Why you should be skeptical of engagement rates

1. Engagers aren’t your customers. General social platform engagement rate is often skewed by hyper-engaged power users that look nothing like your target customer. Just like there are people who just watch tons of linear TV and skew viewership volumes, there are people who spend tons of time on social engaging with stuff. I've seen this at multiple brands. You end up optimizing to the wrong people.

That said: I have had success locking engagement rate metrics only to specific audience cuts (25-44YO, for example), but these kinds of options are increasingly rare in media platforms, and are used as creative diagnostics, not as a judge of winner and losers.

Bonus: It's also true that clickers - outside of SEM - tend to be your lowest-value customers.

2. Platforms optimize for themselves, not you. Social platforms (include YouTube here) are entirely built to keep users engaged and on their platforms - that's their interest. Our interest is to get them OFF those platforms and onto ours. Engagement rate leans into what keeps people on their platforms, not getting them onto ours. Focus measurement on what benefits us.

3. Test the “see/engage now, act later” hypothesis before you believe it. There's a strong POV out there that engagement - especially with "brand" or "story telling" assets - might not drive traffic or purchase now, but will do so later. This would make engagement volume or rate a leading indicator of traffic or purchase. But is it?

Simple test: Take brand and performance assets, stack rank them on traffic conversion rate, then show the stack rank with different conversion windows (1D, 7D, 30D). If this "look now act later" thing is true, then brand assets should lose at 1D and 7D, but win at 30D. Do they? IME, this rarely happens. There are good and bad ads, not good at 1D and bad at 30D ads.Rigorous test: Download monthly engagement rate for your media accounts of choice for 2 or more years. Add to this whatever conversion KPIs (volumes and CPAs, to account for spend changes) you use. Ask AI to quantify the correlation between engagement rate and your KPIs at +0, +1, +3 and +6 months. If engagement rate positively impacts the KPIs, then, by all means, use it! If not, stop talking about it.

Lastly, engagement rate and variants of it (video view duration, etc.) are totally fine as "creative diagnostics" for creative producers, as long as they're clear that those aren't the goals. So, once you're clear on the winners and losers in your creative using the other method, it's totally valid to look at these as secondary metrics.

Again, the risk to avoid is optimizing to what's not in your interest - for example, early/often Brand ID usually negatively impacts video view duration (people see it's "an ad" and bail), but positively impacts business metrics.

We're not a media company here, we're selling stuff.

You will ALWAYS give preference to creative that drives behavior on your platform vs. creative that drives engagement on someone else’s.