The “butterfly effect” myth of linear performance improvement
On December 29, 1972, Professor Edward Lorenz, Sc.D. posed the question, “Does the flap of a butterfly’s wings in Brazil set off a tornado in Texas?” While presented to The American Institute for the Advancement of Science during their 139th meeting, his intended premise was almost universally misrepresented by the media and misunderstood by popular culture. The “butterfly effect,” as his concept’s bastardized version came to be known, took hold.
Professor Lorenz’s actual aim through his provocative question was to assert the opposite point. His idea served as the basis for modern chaos theory: In a complex, dynamic system, “small variances in initial conditions could have profound and widely divergent effects on the system’s outcomes,” The American Scientist, May-June 2017, Volume 105, Number 3, Page 130. In the digital marketing world, this means the bigger your budgets, the greater exponential differences small changes can make, in sometimes unexpected ways.
More importantly, this 1970s bit of chaos theory directly and indirectly illustrates common, crippling digital marketing misconceptions.
Scaling media isn’t linear
To dismantle any misconception, we need to start at the foundation. In this case: the belief that scaling media is a strictly linear, directly proportional relationship to the results you get (all the leads, money, notoriety, sense of fulfillment, love you never received from your emotionally constipated parent, etc.). Scaling media isn’t linear. Increasing budget isn’t inevitably followed by better ad campaign performance, especially if it isn’t paired with intentional action. To illustrate, let’s dive into the basics of digital ad buying.
Digital ad buying is essentially an online, automated auction. If Google, LinkedIn, etc., relied on manual ad buying, they couldn’t handle the literal tens of thousands of auctions performed per second (and there’d be a lot of errors and screaming).
Enter our collective overlord: the algorithm. In Google’s case, a user might search for “best fiber network near me.” Google’s algorithm reviews every auctioneer that has selected “best fiber network near me” as a target. It then determines the order in which those ads appear according to 1) the bid and 2) the perceived relevance/quality of the website competing. Remember that an ad platform is always simultaneously working to ensure relevance to protect their platform in the long term (i.e., if they serve up random, unrelated ads to poor-quality sites, users will stop using them).
Here’s the relevance to scaling media and the hitch in your giddy-up: Algorithms encourage you to do things to improve your “quality score” (i.e., your perceived relevance and possible ranking in the auction.) But not all of these recommendations are good for your brand. For example, weak CTAs or keyword stuffing might boost your quality score, but they can also dilute and cheapen your brand perception.
These potential negative effects become more pronounced at massive media buying scales.
Small optimizations have outsized impact
Lorenz’s red-headed stepchild — the popularized “butterfly effect” — “emphasized the outsize significance of minute occurrences.” While scientifically incomplete, this emphasis brings us to the next misconception we need to dismantle: Impact only lies in dramatic action. Not true.
If you’re spending $50M a year on e-commerce ads, for example, the tiniest tweak to an ad headline or a CTA could result in tens of thousands of dollars in lost revenue at peak times. In other words: The bigger you get, the smaller you have to think. Seemingly minute decisions, such as pinning headlines, determining which products are featured, billboard location, homing in on trade show calendars, and refreshing ad copy matter much more than general budget increases.
Spending $200,000 a day on Google Ads doesn’t mean we don’t need to care about ad headlines because “the traffic will still be there.” It means we should care more. We’re building the brand image in front of more eyes, and we have to play into the algorithm of Google (or any platform) to improve ad ranking. Improving ad ranking improves cost-per-clicks, which improves search engine results page position, which improves click-through rate (CTR), which improves conversion rate, and so on.
At scale, making small adjustments like unpinning headlines, increasing target cost per acquisition within a shared budget, etc., signal permission to Google to find existing demand that tighter controls often limit. Small adjustments give more freedom to Google, which in turn can reset campaigns into a learning phase. However, the learning phases can be costly at higher budgets, again putting the responsibility to the advertiser to be mindful of how and when optimizations are made.
Flexibility doesn’t mean flippancy
James Clear, the author of “Atomic Habits,” wasn’t thinking about digital marketing when he wrote, “The tighter we cling to an identity, the harder it becomes to grow beyond it.” But digital marketing strategy is predicated on human nature. And humans love psychological safety and predictability more than necessary change and growth.
