The Hard Truth About Scaling: You Have to Ride the Wave

Machine Learning, Thought Leadership

In this post

Short-term inefficiencies occur during testing and scaling. Here's why and how you can prep your team for the bumpy ride ahead to success.

Amanda Evans
October 1, 2021

Congratulations! You’ve achieved great results on your ad campaigns from an efficiency point of view. But that’s not enough.

Your sales team wants more leads, your CMO wants more new customers and your executive board wants more revenue.

“How do you take already great results and increase volume?”

This question is on the minds of nearly every growth marketer, CMO, executive team and board. And, what each advertiser wants to see is something like this:

But the reality of scaling is much different. It’s not a smooth ride. There are ups and downs, trials and errors, hard conversations and amazing wins.

The hard truth is short-term inefficiencies occur during testing.

Failing, learning and succeeding are required for long-term growth. Successful advertisers learn to surf these waves while skittish advertisers wipe out when the first wave hits.

Here’s an example of how a typical campaign scales:

During the periods of scaling, the CPA (yellow line) increases along with the conversions (blue bars). After a few weeks of scaling, the CPA comes back down as we optimize based on the learnings occurring during the scaling phase.

Then when it’s settled, we make another scaling push which temporarily increases the CPA before it comes back down again.

Why Does CPA Fluctuate during scaling?

With each scaling push, we’re making some changes to the campaigns — perhaps it’s increasing the budget or increasing our CPA bids to capture high-quality leads. With each change, the machine learning algorithms go into a learning phase.

During the learning phase, the algorithm is fishing in new ponds and testing new audiences, ad inventory and placements to determine the best mix. The inefficiency you see during this time is essential as the algorithm learns what doesn’t work.

A common mistake is to interfere with this process and make another change (even rolling it back) to try and “fix” the inefficiency. But savvy advertisers know that it’s best to leave it alone and let it work its magic.

How Long Is The Learning Phase?

In reality, the length of your learning phase depends on your data volume, audience saturation and bid strategy. In some instances, it may take two to three days. For others, it’s an entire month.

Our rule of thumb for Google when campaigns are on automated bid strategies is to increase roughly 20% every two weeks.

Facebook is much easier to scale with 10-20% increases every four to five days. If you don’t see improvements within that time frame, it’s time to intervene!

Scaling Efficiently

Efficiently scaling is a process, not a one-day event. To ride the wave to success, we recommend the following tips to help ease the apprehensions of stakeholders:

1. Establish tolerances for CPL/CPA

Let’s say you want to double volume while keeping a CPA of $150. Knowing that there will be fluctuations in the CPA during the scaling process, there needs to be an established tolerance for increases.

For a period of three to four weeks, what is the maximum CPA that you can tolerate? This needs to be an honest conversation with your agency and internal teams to establish boundaries and expectations.

The wider your tolerance, the faster you’ll be able to scale.

2. Prepare your stakeholders

Even if the scaling process is intuitive to you, it’s not for all of your stakeholders. Explain the realities of scaling to your executive and sales team so they can understand and prepare for the process ahead.

It is also recommended you also share learnings and improvements along the way throughout the organization.

3. Loosen your grip and trust the process

During the scaling process, the hardest thing to do is to trust the platform algorithms enough to leave them alone. We want to control every little thing and to see those improvements quickly.

But interfering during a learning phase is a crucial mistake. Instead, take the time upfront to ensure tracking is set up correctly and that the scaling test is optimizing to the very best metric. Then loosen your grip and trust the process.

The Key Takeaway

With a proven pattern of scaling — optimizing, scaling, and optimizing — growth marketers can achieve massive success. We’ve seen already well-established companies increase their volume by well over 600% using this process.

Scaling and this type of growth is not for the faint of heart, as KPIs will dip and recover. But a little tolerance, preparation and trust can make massive growth a reality.