The Value of Traditional Advertising in a Digital World

So, this is something I have been thinking about for a while now…and I have talked to numerous growth folks and entrepreneurs about (you know who you are) but I couldn’t quite synthesize something until recently.

In reading about randomness, probability, and chaos theory (something I like to do in my spare time), a couple principles found in the humanities made me rethink how I look at growth for startups.

The first principle is relatively simple, it’s the law of cumulative advantage. That is, what wins today is likely to win tomorrow. Specifically, success compounds out over time, so the winners diverge from the losers on an increasingly positive trajectory. While that may sound abstract, we’ve all heard expressions that reflect the consequences of this principle in everyday life, such as “the rich get richer while the poor get poorer.” In any race, getting a head start is incredibly important, and it turns out the rat race is no different. Now, when it comes to winning market share in competitive industries with low barriers to entry (something we see with marketplaces for example), cumulative advantage works no differently.

Now you maybe thinking that’s great Dan, what the hell does this have to do with traditional advertising?

Well…I am getting there, but first, let’s talk about that second principle.

Zipf’s law is something I recently discovered while reading about the faults of Gaussian math, and the overconfidence in normal distributions to determine probability.

The Achilles heel of a normal distribution is it’s under estimation in the probability of outlier events. For the fellow nerds following this blog, the tails are too thin and the curve decreases toward zero at too great a velocity. One opposing way to look at the world is through power-law distributions, in which there is a greater probability of outliers, and their value is exponentially divergent from the majority of the distribution. To make this concept more tangible, think of the distribution for web traffic amongst websites, income per citizen, or population per city of any given country.  When viewing these distributions, it will be readily apparent that a small percentage of actors dominate the distribution and can easily skew the mean or simple averages.

Now getting back to Zipf’s law, this principle states that when it comes to the use of language, vocabulary is extremely power law distributed. That is, a small % of the words in the English language account for the majority of words used in verbal and written communication (you can learn more from studies of Zipf’s law, herehere and here).

To visualize this, I am excerpting a graph below of the frequency of words used in Wikipedia from the linked study above. You can see the frequency of words follows a power law distribution where most the frequently used words dominate the distribution.


Ok, now it’s time to connect the dots. Given what we know about the distribution of language, we can assume that any given category or topic is distributed the same way. Such that, if we look at the consumer consideration set for companies in ride-sharing or search engines or even cereal, we’ll see extreme power laws. The #1 word that ranks in any of these categories obtains exponentially greater usage than #2, #3 or #4 (the curve is not linear). Now what is rank a function of? As all good marketers may guess at this point, the answer is awareness.

And awareness is a function of advertising and execution of any go-to-market strategy. This is where the value of traditional advertising comes into play. This is where the value of all those mass-reach unattributed impressions is concentrated, into top of mind awareness…most marketers I know would say of course Dan, traditional advertisers have known this for years! That maybe true, but no one I know has phrased it this way, that the consideration set for any given category of consumer products (especially commodities with little differentiation) is power law distributed…therefore ranking high up in the distribution really, really, fucking matters.

And traditional advertising via mass reach mediums (TV, Radio, Out-of-Home print) is a way to go-to-market with incredible speed.

The take-away from all this is that marketers in winner take-all markets need to be more aggressive with their advertising budgets and scale reach beyond digital channels. Establishing market leadership and a “dominant rank” in top-of-mind awareness is critical for success. Direct attribution may be a problem in the near-term future, but the executive team need to be aligned with giving credit to marketing spend for substantial increases in organic demand. It’s relatively simple to determine a baseline rate of growth, and then credit traditional ad spend with any increases in demand that goes beyond that baseline. While this may not sound like an optimal solution, when it comes to the reality of winning in a hyper-competitive market, you simply can’t wait for one.