Marketing-mix models are of limited use at a time of extreme disruption to old consumer habits. Joel Rubinson outlines an alternative strategy that brands may be able to leverage instead.
Marketing in the COVID-19 crisis
This article is part of a special WARC Snapshot focused on enabling brand marketers to re-strategise amid the unprecedented disruption caused by the novel coronavirus outbreak.
We are in extreme times for nearly every marketer: COVID-19 has created panic buying, shutdowns of retail, a coming disruption of food supply chains, a 25% correction in the stock market, and a 20% unemployment rate.
And ad spending is down by double digits.
As I talk with marketers, asking them about how they plan to restart their businesses as the US comes out of this, some of the most persistent big questions I hear are:
- When should I turn my advertising back on?
- Should I focus on brand or performance advertising?
- How can targeting play a role in reducing ad waste at a time when every ad dollar is being scrutinized?
It certainly isn’t business as usual for marketers. And the rules have changed for marketing-analytics teams as well, who now need to:
- Reduce reliance on marketing-mix models
- Increase reliance on user-level attribution modeling and experiments
- Prove and maximize the value of targeting
Reduce reliance on marketing-mix models
Large marketers find themselves in a tough spot because no one can, or should, trust any updates to marketing-mix models for at least a year.
The methodology taps into historical relationships that rely on a “past-is-prologue” principle to drive future actions. But, with an entire global economy on at least partial shutdown, the past has become suddenly irrelevant.
Marketing-mix modeling is broken, at least for this year, and probably into 2021.
Increase reliance on user-level attribution modeling and experiments
In fact, marketers need more help when the past is not prologue.
For the next 12 months – especially as advertising begins to open up – marketers should look to multi-touch attribution (MTA) and controlled experiments for guidance.
Isn’t MTA also susceptible to disrupted advertising-to-sales relationships, you might ask. And I answer, “No,” for a simple reason: MTA and controlled experiments use forward-looking data not backward-focused history. MTA analyzes conversions versus ad serving as data are unfolding. It focuses on going forward during campaigns and flights of advertising.
Controlled experiments operate in much the same fashion: They are designed and executed in the future, not the past. Because they are fresh, the results can help resolve the vexing questions that marketers keep asking:
- “Should I start advertising again?” – this is highly testable.
- “What mix of brand and performance marketing works best to balance brand needs and the need to show tangible results?” – again, highly testable.
Proving and maximizing the value of targeting
“Reducing wasted ad impressions” and “addressing advertising to consumer targets producing 2x [or more] return” are two sides of the same coin.
One is the sickness; the other is the cure.
In both cases, marketing inevitably faces financial pressure to prove that advertising is not just an expenditure, but also an asset that adds to the bottom line.
And, again, in both cases, the way to make advertising campaigns deliver each and every time is by smart targeting.
At least 50% of your wasted ad impressions can be reduced by smart targeting with the elimination of:
- ad impressions served to those with no interest in your offering (no interest);
- ad impressions served to the wrong demos (e.g. males for feminine-protection products);
- licensed segments for programmatic targeting that have low validity.
To explain the last factor, let’s say you are a quick-service pizza chain. If a data store provides you access to a segment of pizza-chain shoppers, what percent of that cohort potentially will be interested in your offering?
The answer, of course, depends on the validity of the segment. Before you can determine potential audience interest, you need to know what percent of that segment truly consists of people who would buy from a pizza chain. You also need to know how interested they are in your particular brand.
Choosing the right segments can make your advertising self-funding and build your brand at the same time when it gives you access beyond your current customers.
To make a smarter choice, you shop around “analytically”: If there are five segments you could license from different aggregators for programmatic targeting, my experience suggests that at least one segment would be no better than random. And, in fact, the best segment predictably could deliver twice the return on your advertising expenditure.
Choose the wrong segment, however, and you are back to ad waste.
Coming out of COVID-19 lockdowns, as marketing reboots, so must marketing analytics move in fresh directions.
Don’t look back. Look ahead. Don’t just do what you were doing before COVID-19. Change your analytical practices to provide the right guidance for a whole new set of marketing challenges.