When Your Product Changes & Results Tank

I work with quite a few start-ups, which means they are trying to nail down some of the particulars on their product or offerings. Things like pricing, shipping times, features, etc.

It’s not uncommon for small businesses in general to pivot what they offer in order to meet market demand. Hitting things just right on the product, price and promotion can be a tough target to hit!

It can create unexpected fall out on the Facebook Ads side, but sometimes not in the way clients or even managers immediately assume.

When You Change the Freebie/Hook

This client changed their offer mid-August, going from a freebie to a small paid trial. It looks like when this happened, the users were having no part of it – check out that terrible CPA and the plunge in conversion volume:

We relied heavily on lookalike audiences in this instance because they have performed fantastically on their own. After this? Not so much.

But, as the weeks passed, their data started to show that overall, conversion was NOT falling across the board. Users didn’t seem to mind paying a few bucks, and they were more committed to purchase more over time, so the value of those users was a lot higher than the freebie-seekers.

We initially scratched our heads, until we realized something:

The lookalike algorithm on Facebook was struggling to match people. We had basically jerked the wheel on it.

The success of the campaigns had hinged on the conversion data, and the users associated with that. While the conversion data was still coming in, the demographic had shifted, and Facebook was all:

 

It couldn’t figure out who we wanted it to find because the past 30 days had such a large shift in the patterns and tendencies of the audience we were sending to it. The offering wasn’t a failure, the targeting was.

We got rid of the lookalikes that were relying on on-site purchase, and instead uploaded a customer list of ONLY those who had purchased the new offer.

Then, magic happened:

Facebook knew who to look for, and started rockin’ it out once again.

When You Jack the Price

Another client also changed their price AND the shipping date of the product to be a lot later. This created a lot of challenges, because we had no idea if people hated one or the other. Either way, the lookalikes we had relied on started failing, hardcore:

We started to make headway with the new audience list, but also had accepted it  may never return to its former glory on the lower CPA and higher volume we’d been generating – the market for the revised offering was simply a lot smaller. This was also confirmed by what they saw from every channel.

Sweating bullets?

There are a few ways to handle product changes and the anticipated impact to your advertising:

Recognize ahead of time your results will shift.

You may not be able to exactly quantify by how much, but be prepared for anything.

Stay in close touch with what’s happening across ALL channels.

This can be a faster indicator of whether something’s broken specifically in your paid setup, or if the audience is responding the same way to the change no matter how they are exposed to it.

Consider pausing lookalikes temporarily.

The Facebook algorithm has what I call a “death spiral,” where sometimes once an ad set gets stuck in a certain vortex, it just cannot pull out of it.

Indeed, in our second example, I launched everything in fresh ad sets and started to make some traction faster.

If possible, you may want to consider pausing ad sets that use a lookalike, and swap out the reference with a newer customer list, or more recent traffic windows once there’s been time for some conversions to happen and data to collect.

Experiment with your window in conversion-focused campaign types.

Something else that can also help is checking out how you have the Facebook algorithm calculating your bids:

You may want to experiment with how long that conversion window is or isn’t on the heels of making large product changes.  You may also want to enable to the option for the platform to focus on link clicks for a bit until conversion data racks up more reliably.

Any other methods I’m not thinking of to help stop the bleeding? Let me know!

You may also like...

Leave a Reply

Your email address will not be published. Required fields are marked *