Optimize later. Experiment first.

By Chris Harges, director of strategy

beef 3

Digital marketers are masters of the small win. And that’s a big deal. Over time, small wins earned through optimization add up. The resulting impressions and transactions help justify the spend and contribute to growth. But there’s a hidden cost in focusing on marginal improvements. You can miss the forest for the trees—and trees don’t scale.

For most paid marketing channels with clear conversion metrics, the efficiency of the channel (as measured in direct ROAS) will degrade over time. If you’re buying retargeted display for example, your early investments will yield a better return than later ones. The reason is simple. The pond starts full of fish. Cast a line in and you’re sure to catch something. But as you harvest early gains, you depopulate the pond and you will land fewer fish with each cast.

“There are only so many fish interested in your bait. And the hungry ones bite first.”

The same is true for acquisition marketing tactics. Though the curve flattens more slowly, it’s a function of the same phenomenon. Tuning and optimizing will help eke out the best performance. But there are only so many fish interested in your bait. And the hungry ones bite first.

There’s an obvious fix—add more fish. For most marketers, this means easing open the aperture on your targeting. But that should be a last resort. Unless you’re marketing a product with overwhelming market share, you’re only reaching a fraction of the audience you know is the right fit. You shouldn’t chase less likely prospects until you’ve maximized share with the likely ones. So where do you turn when optimization fails? Two options to play out the fishing metaphor:

Fish Somewhere Else

Google’s Managing Director for Ads Marketing Matt Lawson puts it this way “… it will be strategic testing—not incremental adjustments—that will take measurement strategies forward by leaps and bounds rather than mere steps.” His experimentation takes the form of media mix shifts. Lawson uses the example of Red Lobster marketers experimenting with Google’s store visits to measure AdWords performance as a foot-traffic driver. Cool but not really new as a strategy. Marketers have been playing with new digital communication tools for decades.

“What’s important to remember in the digital age is that optimizing is not experimenting.”

Change the Bait

Experimenting with creative is not new either. What’s important to remember in the digital age is that optimizing is not experimenting. Here’s the main point of this post: A/B testing similar creative helps optimize the return from the same subset of your audience—the ones who respond to that kind of creative. To get at those who aren’t responding, you need to experiment with radical variants. Here’s an example:

Meat Was Murder

We ran a campaign at an outdoor apparel company to build brand awareness and drive traffic to its e-commerce business. As part of an omni-channel media strategy, we created a landing page with high-value content and an assortment of associated merchandise. Then we did customer-look-alike modeling for display and paid social with creative that teased the content. The creative and the content featured a mash up of cooking and outdoor sports (we knew both appealed to our demo) with the CTA: “See our recipe for the perfect winter’s day.”

“Instead of images of outdoor athletes, we ran a close-up of ground beef.”

Our first flight of creative had lackluster performance. We were A/B testing similar variants and not learning much, so we tried a radical variant. Instead of images of outdoor athletes, we ran a close-up of ground beef. It fit in with the recipe concept but was completely out-of-category in its sensibility. No athletes. No snowy mountains. Just a pile of raw meat. CTAs doubled over the “safer” creative and pulled at about 5x our category benchmark. Conversions as a percent of traffic were the same as the “safe” ad and our audience demographics were identical. We were getting more of the fish we wanted by changing the bait.

Radical variants don’t have to be shocking. But they do need to take a very different tack from your initial creative. One way we allow for this at Tank is to go wide on our creative concepting. That gives us two radically different finalists—both right for the campaign, but far enough apart to allow for experimentation, not just optimization.

Chris brings 20 years of experience as a brand marketer and strategist, helping build awareness and drive sales through story-driven, omni-channel, go-to-market strategies.