Using social algos for better ads...

Just shared something in the GM+ community this morning and I thought I would also share it here as it could be pretty useful for you as well.

As mentioned in this, I'm looking at how to automate this at scale for myself in GM+ - if you wanna help with the buid or be in with the initial testers, you can join us here.

Now, onto the post...

---

Big thanks to MT for turning me onto the guy I’ll be breaking down in this post.

Got me really thinking about how to improve my social marketing strats, but also helps illustrate a point I think might be the future of better ad spend.

I’ll start with the latter.

Using social algos to improve ad strategies

So, when building the social system for VP, I took a slightly different approach with the post strategy.

I found competitor accounts and looked for the posts that had views, prioritised them by the number of views against their follower count.

Any post that got 2X+ the follower count was worthy of “stealing” (I’d copy the format and create my own version).

The reasoning here is simple.

If an account has 10k followers, but a post has only 5k views, then it wasn’t good enough to break out of their follower count and leverage the algo.

If that same account has a post that did 20k views, it broke free of their follower base and the algo started showing it to people outside their existing base… essentially growing the reach and audience.

These are what you should be looking to copy and replicate for yourself. These are proven formats that work well on the platform.

Where do ads come into this?

When you run ads, you are limited by your spend.

Starting off on a $20 / day budget will limit you to maybe 2k impressions (average CPM being $10 according to Google).

The algo isn’t gonna do you a favour here. It’s not gonna give you more reach or see how it can help. It’s not gonna grow your reach beyond that initial 2k people cause that’s all you’ve paid for.

So, the creative testing is up to you, and the speed of the data is down to your spend.

For years, I liked this as more spend = more data.

But when you pay for ads, you guarantee placement, but limit reach.

Which means the data you get back is limited to the budget you can afford. And if you’re running a bad ad, it might take you a few days to figure it out. A few days that lighten your wallet.

The algos on the other hand already have all the data for what is going to work on the platform.

They’ll never tell you what that secret recipe is though.

You have to figure it out for yourself.

The easiest way to do that is to look at what’s already doing well craft your own version of it, and then run it to see what happens.

The feedback loop after publication is pretty quick.

I noticed when I had some posts that were destined to do well, you’d see it in the first 30-60 minutes as the algo pushes them hard.

The downside is the social world moves on fast.

A post that’s killing it this week will likely die off quickly next week.

It’s why so many creators have very lumpy cashflow.

And that’s where ads come in.

You have a proven post that gets reach, engagement, and clicks. When you know what the audience wants, turn the post into an ad and make sure it keeps getting placement with your ideal audience for weeks/months.

I think this is the better pipeline and something I’m testing now.

The algos will tell you what works well for your audience.

Use that to then find the winning ads.

An example…

1m+ views on a 3k account

MT sent me this recently. Commenting on what a great post it is.

Dude’s obviously selling some kind of automation service, but the ad is just a fun thing to watch on Insta (best ads are not immediately identifiable as ads)

​https://www.instagram.com/reel/DahhGBGJO6v/​

Funny thing is, this guy has an account which has 3 posts, 3200 followers, and this post with 1M+ views.

Man’s getting 1k followers per post right now.

And the post above has like 2.5k comments.

With the CTA being “comment Bali for thing” we know it’s a ManyChat lead gen type piece.

If we assume half of those are him replying with something like “DM sent”, we’re at 1250 top level comments.

Average CVR for comment to DM for lead gen is 25-55% according to Google.

So this fella likely pulled 250ish people as leads from this one post.

It doesn’t appear as though he ran ads to this, so he could keep this going as a lovely little lead gen machine with a modest spend behind it.

--

If you want to help me with this build, or get early access to it as a GM+ member, you can join here.​

Speak soon,

Pete "automating social" Boyle