CaddieOpen in CaddieDownload for Mac

The Influencer Marketing Math That Made Cal AI Profitable

Shared by Caddie Library · Updated
App Masters10:24Open on YouTube
Chapters show their main points

The bottom line

Keep your CPM below your RPM, the revenue you earn per thousand views. Cap view bonuses near one million views, so viral upside offsets the many videos that lose money.

Key takeaways

  • Influencer views convert worse than paid ads, so they have to cost less. 1:31
    • A paid ad links straight to the app. An influencer viewer must leave the social app, open the app store, search for the app and download it.
    • UGC from brand new accounts must be cheaper still, because the account has no earned brand trust.
  • The most you can pay per thousand views is your own revenue per thousand views. 6:55
    • If your RPM is $5, keep your CPM under $5. A copied benchmark ignores what your business earns.
    • A clothing brand with a $200 order value and $1,000 lifetime value can afford far higher CPMs than an app with $30 and $50.
  • Most videos lose money, and the profit comes from the few that go viral. 4:04
    • A $20 UGC video that gets 500 to 1,000 views does not pay back.
    • One or two videos a month reach millions of views and cover the losses on the other 28.
  • Cap view bonuses near one million views. 4:55
    • A video at one million views often grows to 1.3 to 5 million, and that extra reach costs nothing.
    • Wider audiences are less targeted and convert less. Pizza rat got 100 million views that were worth little.
  • Tracking downloads gets harder as videos pile up, so learn from the early data. 8:48
    • Millions of videos, some years old, still get views and blur which post drove a spike.
    • Early on, each video showed up as a clear jump, such as a baseline of 100 downloads an hour rising to 200. Those patterns and comment sections help estimate later results.

The CPM doesn't matter. All that matters is that your CPM is lower than your RPM.

Jake, Cal AI

Chapter summaries

0:00 Getting creators to cross post

Cross posting is easiest to win when it's treated as a small ask.

  • Presenting cross posting as no big deal gets creators to agree. 0:14
    • The offer is $400 for four posts, with the same videos going on both platforms.
    • The creator is already making the video, so posting it twice feels minor.
    • Jake borrows a trick from arguing with his wife: if you make it out to be a big deal, they'll make it out to be a big deal.
    • Most creators accept it. A few are briefly surprised and say "wait, what?"

1:04 Paid ad CPMs are the ceiling

A Facebook ad CPM is the most you could pay, so influencer rates must sit well below it.

  • Paid ads, influencers and UGC are three channels, and each tolerates a lower CPM than the last. 1:31
    • Steve starts from the CPM he gets on Facebook ads and works back to what to pay creators.
    • Jake agrees, but says the influencer CPM should be significantly lower than that.
    • A new UGC account has no audience that knows it. You're not Mr. Beast, you don't have any brand equity.
    • The order is paid ads highest, influencers next, UGC lowest.

3:03 Base rate plus capped bonuses

Pay a modest base, add view bonuses to reward virality, and stop the bonuses at a ceiling.

  • Each creator gets a base rate and can unlock a few view bonuses. 3:39
    • Base rates run about $350 to $450 for four posts.
    • Bonuses give creators a reason to aim for a viral hit.
    • Steve built a spreadsheet of view tiers with a base rate and a bonus for each. Jake called it a great outline.
  • Capping bonuses leaves everything past the cap as pure upside. 4:48
    • Paying for every view makes it hard to capture the upside of a hit.
    • Jake usually caps at a million or a little more.
    • People can't grasp big numbers. Once a post passes a million views, 1.5 million and 2 million seem about the same to a creator.
    • So a cap costs little in creator goodwill.

5:13 Where to stop paying for views

Creators are happy with a capped bonus, and views beyond the cap are worth less anyway.

  • Paying past the cap is unnecessary and wasteful 5:19
    • Once creators hit the bonus ceiling, they are content, even if uncapped pay would be twice as much.
    • Jake's rule is to stop at one million views.
    • The first viewers of a viral video are fans of that niche. Later ones are the general public.
  • Viral reach reaches people who never buy 6:30
    • A rat video's first 100,000 viewers might subscribe to its Patreon. Hundreds of millions of casual viewers would not.
    • Jake saw pizza rat without ever watching rat content.
    • Past a certain view count, stop paying, since those views are probably worth little.

6:52 Price against your RPM

A fixed CPM target is meaningless, because affordable prices depend on what each business earns.

  • High-RPM businesses can outbid everyone 7:36
    • Telehealth companies also earn so much per customer that their affordable CPMs dwarf a consumer app's.
    • Measure RPM by taking a video's views and the revenue it produced, then scaling to a thousand views.
    • Whatever your RPM is, keep every CPM underneath it.

8:35 Tracking results from viral videos

Attribution gets blurrier as you scale, so early patterns have to guide later assumptions.

