How to scale up my Facebook advertising campaigns without hurting ROI?
One of the biggest problems I come across with 99% of people is that they want to scale their campaigns … but when they do it fails miserably!
The questions they often ask me are:
1) When I try and scale, I lose ROI, why is this?
2) I’ve tried to scale by duplicating my ad-sets, but soon after, it crashed and burned, why?
3) I’ve had a couple of good ad-sets, but when I raised my budget by 20-30%, my CPA went up too high, what can I do?
4) I found a winning product, and I’m getting 3x ROAS, but as soon as I try to scale it tanks every time, what am I doing wrong?
Let me explain a few things and share some needed wisdom … And you may not like what I’m about to say, but you need to hear it anyway …
First, you can’t scale crap. If you have a product that is poor quality, doesn’t do what it’s supposed to do, takes forever to ship and misleads the prospect into buying it, then you’re going to be wasting your time, energy, and money.
In order to scale you have to have an awesome product that resonates, that is highly relevant, fixes a major issue (or provides an amazing pleasure/entertainment) and that it’s perceived to be really valuable in your your prospect’s eye, so much so that they buy!
You don’t have the audiences built correctly to scale up into. You’ll all be aware of custom audiences and how you can expand these into Lookalike audiences. But the issue here is, people don’t know what type of CA to build, when, or which Lookalike Audience to create, or which to use first.
Generally, most people have to start with interest targeting in the beginning … and this is great, but you need to build your CA and LAL’s from day 1 …
As the data is being captured (both Pixel events and within your CAs), you need to be segmenting your initial targeting into better audiences to scale into. Your interests may be good to begin with, but I can guarantee, it won’t be long before your CPMs rise and your ROAS drops…
Your CAs allow you to segment the traffic … But most people don’t create them based on where the event is taking place in the funnel, nor use timed CAs, and timed LALs.
Let me explain:
As you drive traffic, you should be placing different event triggers, and be slicing (segmenting) these further for two reasons.
1) First targeting based on ‘Intent’. You should be retargeting immediately.
2) Second targeting based on ‘Time’. As time passes buyer intent decreases.
Let me give you an example:
Let’s say Sarah visits my store, and she places a product in the cart and abandons.
You should have a CA for your Add to Carts. Sarah should now be in that Custom Audience. So she’s shown a level of ‘Interest’, but for whatever reason she bailed … Your job now, is to retarget Sarah by using that CA, and try to recover the sale.
Most marketers do a poor job with retargeting, if they do it at all … Don’t be like them.
You can do sequential retargeting, by creating ‘timed’ CAs. Example:
3 day, 7 day and 10 days. A simple retargeting follow up sequence could use those CAs, and you could offer an incentive across all of these, like a reminder of the product day 0-3 days, a 10% discount day 4-7, and offer 20% days 8 through 10 …
Did you know these audiences are dynamic … as you drive traffic, people are entering and exiting these different segments …
You can create Lookalike Audiences for these segments too.
Do you think a 3 day Lookalike of buyers would work?
How about a 7 days or 14 days?
How about your Fanpage Engagement segment of 7 days … Would a 1% LAL work …
Let me also explain something to you … If you try a 1% lookalike audience of 180 days and it works … Why wouldn’t you try a 90, 60, 60, 30, 14, 7 and 3 day 1% Lookalike?
How about putting together (combining into an ad-set audience) several percentage lookalikes to scale:
1% LAL 3 Day Buyers
1% LAL Fan Page Engagers
1% LAL Ad Post Engagers
If you’ve got Lookalike audiences working and your pixel is working hard and finding your buyers from the expanded audience, why stop at 1% Lookalikes.
People assume, because they’ve tried a Lookalike audience or a 2% or 5% Lookalike audience before, and it failed that it doesn’t work …
Remember, they are dynamic … They are forever changing based on the CA data it’s built from.
And (from my own and many other top marketers), they try higher level Lookalikes too, just because the 1% didn’t work this week doesn’t mean a 5% isn’t going to work …
Same with if you tried them before (weeks/months ago) and it didn’t work … It’s dynamic … so make sure you retry them …
Heck, I’ve even re-built them when they’ve stopped working, and they start working again.
Some of my best working audiences are a combination of 1% and 8% Lookalikes created from different timed CAs.
John Hutchison is a Serial Entrepreneur, an online marketing and eCommerce expert.