Playing to the algorithm necessitates flexibility. Many marketers conflate that with “flippancy.” Not at all. Playing into the algorithm for better results also means doing more work on the back end. If Google is randomly selecting headlines, for example, you’ll need to refine them to make sure they’ll all pair well. If you’re going to cut the apron strings, your ads will need complete congruence: immaculate design, succinct sense, and extreme relevance, all sandwiched in solid brand image protection.
To illustrate, a B2B/B2C client was using a single pinned headline in their Google PPC campaigns. During a peak season, Google began designating these ads as “poor,” and CTR plummeted. Removing the pin and allowing Google to choose headlines restored ad ranking to “good,” and CTR exceeded normal levels, allowing ads to hit YoY traffic goals.
This “adaptability” extends to your opinions and assumptions. If you only stick with what you think you know, you’ll miss out on opportunities to grow. As an example, a retail shoe client was adamant about not including kid channels in their YouTube strategy. Normally, kid channels amount to a lot of wasted impressions and clicks/spend because they aren’t relevant to most businesses. In this case, after all, parents are buying shoes for their kids all the way up through college. While it initially seemed “safe” to exclude kid channels, the agency’s media team knew that parents would be the ones paying attention to the ads, so despite the mental hurdle, the client allowed us to place ads on some kids’ channels, with good results.
Flexibility means determining your comfort threshold between playing ball with the algorithm and considering lifetime brand value. Flexibility isn’t brand dilution, low-quality creative, or generic messaging. It means putting the work in beforehand so we can release control to an algorithm that can then work better toward your goals. Flexibility is key to better results.
Strategy isn’t static
Which brings us to: Strategy can’t be “one and stick-a-fork-in-it done.”
If you never refresh your ad copy and you’re showing the same ad to the same people, that won’t do much to build your brand or strengthen brand loyalty. Six months from now, customer acquisition will likely tank.
While you don’t want to risk anxiety-hopping from one approach to another without the chance to gain meaningful data, there are areas you should regularly be assessing and refining:
- Site UX (if visitors can’t figure out what to do on your site, they’ll bounce)
- Possible friction points (e.g., consider whether gating content will be helpful or hurtful to your goals)
- Testing (always-on testing strategies don’t mean wasting budget; they allow us to safely test new and additional strategies at scale)
- Your assumptions (especially if they lack nuance or are solely based in fear)
- Optimizing by channel (demographics, intent, use, and optimal content type and content specs vary significantly)
While we’re all at the mercy of the algorithm, it doesn’t follow that we’re powerless. It simply means we need to be constantly homing in on what’s working and what could work in the future.
Expert media strategy rests in intentional consistency
Any media buyer can spend more money, but doing so without intentionality is like a real-life, reenactment of “Everything Everywhere All at Once.” What’s more, if you just add money, you’ll eventually max out the users that you’re targeting. You can’t create demand with money; you’ll just be paying more for the same user.
Besides that, focusing on budget over hitting touchpoints along the user journey is a huge missed opportunity. That said, money is an “easier” subject than “user journey,” so the aversion is understandable. We’ve probably all seen those user journey illustrations — a giant convoluted path that raises our blood pressure at a glance. Which leads us right back to intentionality: Don’t try to hit everyone all the time with absolutely every bit of messaging. Is the user journey messy? Yes. Is the user overwhelmed? Yes. But your (our) job is to hit every touchpoint with consistency and meaningful connection.
One of Luxid’s media specialists is an avid sewer. In sewing terms, she thinks of hitting touchpoints as the same as carefully weaving her way along a piece of cloth, being mindful of where she makes each stitch.
An effective modern media strategy is:
- Based on expert judgment: This includes knowing which levers to pull and when and how to push performance to achieve your goals.
- Algorithm-driven: Google, LinkedIn, Meta, and others actively interpret and respond to the signals advertisers give them.
- Proactive: Take ownership for brand image while balancing flexibility to improve performance. Platforms don’t inherently protect brand tone, context, or long-term perception. That’s on us.
These tiny “flaps of a butterfly’s wing” are the difference between money spent and exponential impact. (We’re pretty sure you’re looking for the latter.)
