Productivity Growth

There are at least two systems of achieving productivity growth: path dependence and disruption. What if there is a third way?

This post unpacks that paragraph and explores ways through. It will start with explaining lock in and path dependence. We’ll cover the application narrow machine intelligence in a very narrow industry. It will end with a small scenario and a few what ifs.

Lock In

Consider banner advertising. This is a relatively old industry. Its roots predate the Internet by at least a couple hundred years. It may have started thousands of years ago.

It starts out with a person with a problem. They need to get the word out about their product or service. Reframed, they need to capture some group of peoples’ consciousness long enough to change it. At the root, there are two jobs here: aggregating consciousness and changing consciousness. Where there is value, specialization, and security, there are markets, specialists and progress.

There is an industry of people who specialize and sub-specialize in capturing and aggregating your attention for purposes of affecting it. These are producers, journalists, editors, writers, painters, programmers, product managers, videographers, editors, musicians, developers, photographers and on and on. This is the artistic side.

There is an industry of people who specialize and sub-specialize in changing peoples’ consciousness towards products and services. These are copywriters, creative directors, agency presidents, developers, marketing scientists, programmers, art managers, client managers, planners and on and on. This is the marketing side.

These industries can be intertwined. Some societies mix art with commerce quite easily. Others do not. These attitudes and conventions can become locked in.

The agreement between artists/journalists and publishers in some societies was that marketing could happen along side writing – separate, and over there. Over there in boxes that appeared alongside writing and photography. Over there on pages of written ads that were classified into categories. They needed to exist. And they were over there.

This pattern was carried over to the Internet when it became commercialized in 1993. There was the content that people wanted to surf, and then there were visual display ads that appeared over there, along the side or at the top as a banner. Some publishers enabled visual ads to pop-up and interrupt the experience. It took a coalition of technologists to re-establish the pattern of ads alongside, not ads in spite of, content.

The convention of ads in front and in between video and audio, as established thousands of years ago in theatre, continues.

This is one way of introducing the idea of Lock In.

The other way the introduce Lock In is to refer to the QWERTY Keyboard used in the Anglosphere – a deliberately inefficient keyboard layout intended to keep typewriters from jamming – is now completely locked in. The social costs of moving to a new design are extremely high and the perceived returns are too low.

Banner Advertising Lock In

Consider the relationship between marketers, publishers, and agencies. Marketers want to capture attention for the purpose of changing minds about what they have to offer. Publishers want to capture attention for the purpose of changing minds – and need a source of income to do that. And agencies can be thought of as coalitions of people with one foot in either world.

There is a dynamic that is locked in with respect to banner advertising. Marketers tell the agency about the product and the target market. The agency generate some creative on their own, and get some creative from the marketer and they resize it over and over again into different formats. And then the agency fights for margin with a technology giant to place it on a publishers website.

At any given moment, a marketer may have several stories they want different target audiences to believe. They may want pregnant women to buy a financial product for their soon to be born child, and they may want a first time home buyer to immediately thinking of retirement. One set of messaging in English. Another in French. Two versions. And at least 14 size formats. It is a lot of art that must be manually assembled. This is great for the agency, because they earn money both on the commission on the working dollars, and on the creative services they offer (the non-working dollars).

This sets up a conflict between agencies and marketers that you can read about every month in Adweek or you can read a solid version of it in Claude C. Hopkins’ My Life In Advertising. The short summary is that the marketer often tries to shed risk onto the agency, and the agency resents it. The marketer often believes that the agency could be more productive, and understand that they don’t have many incentives to cause productivity to increase.

Productivity Growth: Morphing Banner Advertising

In Morphing Banner Advertising (2013) Urban, Liberali, MacDonald, Bordley, and Hauser explore a novel way to increase the advertising effectiveness of display advertising. They increased the effectiveness of display ads through some deliberate targeting and matching ad creative to cognitive state. This increased the amount of ad creative required.