  • Learn what converts early, then extrapolate 9:20
    • Note which creator types cause spikes, their content format and what their comment sections look like.
    • In Cal AI's earliest days, no influencer video meant no downloads that day.
    • A spike could be matched to one creator, such as Steve's video that had just posted.

In the growth library

Jake, Influencer marketing lead at Cal AI. Cal AI prices influencer and UGC deals by comparing what each thousand views earns (RPM) with what a thousand views costs (CPM), instead of chasing a benchmark CPM. The episode covers base rate plus capped view bonuses, why viral views lose value, and how to attribute downloads once millions of videos are live.

Lessons that draw on this:

More from App Masters · The whole library, and ask it anything

Transcript319 lines
0:00

When you looking at influencers and negotiating, are you thinking a single platform like Instagram or have them cross post [music] say Instagram and TikTok? >> Yeah, so we always try and have people cross post and I kind of use like a technique I use like with my wife when she's like mad at me about something like she's like, "Oh my god, like it can't be written down." I was like, "Oh no, it's not a big deal. Don't worry about it. Don't worry about it. It's It's all good. Chill."

0:27

Um and that's kind of what I do with the influencers where it's like I'll be like, "Oh yeah, you know, we'll pay you $400 for four posts." And they're like, "And yeah, just cross post it on both platforms. No big deal." And then they're like, "Wait, what?" And I was like, "Yeah, it's Don't worry, you're already making the video. Just post it. It's all good." And like >> [laughter] >> It works It works well. Like they're they're they're cool with that most of the time. It's like if you make it out to be a big deal, >> [snorts] >> they'll make it out to be a big deal.

0:52

But if you just like, "Nah, it's all good." >> [laughter] >> I want to know how long that that works with your wife, okay? Come back with me a couple months. >> It worked when we were kids, not so much anymore. >> 100% [laughter] When you are negotiating some of the prices, 400 for four posts is really good and I'm going to break down some of the tiers that you have in here, the S tier and the A tier. I really want to understand what that means. But you know, what I've done with other ways of doing influencer marketing is kind of looking at the CPMs. Are you guys kind of looking at that, too, and being like cuz I'm like, "Well, I can get this CPM within Facebook ad. So, here's where I would like to be with this, you know, when I'm paying out the influencer."

1:31

>> Absolutely, yeah. I think you want to think of it in like a I I think there's like three primary distribution channels, right? You've got paid ads, you've got influencer marketing, and then you've got UGC marketing. So, paid ads you can afford the highest CPM. You can pay the highest CPM and still be profitable because when people see that ad, there's a link they can click on directly that takes them to your app immediately. So, like that's like the least amount of friction, meaning it's going to lead to the highest conversion rate. So, you can afford a higher CPM than you can with influencers, where when people see an influencer ad, let's say they they see your ad and it's oh my god, that app seems really cool. I want to go and download it. Well, if they want to do that, they have to exit out of their social media app, they have to go to the app store, they have to search for your app, they have to find your app

2:16

within the search results, and then download it. There's so much more friction there. So, like given that added friction, the CPM better be significantly lower than what you're getting with paid ads. And that And then with UGC, these brand new accounts that aren't like established influencers, the CPM has to be even cheaper there than with influencers, because not only do you have the same amount of friction that you have with influencers, but you don't have the brand trust that that influencer has earned over the their years of posting content. You're a brand new account that was just created last month, and no one knows who you are.

2:54

>> Right. >> You know, you're not Mr. Beast, you don't have any like brand equity. So, if conversion's naturally going to be worse, so the CPM better be lower. That's the way I think about it. >> This is so educational for me. So, do you tend to have like a base rate with a bonus? And I can kind of pull up some chart. Some I'm a nerdy spreadsheet guy. So, I'm going to pull this [clears throat] up and kind of give you guys like the way I'm going to think about it. And Jake, like poke holes on this and say like, "No, Steve, I wouldn't do it this way." So, I'm game for that. But like, you know, here's how we started thinking about it. This many views, we have a base rate and bonus.

3:24

What do you think about it? What was your structure, and then we'll get into this dumb spreadsheet that I put together. >> Yeah. No, I I think this is this is a great outline. And essentially like Yeah, you have you you want to come up for influencer, you want to come up with like a base rate. Like, yeah, it's we'll do $450 or $400, $350 for four posts, and then like they they can unlock a couple of view bonuses. One thing I will say though is like what you do you here, you want to cap out, or at least I always wanted to cap out the view bonuses because what you're getting, like the the main value prop for influencer and UGC content is the opportunity of going viral. And like cuz most of if you're doing like UGC in particular, most of your videos are not going to be profitable cuz like they're going to post it, it'll get 500 to 1,000 views and like you know if you're paying $20 a video, like you're

4:19

not making money there. Where you actually make your money is in the one or two videos a month that go viral and get millions of views. So, in order to offset the losses you had for the 28 videos that didn't go viral, you need to capture all of the upside of that viral post and if you're paying for every view, it makes it hard to get that upside. So, I like it's I think it's good to have view bonuses so that the creators are incentivized to go viral, but like you want to cap it at a certain point. That way like everything past that is just gravy for you.

