Monthly Archives: October 2015

Our Economic Future: Driving Innovation Through Better Public Policy

The following are the notes I pulled together for the first panel in ADC‘s Future Summit on Monday September 28th.

The major opportunity for Australia is to find and exploit new production systems and consumption models that are cheaper, simpler and more “digital” than the highly entailed product-creating systems that are the legacy of the previous industrial era. We also need to see this as socially driven change, rather than a technologically driven change.

Two quick examples of this in action.

First: cars.

There’s a lot of talk at the moment around self driving and electric cars. Tesla has built an expensive but unprofitable electric car on the back of over USD 4 billion of government grants, while Mercedes, Google et al are out there with prototypes for self-driving cars that look like a technoutopian’s fevered dream.

In the case of Tesla, on the production side, the firm is better thought as the ultimate expression of an industry structure established roughly 100 years ago by Henry Ford; but it might not be an exemplar of how we will build cars in the future. A better example of where car manufacturing might go is iStream by Gordon Murray Design in the UK. iStream is a new production process, one based on established and well understood technology, but which removes 80% of the cost from the factory, slashing the cost of cars in the process. The production process Ford, Toyota et al are using needs 150k cars from a single model to be profitable, which means that Austrlia was lucky to have an old skool car industry for as long as we did. iStream is profitable on 12,000 cars, and would be commercial viable in Australia.

On the consumption side, viewing self-driving cars simply as autonomous versions of manually operated cars ignores changes in consumption patterns where consumers are preferring to consume many products as (value-added) services (think Spotify et al). The car equivalent is Flexicar or GoGet (car-by-the-hour).

If we put the two of these together it’s possible to imagine a new public transport model based on cheap and flexible, locally built and supported, autonomous cars. Some of the cars might be contributed by the government. Some by private operators (Flexicar et al). Some might be from individuals who are contributing their cars to the common pool when they don’t need them (during the week when they work, or when on holidays) via something like Uber.

Second, a local example: the transformation of the  building industry.

Building mid-rise buildings—office blocks, hotels, apartment buildings, &c.—is currently a craft-based process. Design a building, create holding company, buy land, put together consortium, get funding from bank(s), and then go onsite and incrementally add value to the land by hammering in nails, pouring cement, running wires etc. There’s a lot of talk about new technologies “disrupting” building, such as 3D printing. This is unlikely. Buildings are complex structures with many interwoven parts. You might be able to 3D print a wall, but you still need to integrate the services, render it &c. While these new technology might make elements of the process more efficient, they’re incremental improvement at best.

Enter Unitised Building (UB), based in Melbourne. UB have created a new production process that enables them to build a mid-rise building in a fraction of the time at less than half the cost. A good example is 3:East, built in 11 days. UB takes a complete 3D model of the building—including services &c.—and uses digital tools to split the building into a number of units (the model has been “unitised”). A second layer of digital tools takes that unit models and splits out the files required by CNC machines. The units are built in a factory and then transported to the site where they are lifted into place (one every 8 minutes) and snapped together. The only requirement is that you need a crane on-site, which, practically, means that the UB approach is dramatically faster and cheaper once you hit 3 floors (and need a crane regardless).

What UB have done is create a new process that moves the complexity of building from the physical world to the digital world. Indeed, their CNC requirements are quite light and they need few machines, so their factory (in Brooklyn, in Melbourne) has a very small footprint by manufacturing standards. They’re even exporting by finding contract manufacturing facilities overseas and transmitting the digital files to the CNC machines in the remote factory.

Creating these sorts of system changes has a couple of problems.

First, the old industry / sector structures we use to frame regulation and government support make no sense in this new world, as these new solutions span industry sector boundaries and have different requirements. (Supporting manufacturing, for example, has traditionally been a question of ensuring that the manufacturers have lots of land, but the new generation coming through don’t need much land, while they do need access to lots of network bandwidth.) This miss-match between the demands of the new and how government frames public policy makes it difficult for the two to engage.

In the case of UB, two of their challenges have been getting the banks to fund buildings when the current building risk model (based on incremental value creation on-site and quantity surveying) doesn’t match their building process, and the challenge of accessing government support when they don’t fit in any particular sector/industry (Are they a builder? Or a manufacturer?). These new firms span sectors / industries, deliver products as services, and do a bunch of other things that don’t fit with the old industry models. If we’re to frame policy and regulation for the future then we need to set aside the old industry/sector-based view of the world. Fundamentally, we need to stop muddling through as incrementalism won’t fix this problem. There are signs of change though, such as UB winning this year’s “Victorian Large Manufacturer of the Year” award.