“However, there is a fixed cost to the development of multiple banners for use in morphing. The targeted banners for the automotive experiment would have a cost of $250,000 to produce if done at market rates (based on a quote from an advertising agency to GM to do this work).” (p. 16)

Glen L. Urban, Guilherme (Gui) Liberali, Erin MacDonald, Robert Bordley, John R. Hauser (2013) Morphing Banner Advertising. Marketing Science, pp. 1-20.

This is a situation where increased diversity in creative can increase returns, and there is a hard cost in creating it.

Path Dependence and Disruption

There is a path dependent, standard response, to reducing ad creative cost (increasing productivity) over the past 25 years. First, you hire them increasingly younger, pay them less, and drive overtime. Then you nearshore the work. Then you offshore. Different agency presidents are in different stages of that process.

There is a disruptive path to reducing ad creative cost. Google’s response to this problem was to make the cost go to near zero with text ads. By doing so, they captured all the Small and Medium businesses that had relied on the classifieds. It was only later that they got into ad delivery. They’ve avoided creative entirely.

A fairly obvious path to solving the fixed cost problem of ad creative would be disruptive. You’d train a machine to generate the creative, with an eye towards personalization at scale. You could take the orthodox path to disruption, with a large R&D investment upfront, and with a focus on displacing the middle part of the market. There’d be a significant problem around managing creative-machine tension. That’s the ultimate answer as to why Google just doesn’t do it themselves – it’s because that isn’t quite how they systematize knowledge.

There are very good reasons for lock-in and path dependence. There are very good reasons for disruption. Each system have very well defined, very well optimized, checklists, theories, beliefs, assumptions, and cultures. One check list (path dependence) is easier to run than the other (disruption). Both sets are linked back to deep human memory and cultural tradition. 90% of agency presidents could be relied upon to act in a predictable, path dependent, way. 90% of startup founders could be relied upon to act in a predictable, disruptive, way. And maybe, for that reason alone, there is no reason to wonder about a third way.

A Third Way?

I still wonder about a third way.

If there is a third way, it would have to be because of people.

I agree with Kolb that there are at least four different ways that people tend to learn. I’ve noticed that some people seem to be relentless doers. Some people see systems and patterns in everything and make connections. Some people are endlessly divergent thinkers – they’ll go on discovering new things and will never synthesize or try to converge on anything. And some people are relentless synthesizers. Put these four people in a room to solve a problem and watch the fireworks set the grease in the dumpster on fire.

Path dependent cultures seem to recruit and retain tactical doers. It’s very efficient to have a large set of executors…execute. They don’t seem to get into much conflict with each other, and tend to stay out of each others way. I’ve noticed this about established industries in hardcore optimization mode – airlines, mining, hospitality, rail, and restaurants. They tend to be very homogenous each in their own way. And I think that’s beautiful.

Disruptive cultures seem to recruit and retain all four types and put them on the same team. They got great one or two great divergent thinkers, a few great convergent thinkers that bring a focus, amazing systems thinkers, and excellent doers. Teams that contain each of those four types, who can trust each other, tolerate each other, and learn together, can achieve some pretty great things. It’s very hard to do. These four types of thinking styles drive each other crazy.

I wonder if there’s a way of organizing teams and their experiences in such a way that explore and exploit cycles are held sustainably in balance. Profitable organizations tend to fall into an equilibrium of general exploitation. I don’t think that this is purely a function of short term shareholder value delivery – because I see this in environments where shareholders are absent. It’s as though…when there is more there is to lose, there is less incentive and more risk against learning. I wonder how often the words “You don’t get paid to think” start to appear in the shared culture over time.

At the same time, if there are too many exploration cycles, too many people may become burnt out, eroding the social capital of the firm and slowing down the rate at which a firm learns. There may be a design pattern where a firm can moderate how much learning a team is doing by managing participant flow. There may be a pattern of shaping problem flow.

I wonder if there’s a third way that is pareto optimal, between the predictability of path dependence and the chaos of disruption — or if it’s all just all traps.

It’s very hard to image an agency president, or an ex-Googler founder, behaving in a way other than what they know. They’re locked in. And that’s very neat about asking about a Third Way. It would be a kind of leadership that I don’t have a name for. And just because I don’t of it…it doesn’t mean it doesn’t exist.

I wonder what, and if, it is?