4:52

>> I like it. I like it. What do you guys usually cap it at? A million plus, two million? >> Yeah, usually a million plus cuz like the thing is it's weird like humans naturally are unable to fathom large numbers. Like a million is so much and like but like once you get to a million people like oh yeah like a million, two million, 1.5 million, like they don't think there's a big difference there, but it's like that's literally like two times as much, but they don't see it that way.

5:18

So, like you don't necessarily need to pay it pay for all those extra views. Like once they get that million bucks, they're cool. Like And if a video gets a million views, it'll probably grow to 1.3 to four, five million views and like it's amazing if you don't pay for that. So, I I would usually stop at a million. >> Do you try >> Also, the other thing sorry. >> Yeah, yeah, go for it, dude. >> The other thing is um as videos go more and more viral, they reach a wider and wider audience, a less targeted audience, it's less likely to convert. So, those views are like legitimately worth less. A prime example of like Do you There's this like a rat that went viral a long time ago in New York called pizza rat. And it was like a little rat that was like carrying a piece a piece of pizza. And I bet like at the start like there there's a lot of rat content in New York, right? But that's only getting shown to like the people that

6:13

are into rat content of New York. But pizza rat went so viral I [snorts] saw it, right? And like I'm not a I don't consume rat content. >> Why would you? >> And it's like so when a video gets hundreds of millions of views, it starts to penetrate the gen pop. And then like it's like I wasn't going to subscribe to the rat guy's Patreon account, but like the first 100,000 viewers might. So, it's like the the while he did get a 100 million views on the pizza rat video, they weren't really worth anything. So, that's why like after a certain view threshold like you should stop paying cuz they're probably not going to be worth much.

6:50

>> I like it. I like that breakdown, dude. Is there a CPM that you try to aim for? >> Yeah, so and this is the key thing here is like the CPM doesn't matter. All that matters is that your CPM is lower than your RPM. So, when you your RPM is your revenue per thousand view. So, when you run a like influencer radio, you can see like how much we this video got a thousand views or a million views, whatever, how much did we make from that? And then like you calculate that to figure Oh, yeah, we make you know, $5 per thousand view. And if that's what your RPM is, then you just want to make sure your CPM is below that. So, like you know, some And that that's like the one of the tricky things about influencer marketing is like depending on the space that you're operating in, you could be competing against brands that have ridiculously high RPMs. For instance, like a clothing brand, right? Their average order value is probably going to

7:47

be like $200 and their lifetime value of a customer is going to be like $1,000. And if you're an app that has an average order value of $30 and your lifetime value is $50, it's going to be very hard to compete on an RPM basis with that clothing brand because they just make so much more per thousand views. And um yeah, same thing with any like telehealth company. You know, they just make so much money with their customers.

8:15

So, it's you you just make sure the the CPMs that they can afford are so much higher than the CPMs like a normal consumer app can afford. So, I I like I think the main thing people should take away is like don't focus on like a set CPM. Just figure out what your RPM is and make sure your CPMs always underneath that. >> Okay. Well, speaking about RPM, then uh I have this question, too. Nurik says, "Jake, how are you tracking users if it's coming from viral videos?" So, how do you how are you guys measuring the RPM?

8:45

>> Yeah. Yeah, so um this is something that gets increasingly difficult as you become increasingly successful because and also as you have more videos >> Yeah. >> like up. So, like for instance, like Cali we've got millions of videos now that have been posted. And like there's videos from years ago uh [snorts] that are still getting views, right? Which is like polluting our data. So, it's hard to know like if we see a spike in our downloads one day, it's like did that come from what was posted today or did that come from what was posted a month ago, 2 months ago, a year ago?

9:18

>> Interesting. >> So, the the key the thing you got to do is like very early on when you're like first launching the app, you want to see like what types of creators like cause a spike and like what their comment sections look like, what the content format looks like, so that you can kind of like normalize and like generalize like oh, this type of content converts like this and like when it can when the comment section looks like this, that's like a good sign.

9:42

And yeah, like that that's like the main way that you're kind of you have to do a lot of like extrapolation and make a lot of assumptions based on how your data looked early on when you know, in the early days of Cali, I we'd if we didn't have a influencer video go live, we had no downloads for the day. So, it's like every influencer video that went live, we would see a spike. And then like also as we grew, we may have like a baseline of like 100 downloads an hour and then when a video would go live, you'd see that spike to 200 and we'd know it's oh, yeah, that correlates with, you know, Steve's video that just posted.