In this issue: Whatever happened to test and learn? Why we shouldn’t let algorithms tell us how we test. And some lessons from the founder of scientific advertising.

Back in the heyday of Direct Mail, were Test and Control cells negotiable? No, never.

Each DM piece cost at least $1. That's a $1,000 CPM. So of course we had control and test cells, because our jobs were on the line. If you got it wrong, that was a hell of a lot of money down the drain. Even the postage cost a lot of money, and somebody would have to pay for the mistake.

Digital? $10 CPM! Suddenly…hi we’ve gotten lazy.

Did we collectively decide that just because digital impressions are cheaper, that those ‘direct response’ principles, honed over 100+ years in ‘direct mail response’, don’t apply to ‘digital response’? That the risk of being inefficient in optimizing large ad investments, were somehow smaller?

Digital marketers gripe about CPCs and acquisition costs, yet skip the fundamentals: rigorous testing, matrices, hypotheses, or confidence intervals.

We’ve tossed aside the ‘boring’ “direct response” playbooks that ‘actually’ drive returns. Why? Because digital impressions are 1000x ‘cheaper’ than direct mail.

Spoiler: they are not.

A flood of impressions at a fraction of the cost doesn’t equal easy RoAS.

Minimizing the cost of acquisition in digital is harder than ever!

If you don’t know where you’re going, your journey won’t be cheaper just because gas prices fell. Because you’ll still be going around and around in circles.

But a lot of digital marketers have stopped testing. And that means they don’t know where their campaigns are going. They’re driving around in circles. But they’re not worried, because the price of gas is cheaper than it used to be.

But they’re not learning anything. And they’re not going anywhere.

Which is why so many of them barely break even on RoAS.

Want to squeeze more juice out of your digital ad spend?

Start testing. Test aggressively. Same rules, new platform.

Yeah, it’s hard.

Designing test matrices? Time-consuming.

Crafting hypotheses? Takes creativity.

Tests that don’t deliver lift? Takes courage to admit, and insight to reflect on the lessons.

Discipline to stick to 1 control and 3 test cells with just one variable monitored per each such test? That’s where the magic happens, but yes, it’s a grind!

It's tempting to hit the ‘easy button’ and let Google’s PMAX or another algorithm take the wheel.

Yes, automation is essential – we’ve been living in the age of machine learning since 1952! Google introduced machine learning into Adwords in 2013.

But performance marketers who care about RoAS? They don’t stop and let the algorithm drive performance on auto-pilot .

The ROI lies in the hard stuff – good old-school test matrices.

Over time, small, consistent tests boost RoAS by 5-10% every month, doubling or trebling returns in a year. That’s what you call adding Value.

When done with excellence, RoAS increases, and digital ad budgets scale up at the same time. So your brand makes more money and your marketing department gets more budget to manage.

Testing works. It’s just not for the lazy. It takes time, tenacity, and grit. The upside? Creating more value.

Here is how I think about how we all can align our actions around our Family ‘life’ in 2025, with ‘the more we put in, the more we get out’, as the North Star.

(btw, this framework follows my Return on Marketing Career post)

Step 1, Get out of your Comfort Zone: I know that managing digital advertising is just one of your core accountabilities – it’s not your only focus. Yet digital ad dollars are probably a large expense in your budget, and you likely get many questions from the CFO about how those budgets are managed. (I certainly did when I was leading marketing at Bath & Body Works.) With that in mind, what if you had a hypothesis, that doubling returns from digital ads was … possible?

Step 2, Develop a New Skill: Create what appears to be an ‘unreasonable’ goal. Ask your team, and your agency, what would need to happen, to double RoAS from digital ads. What would need to happen for the ‘unreasonable’ goal to become real? How long would it take? What steps would have to happen?

Step 3, Create More New Value in your Role: When you shoot for the stars, you get to the moon. Clearly, doubling RoAS adds Value. But how do you get there?

The answer is simple, even though the process will be hard. You do it by delving deeper, truly understanding the process, the small consistent steps, that are the foundation of having ‘test and control’ cells to squeeze max Value from online ads. You don’t try to solve the entire problem in one go. You succeed by breaking it down into bite-size pieces. As the old saying goes, you can only eat the (strictly metaphorical) elephant in chunks. (Not that I would ever harm an elephant, of course.)

Step 4, Set Yourself Up for a Bigger Future Role or a Promotion: Leaders who move up faster know to leverage the latest tech, including AI that’s built into media buying platforms to optimize returns. But they also always keep the pressure on their team, to use human intelligence, to design high-impact tests, that AI just can’t think up on its own.

In 1923, Claude C. Hopkins wrote Scientific Advertising. Although over a century old, its principles still apply today, especially the emphasis on rigorous testing and measuring the impact of marketing efforts.

Here are the top three principles from the book:

Consistent, disciplined testing is key to optimizing returns and scaling budgets effectively in marketing.

If you’d like to discuss your career journey with me one-to-one, please feel free to email me at [email protected] or message me on LinkedIn.

Thank you for reading.

If you know someone who you think would appreciate this newsletter, please forward it to them.