The logarithmic scale of 1 to 10, with 10 being the highest. This uses the same logic as Google’s Quality Score. It is driven by CTR (Click Through Rate) / engagement as the primary factor rather than negative feedback, landing page quality, and so forth. Facebook shows only for ads, but we know they calculate it for all posts, boosted or not.

They used to reveal this by utilizing a News Feed point score, much like a batting average. Since video generates more engagement than all other post types, you should have higher in video, especially since Facebook counts a view as engagement.

I’ve long believed that engagement of any sort requires a click. My recommendation to Facebook has been to exclude autoplay views under 10 seconds.

There is no single bar to meet. Most people ask me for a particular number, so they can use this magic threshold to decide when to kill or add budget to an ad. Instead, you want to have at least a 6 for the top of the funnel, at least a 4 mid-funnel, and at least a 2 at the bottom of the funnel. And, yes, you can use video at every stage of the funnel.

Video will help convert and persuade users better when you add it anywhere, just like adding bacon makes everything better! If you trigger a video based on a custom audience, you might get a lower VTR and if you are using the same videos over and over or using videos that may look the same (similar thumbnails).

Burning people out on video is easy to do when your budget is high relative to the size of your audience. For example, if you are over a frequency of 2 per week per ad set, you risk burning out that audience.

All of the metrics above are your “check engine light” letting you know if one of your videos is approaching the burnout point. Instead of killing the video, usually, you can reduce the budget and keep that ad evergreen, allowing the frequency, VTR, and other metrics to stay healthy.

If this video is driven by a custom audience (perhaps folks who watched video A were on your website yesterday, or just bought something), then growing that upstream audience will let this particular video spend more and perform better. So be watchful of how audiences move down the funnel when you sequence videos together, like the bulge in a snake that has swallowed a mouse.

A lot of people who don’t spend time analyzing their video performance like to say that Facebook’s video metrics are inflated or aren’t accurate. For those who do, the top complaint is that the default view is counted at only 3 seconds instead of 10 seconds.

We don’t believe there is a magic number since what’s important is the view curve, which shows how long people stayed from start to finish, showing the decay from beginning to end.

The kinks in this curve are what matter if you care about improving your performance since any particular metric doesn’t tell you the whole picture. Plus, averages lie; they wash out the underlying trends.

YouTube marketers like to say that YouTube yields “better” quality views since a TrueView is counted at 30 seconds instead of 3 seconds. I would generally agree with this but have to point out that paying $0.50 cents for a 30-second view isn’t necessarily better than spending $0.02 cents for a 3-second view.

What matters is what your decay curve looks like and whether your content is arranged correctly to take advantage of the differences between Facebook and YouTube.

For example, the best practice on both Facebook and YouTube is to have a “hook” (attention-getter) in the first 3 seconds (not your logo or bumper), which leads into igniting pain or pleasure.

On YouTube, you want to explicitly disqualify people from watching from 15-30 seconds, since you are charged for a view only when people stay for 30 seconds. In other words, you might say “If you are facing X as your issue and understand that it costs $Y, then please continue watching since this message is for you. If this doesn’t describe you, please don’t continue.”

This tactic of explicit disqualification by mentioning your important buyer criteria works great on YouTube. However, copy this same video to Facebook, and you might wonder why it’s not performing, perhaps even claiming Facebook “doesn’t work.”

The same is true for brands who take their 30-second and 60-second TV spots and copy them to Facebook. Seems logical, right? Not when most of Facebook is autoplay, sound off, and on mobile. These users have shorter attention spans, different expectations, and can skip (except for the new in-video ads).

TV ads are based on the fundamental assumption that you have a captive audience who is sitting down to watch a show, instead of giving you partial attention while they’re on the go.

Another bucket of general misunderstanding is around derived metrics, which are the calculated metrics. For example, you can’t add up organic and paid reach to get to total reach. Some users were exposed to a video organically and paid, so they should be counted only once for total reach.

This is still a bug on Facebook’s side and has been for years. The same goes for adding up total watch time for one or many videos and expecting it to equal the average view time multiplied by total views. Remember that Facebook doesn’t count the time that is under 3 seconds when counting views.

