw off year-to-year comparisons. So if you are comparing two time periods, think about why they might differ.
  • It’s always best, if you can, to compare either two different campaigns over the exact same date range, or your current campaign this year compared to the exact same date range the previous year. You’re more likely to get an apples to apples comparison, but…
  • When comparing year to year, keep in mind whether the industry, the markets, the economy or anything else has changed significantly. Also, CPM tends to go up each year on both Facebook and Google, so you should expect some change there. Sometimes we’re talking 5-10% but sometimes these changes are huge, as when AdStage reported Facebook CPM’s nearly tripling (+171%) from 2016 to 2017. So look for industry reports on metrics, and also benchmark year to year changes across multiple Facebook ad accounts if you have that kind of access.

STEP 2: Compare CPM, Which Is the First Source of All Subsequent Cost Metrics in the Funnel

First, we want to see if the basic cost of displaying ads has gone up. This is measured in cost per impression (CPM), or cost per 1,000 ad showings.

There are a few different causes of higher CPM’s:

  • Year over year platform cost increases, as described above.
  • Changes in your ad set targeting; you may have added new targets that are more expensive.
  • What your ad sets are optimizing for; if you’re optimizing for impressions, Facebook will find you cheaper ones. If you’re optimizing for conversions, it may be more expensive to show to that higher quality subset of people.

Now compare: Look at the change in CPM, and compare that to the change in your CPR. Is it close enough to completely explain the CPR change? Or was there minimal change in CPM, or even in the opposite direction of the CPR? For example, if CPR went up but CPM went down, CPM is not the culprit.

Check out the following example:

No alt text provided for this image

In the first campaign (first row), CPR went up 40.95%, but CPM only went up 13.9%, so CPM played a part, but there’s more to the story.

STEP 3: Compare CTR (Clickthrough rates)

The combination of CPM and CTR is going to give you your CPC. So we want CTR to be as high as possible. It’s reasonable to expect CTR’s as high as 8% at times, but 1-3% is more normal.

Clickthrough rate can go down when:

  • Creative burns out: you can’t run the same ads, videos or text forever in the same audience target, because eventually everyone has seen it and it no longer works. Create more stuff!
  • Frequency is too high: If your spend is too high for an audience size, your frequency may go up to 2, 3, 4 or higher, and this overexposure can lower CTR. You might want to do that in a reach-oriented campaign, but in conversions, it could hurt your CPR.
  • Spend is too high for audience size: When you have a good campaign CPR, you want to “turn it up,” as in spend more. And you can, to a point. But you’ll eventually hit a threshold where you’re dipping into lower quality parts of audience, and CTR will go down, and possible CR as well.

If you are seeing a big CPM increase, then if you can lower CTR, you will lower CPC.

But if your CTR remains the same with a higher CPM, CPC will go up, too.

STEP 4: Compare Cost Per Click

Ok, admittedly, this one is not super useful, because CPC is a function of CPM and CTR.

No alt text provided for this image

CTR changes and their effect on CPC

In the above example, you can see that usually when CTR goes down, CPC goes up, and vice versa. But these numbers are not exactly the same, so we know that CPM changes also played a role.

Most of our solution is going to be focused on

  • Controlling CPM (through audiences),
  • Controlling CTR (through creative), and
  • Controlling RR (through audience, creative and website).

Check your CPC changes, but focus more on the three metrics above!

NOTE: one thing you may want to do with CTR is make sure you’re hitting industry and overall Facebook benchmarks. In the example chart above, all CTR’s are below 1.0%, which indicates there may be a problem with the creative not being compelling enough to the audiences.

STEP 5: Compare RR (results rate), the % of people who visit the website and convert

It’s important to find out if the percentage of people converting on your website has gone down. Sometimes this is a temporarily seasonal type of thing, but sometimes it’s a new competing offer in the marketplace, or perhaps the website has accidentally been changed and you didn’t realize it.

Always check the website and make sure everything looks right and is functioning correctly.

If RR goes to zero, make sure the Facebook pixel is still there and the conversions are still set up correctly.

Nerdy point: RR is results per impression, which is the % of people who saw an ad and converted. Most people also look at conversion rate (CR), which is just the % of website visitors that convert. This is a more focused measure of your website’s ability to persuade. CTR and CR can balance each other out sometimes- one goes up and the other goes down- what you want is for both to go up. When you have relatively similar CPM’s but the CPR is dramatically different, sometimes both CTR and RR are better, combining their effects for a much lower result rate.

No alt text provided for this image

Even when CPM is higher, the best cost per results may come from the combo of the best CTR and best CR.

STEP 6: Focus on the area of biggest change when devising your performance solution

When you find the biggest problem area, you can focus your solution there. For example:

  • Check your ad sets to see if specific audiences have high CPM’s and are dominating the spend too much. Look at their CPR’s, and if they’re not good, you may want to pause them.
  • If CTR is the problem, look at which ads have better and worse CTR, and go more in the direction of the better ads. Create more new ads along those lines. Pause the bad ones.
  • If CR (response rate) is the problem, it could be your offer, your website, or it could be that particular audience. Look at your ad sets to see if some of them are not converting very well. But if all CR’s are low, it is more likely to be the entire offer, or the website.

And that’s it!

You now know how to diagnose where the biggest problem with your performance is coming from.

And diagnosis always comes before proper treatment… so now you can go forth as a Facebook Ad Doctor and avoid malpractice!

要查看或添加评论,请登录

Brian Carter的更多文章

社区洞察

其他会员也浏览了