Recap:

In the first post, you learned the 3 Rules Of Paid Traffic and why you need to become a master at lowering your CPAs to create "room in the math" to scale.

In the second post, you learned how to break down your math to figure out WHERE your greatest problems / opportunities lie and how to build useful priority lists for your team.

In the third post, you learned how to take a step back to develop hypotheses as to WHY your math is breaking down where it is.

Now that you have your hypotheses, we need to talk about how you develop and deploy experiments to figure out what really "moves the needle."

And that's what we'll cover in this post.

The Truth About Developing Business-Changing Experiments

I would love nothing more than to give you a step-by-step formula for turning your hypotheses into guaranteed winning experiments.

But the truth is, that formula doesn't exist.

The way you find wins with paid traffic is by developing sound experiments, deploying them in a smart and consistent fashion, and being sure not to ruin them by scaling too fast.

All of the work done in Phase 1 and Phase 2 helped develop hypotheses that you can be confident WILL improve your ability to scale if you figure out how to change the numbers.

What you need now is a process for consistently creating and deploying experiments to determine which of your ideas actually moves the needle.

To help you do that, I'll introduce you to our 3 step process.

The 3 Step Process To Develop & Test Experiments To Work At Scale

Step 1: Spend Time Designing Your Experiments

We are all about doing things quickly and cheaply to see if an idea has the ability to produce results.

But when it comes to designing experiments, taking a little extra time is REALLY important.

And there are 3 things you need to consider:

Consideration #2: How You Say It:

While the words are vital to ensuring a clear message, you need to think about how to most effectively translate those words into pixels on a website that deliver the message.

This includes considering things like:

You won't be able to make bad copy convert with good design...but you absolutely can prevent good copy from converting with poor design.

Consideration #3: Trackability:

The #1 advantage of digital advertising is the ability to track everything...yet so many companies don't have tracking engrained in the DNA of their company.

Your experiments MUST be designed with tracking in mind so you can review the data to see if your experiments really created the results you were hoping for across the entire sales process.

Tracking Pro Tip:

I HIGHLY recommend deploying your experiments (if possible) as split tests on the ad network so each creative / sales flow has its own unique Ad ID.

This is really important because it will help you see if your experiments improved all of the metrics you wanted or if it improved one but decreased another you didn't want.

Lessons From Our Failures:

More than once in the last decade, we've deployed an experiment that greatly dropped CPAs...but AOV dropped as well.

If we'd simply run a split test on the landing page, this data probably would have gotten lost in the shuffle and resulted in us pushing out a "solution" that did decrease the cost of a customer but also decreased the value of one as well.

Our goal is to find wins that improve the overall profitability of a customer acquisition process which means we need to be able to monitor all of the variables...and deploying with unique Ad IDs helps us ensure this.

Example:

In this episode of our podcast, The Scaling Game, I lay out how we went to great lengths to develop the best possible lead capture experiment that served our needs for marketing, sales, customer experience, and tracking.

Step 2: Test The Same Way, Every Time

One of the biggest issues we see in clients that have gotten stuck in the scaling process is that they allow their media buying to be another variable in their testing.

There are no SOPs in regards to how a test is to be conducted which means the choices the media buyer makes are likely to impact the outcome of the experiment.

We ensure all of our experiments are built the EXACT same way every time:

All of these cemented parameters allow us to run experiments that put 100% of the outcome on the creatives themselves.

Media buyers don't get to make any decisions about how to run / scale campaigns until a creative graduates out of the testing phase.

Step 3: Scale To The Math

Once you find experiments that are showing "smoke" inside of your cookie-cutter testing process, you may feel like you've "cracked the code."

And you might have.

But until you start to run something at the same budgets as your controls, you can't be sure that all of the early returns are long-lasting.

You need a systematic approach for scaling your budgets to ensure that the returns track right along with it.

We call this process "Scaling To The Math."

To ensure we do this every time, our head of media buying, Ross Pieterson, created what he dubbed "The Scaling Ladder."

It's a set of rules our media buyers follow for scaling campaigns from our testing budgets up to the highest levels of spend that the funnel performance allows.

Every client has a different "ladder" that's based on their CPA goals and allows us to scale budgets in a way that keeps us out of the "learning phase" and in control of the numbers.

And our rules are very simple:

These simple rules allow us to manage a ton of high-performing experiments for our clients that produce consistent volume because we're not making ad-hoc decisions about what to kill and scale.

Three Additional Benefits Of "Scaling To The Math"

Our process of "Scaling To The Math" not only ensures your experiment's performance will actually hold at higher budgets, it also provides three additional benefits in the long run.

Benefit #1: More Sales Consistency:

One of the major issues I've seen clients face in the pace was the inconsistency that came from paid traffic and affiliates.

They'd see ups-and-downs because media buyers would try to "get while the getting is good" which would create huge boom-or-bust periods for the business.

Our Scaling To The Math process allows for far more consistency because every ad is being run at its maximum ROI point (and then left alone to continue to produce).

Benefit #2: Less Creative Burnout:

I've run several teams that have had to create hundreds of ads per month to keep up with media buyers who "burn out" creatives by scaling them too quickly.

By Scaling To The Math, you reduce creative burnout because you're not "red lining" ads with the goal of squeezing every last dollar out.

Your goal is to run as many ads simultaneously for as long as possible in order to maximize your ROI.

Benefit #3: Easier Migration

I've always been a big believer that platform diversification was important to long term scalability but it wasn't until we put this Scaling To The Math process in place that I got to see its full potential.

By being more measured in our testing and scaling process, we are able to move proven creatives to new platforms and determine if they will be winners far faster.

We're essentially able to completely remove the question "does this ad / funnel work?" from our thought process and just focus on optimizing the economics.

Example Of This Process At Work:

In the video below, I've explained the process we used to test and scale campaigns for a client that required 17 rounds of iteration over the course of 27 days and ultimately reduced their CPAs by 97%.

Key Takeaway:

The main lesson you should takeaway from this post is that turning your hypotheses into well-designed and consistently executed experiments is the key to long-lasting scale.

If you try to "game the algorithm" you won't be able to produce wins that you can plug-and-play across ad platforms because your solution (if you find one) will only work on one platform.

Phase 4: Why You Need To REPEAT This Process Regularly To Continue Scaling

Now you know the 3 phases of our process that has helped us continually find ways to help clients scale over the last decade.

But if you want to continue scaling month after month, you need to learn how to repeat this process to continue to identify opportunities (and threats) to scale.

And that's what we'll discuss in the next post.