How do you know if your video is performing? By checking your statistics against these benchmarks:

• Reach - this will vary based on the target audiences you've chosen
• Views - ideally this should equate to 50% of your total reach. So for every 1,000 reach you'd want 500 views
• View Length:
• Less than 3 seconds - poor
• 6 seconds - average
• 10 seconds - good
• 15-20 seconds - great
• 30+ seconds - you're a Unicorn!

From there, we like to optimize further. We want to see views turn into 10 second views as follows:

• Fewer than 30% watch the first 10 seconds - poor/average
• 20-40% watch the first 10 seconds - good
• 50% watch the first 10 seconds - great

We're focused on producing and optimizing 15 second videos, so we want to look at completion rates. If we are over 20%, we've got a winner.

The longer the video, the lower the completion rate. So that's why we are making it so short.

Short-form video primary metrics:

• For Lead Gen: $3 CPL.
• E-commerce: $10 CPA.
  

This section will help you understand the basics of a framework called #MAA (Metrics > Analysis > Action) in conjunction with #CID (Communicate > Iterate > Delegate).

Before we can start using the #MAA Framework, you'll need to ensure you've installed the TikTok Pixel and you'll want to ensure that you've set up conversion tracking correctly in your Google Analytics account with TikTok as a source. 

In optimization (to help us with conversion), you’re using data to answer these questions:

• How are my different marketing channels performing relative to one another, and where am I bleeding?
• What is the right amount of effort and spend to place on Pay Per Click, email marketing, Search Engine Optimization, offline campaigns, and social media?
• How much provable profit, Return on Investment, and margin are being generated by each of my marketing campaigns?
• Based on the data, what are the top 10 specific things that my organization can do to improve metrics?

So how do you move from metrics to analysis? The answer is in Metrics Decomposition. Metrics Decomposition is what we use to help make sense of changes in results we're tracking.

A seemingly complex concept, Metrics Decomposition describes the relationship between many different metrics. This relationship enables you to make sense of changes in results, like a sudden change in cost per action or a spike in conversions. It keeps you calm because you will be able to figure out what went wrong/right.

• This is the profit formula behind Facebook ads -- not sales puffery, but a mathematically derived formula.
• It starts with applying the conventional profit formula: Profits = Revenue - Cost.
• Revenue from TikTok = Value per conversion x No. of conversions. If an apple costs $5 and you sell 5 apples on TikTok, your revenue is $5 x 5 = $25.
• Cost from TikTok = No. of conversions x Cost per conversion. If you spent $3 to sell an apple, your cost is $3 x 5 = $15.
• Your profit is thus $25 - $15 = $10.

Unfortunately, this is where many beginner marketing consultants stop. Let’s say, today, you sell 5 apples at $5 each, so you don’t make a profit. How would you go about finding out what the problem is?

Here’s how we would do it:

• Compare the conversion rates today versus yesterday. Did fewer people buy per 100 visitors?
• If not, compare the visits on both days. Did fewer people come along?
• This or the previous point must be true.
• If conversion rates stayed the same, but visits dropped, compare click-through rates on both days. Did fewer people click on my ad per every 100 impressions?
• Otherwise, compare the number of impressions served. Were there fewer impressions?

We could go on and on for the next few paragraphs, but you get the point.

Troubleshooting Conversions

Your profit is a function of cost and revenue, maximized where marginal revenue equals marginal cost. A bit of first semester economics will help you find the spend level that maximizes profit, which is not the same as maximizing margin or revenue.

Troubleshooting Conversions

Summary

• Just like on Facebook, instead of trying to promote 50 random things in 50 random ways— keep honing the one thing that's performed the best.
• TikTok ads are driven by consistently sharing engaging content. You need a constant stream of content to target your audience and feed your TikTok ads.
• The best videos for TikTok are 15-23 seconds long and involve movement, recorded on your smartphone with clear headings, text, and call to actions.

“I tried it, and it didn’t work” is one of the most bone-headed things we hear from clients and wannabe marketers. One shot, and they’re done.

In digital marketing, there is no penalty for taking many shots on goal. We learn quickly and iterate constantly. We expect failure 90% of the time.

Make another ad, try another subject line, choose another set of keywords, adjust your budget, tweak the landing page, and so forth. Via many cycles of experimentation, we’ll undoubtedly find some winners.

The Optimization Course will help you efficiently go through this testing cycle via a framework called MAA (Metrics > Analysis > Action) in conjunction with CID (Communicate > Iterate > Delegate).