Just Do It! Experiment for Incrementality
We all know a head-in-the-clouds bookworm who obsesses over accumulating all the comprehensive information pertaining to a matter but never takes action; and we all know a boots-on-the-ground doer.
In your experience, from these two sorts of people, who do you see being more successful?
I am fortunate enough to work with the founders and owners of many great brands. It’s inspiring to work with such exceptional people, and they are almost always doers. This is great, but it also has a downside. They are exceptionally busy. Hiring, sourcing, researching, networking, consulting - being a doer is a 24/7 job.
The reason these doers are so attracted to the type of advertising I do is that it directly impacts the P&L; What you don’t spend is profit. That’s why I’m often asked:
“How can we spend less on advertising and make more money?”
It’s the Holy Grail of business that we are all searching for.
To lay our hands upon this miraculous relic, we need to figure out what advertising (e.g. channel, audience, message, creative) is creating profit that wouldn’t have happened regardless, AKA incremental revenue.
For example, Google’s Pmax and Metas Advantage+ have been getting notoriously good at finding people who were ready to buy anyway - but not at driving incremental revenue.
So how can we measure incremental revenue? In the golden era of performance marketing (over 10 years ago now), tracking users end to end was simpler. Determining a campaign's performance was simpler. I’m not saying this was better than the practices we have now, but it worked for the biggest brands in the world for a long time. You could see if something returned a profit and, if you spent more on it, you made more money. Simple.
As that got harder (due to iOS 14.5, platforms using different models, etc..) bringing various channels together in the same place (either through a more advanced attribution model, CDP, marketing mix model, etc) gained popularity. But this complexity requires more time (and money). It seems like too much time sitting and thinking for these doers.
So what might be another solution?
One thing I don't hear many people talking about is experimentation. The truest way to find out if something really worked is by doing something to test it. The anglicised french word “entrepeneur” means “to undertake”. And that’s what we often need to do. Emphasis on the word do.
Imagine you want more sales with a new seasonal campaign but - as it’s not a sales period - your price isn't a lever. What could you do? I’d suggest you run an experiment:
At the beginning of the month, set a sales objective for next month and define the dimensions that will help you get there (e.g. Creative, Audience targeting, Advertising Channels)
Run the campaign.
At the end of the next month, check if you achieved your targets and review each Creative, Audience and Channel to see why you did or didn’t. Often this will give you some clear insights, ready to test next month.
This might not be the most intellectually acute, but it’s definitely the most practical. It’s the doers way - achieving actionable insights with proper speed of execution.
Don’t get me wrong. I love a good Marketing Mix Model and the work that has been done in that field, especially since its recent revival, has been truly impressive. However not everyone can afford the time or money to spend on more and more complex analyses. You can get stuck trying to find the exact truth. Whereas experimentation satisfies my clients’ needs for both insight and action. It’s the doer's way.
