Rand Fishkin posted an essay a few months ago on how something is wrong with the advertising industry. Okay, I get what you’re thinking. This is not the first time someone has questioned does advertising work?
The essay also directed me to an hour-long (2-part series) podcast from Freakonomics that attempts to answer the same topic. However, this time with more data points than anybody has ever assembled.
The longer I listened to the podcast, the more the question of whether advertising really works stuck in my craw (somewhere in the stickiest part of my limbic system).
After all, we’ve all heard the famous John Wanamaker statement, “Half the money I spend on advertising is wasted; the trouble is I don’t know which half.”
However, before we attempt to address the issue, let us first seek to understand how does advertising work and how we presently evaluate the efficacy of advertising.
How Does Advertising Work?
Advertising works by triggering an emotional response in consumers, influencing their buying decisions, and thereby building a long-term connection with the brand.
The key aspect to note here is that various forms of advertising pay off over varying time periods. According to a study conducted by Les Binet and Peter Field at the IPA, although the influence of brand advertising on sales may be less immediate than that of a direct-response campaign.
But if your advertising effort persists over time, it can be more effective in the long run.
For advertising to truly work organizations have to balance between long term and short term objectives.
How to evaluate advertising effectiveness?
Various methods ranging from basic regression analysis to market mix modeling have been used over the years to evaluate and analyze advertising effectiveness. Let’s talk about the two most popular approaches why they don’t work anymore.
1. Market Mix Modelling
Market Mix Modelling is one of the earliest statistical methods to evaluate the efficacy of advertising. The model breaks down business metrics to differentiate between contributions from base sales and incremental sales.
Base sales are generated without the use of advertising. It is the sales generated as a consequence of the brand equity established over time. Incremental sales are produced as a consequence of marketing activities such as television and print advertising, as well as digital spending.
Marketers can no longer rely on Market Mix Modeling as it doesn’t measure “synergies” across media channels, in which the effect of two or more simultaneous media activations exceeds the sum of their independent parts.
Market Mix Modeling was created in the 1980s by Procter & Gamble and has seen a few avatars in subsequent years. When Procter & Gamble shifted its Marketing Mix Modelling with Nielsen and DemandTec in 2011, the work was pegged at $15-$20M.
Large corporations invest millions of dollars in determining the efficacy of their billion-dollar advertising budget.
2. Randomized Controlled Trial
The other alternative to market mix modeling is the Randomized Controlled Trial (RCT). When you think of RCT, the first thing that comes to mind is a clinical trial in which one group receives drug treatment while the other group receives a placebo.
Randomized Controlled Trial has become a gold standard when it comes to measuring advertising’s lift. In advertising, lift measures a user’s likelihood of purchasing a product or service as a direct result of a promotion.
However, most marketers do not conduct RCTs since they are very time intensive to execute. These studies need a team of engineers and data scientists working extra hours without any additional pay.
A study conducted by Kellog School of Management found that in the absence of an RCT, it is difficult to determine an ad’s lift with any degree of accuracy. There are powerful factors at work against observational methods, making any efforts to establish causality inaccurate.
They discovered that observational methods raised the lift to 416 %, while the RCT suggested a lift of about 77%.
What Should Advertisers Do?
If you’ve listened to the Freakonomics podcast, you’ve probably heard about many studies that indicate that advertising doesn’t work.
One of the most discussed studies on the podcast was one performed for eBay by Steven Tadelis. The study highlights that brand-keyword ads have no short-term benefits and that returns from all other keywords are a fraction of conventional estimates.
While the podcast acknowledges that Google responded to the study and provided other studies demonstrating the effectiveness of advertising, it omits the fact that eBay lost almost 80% of its organic ranking a year after the report was published.
Panda 4.0, a Google algorithm change update released in 2014, impacted almost 12% of search queries. The most significant lesson from the whole scenario was how retailers which followed eBay’s lead and reduced their AdWords advertising spend were forced to reverse their decision overnight.
If we believe that advertisers are at the mercy of large platforms that own almost all of the web’s reach and inventory, what options do advertisers have?
1. Diverse & Improve Media Mix
Advertisers must find alternative routes to conventional media platforms such as Google, Facebook, and others. This may involve retargeting/re-marketing, display, and programmatic advertising via third-party platforms.
2. Leverage Emerging Channels
While you continue to invest with several major platforms, also experimenting with new channels such as direct, 1:1, non-programmatic website ad buys. Discover and create content syndication opportunities with podcasts, email newsletters, and video/stream sponsorships.
3. Conduct Efficacy Studies & Audits
Audits are not just for identifying advertising fraud, brand safety issues, poor quality scores, and improper ad placements; they are also for ensuring that you are receiving the most value for your advertising dollar.
Efficacy studies, particularly a Randomized Controlled Trial, will enable you to assess the impact of your advertising efforts. If you’re doing a controlled study, consider including 10-12 percent of your audience in the control group (without exposure to advertisements); if you’re a large marketer, this number might be as low as 3%.
Advertisers can now utilize RCTs to not only discover incremental sales but also to optimize ads across multiple channels, especially due to emerging technologies.
So does advertising work? As marketers, we are convinced that advertising works and have been conditioned to believe that our beliefs are true. However, as Scott Fitzgerald puts it, ” The test of a first-rate intelligence is the ability to hold two opposing ideas in mind at the same time and still retain the ability to function. ”
At times, we must hold two opposing concepts in mind in order to cast doubt on some of our beliefs and make the best choice possible based on facts and intuition.