Second, we need to acknowledge the these innovations are not the result of light-bulb moments or heroic individuals—they’re the product of trial-and-error and collaboration. By definition, they’re a social process. There’s a tendency—particularly among the technology crowd—to frame the debate in terms of technological determinism. Or, put another way, futurism has a technological blindspot. Just because we invented nuclear reactors doesn’t mean that we’ll have one in every home, or every car. No technology has ever survive contact with society intact.

We need to acknowledge that while the shape of society will change in response to technology (just look at what the modern smart phone is doing to our sense of identity!), society will, in turn, shape the technology it adopts (note that many people now find phone calls rude as they interrupt the recipient, whereas messaging is async).

The current obsession with disruption is a case in point. (And first we must acknowledge that Clayton Christensen’s “disruption theory” is looking less like a theory and more like an interesting idea.) There’s cries that we should let these disruptors usher in the brave new world by allowing them to skirt existing regulation. This assumes that all regulation is bad, or the more nuanced version, that techniques such crowd sourced recommendations are superior to regulation in many instances (why have certification when you can have ratings?) This point of view ignores the fact that regulations are one of the tools we use to encode what we see as the socially acceptable uses of technology. Nuclear power is a really cool technology, but do we want people driving around with small nuclear reactors under their bonnets?

With regard to Uber, and the taxi industry, it’s worthwhile considering the following:

    • allowing taxi licenses to be transferable and limiting their number was probably a mistake, however
    • we provide taxi vouchers to pensioners, partly to to encourage them not to drive, and partly to help them stay mobile and engaged with society: should we compel (i.e. regulate) Uber et al to accept taxi vouchers, or will we allow the death of the taxi industry to disenfranchise these pensioners?
  • Uber separates the payment from the provision of the service, and some parents are using this as an opportunity to give their under 18 (even under 13) kids Uber accounts so that they can get themselves home from school &c. rather than need mum or dad to pick them up: does this mean than we need to compel (i.e. regulate) all Uber drivers to have Working with Children checks?

It’s best to think about three types of policy:

    • Enablers, what do we need to put in place to enable the society we want. One of the biggest boosts to start-ups in Silicon Valley, for example, was Obama Care, as it means that individuals in startups could now access affordable health care. We undervalue policies such as Medicare and HECS as tools to enable as many people in society as possible to engage in the trial-and-error innovation process.
    • Drivers, how can we encourage new developments / ideas that create new value, given that government has a poor record of picking winners? This comes down to how do we use policy support the demand-side to help society to pull in the technology it wants in the way it wants. Admitting that we will regulate driver services, and we will require these services to accept taxi vouchers, and their drivers to have working with children checks, are good examples, as is the policy in Tasmania to provide interest free loans to individuals who want to by bespoke products from makers. Germany’s high feed-in tariffs for solar are another example.
  • Barriers, where do we draw the line? Do we want nuclear reactors in cars? Do we want full-timeAustralian for-hire drivers earning under the minimum wage?

There is a lot of opportunity out there for everyone and Australia, as one of the most voracious adopters or technology in the world, is in a position to capitalise on these opportunities. However, we need to accept that we’re seeing with “digital disruption” is the leading edge of a massive social change, rather than a technological change. The future will not be determined by the disruptors. It will determined by how we, as a society, choose to engage with this change.

Image: Steve Gibson

Be careful what you measure

Tyler Cowen has an article over at MIT Technology Review, Measured and Unequal, that discusses how improved measurement of workers might be a fundamental driver of inequality in the workplace of the future.

Consider journalism. In the “good old days,” no one knew how many people were reading an article like this one, or an individual columnist. Today a digital media company knows exactly how many people are reading which articles for how long, and also whether they click through to other links. The exactness and the transparency offered by information technology allow us to measure value fairly precisely.

The result is that many journalists turn out to be not so valuable at all. Their wages fall or they lose their jobs, while the superstar journalists attract more Web traffic and become their own global brands. Some even start their own media companies, as did Nate Silver at FiveThirtyEight and Ezra Klein at Vox. In this case better measurement boosts income inequality more or less permanently.

The assumption behind this sort of piecework measurement is that all the value realised by an article is due to the sweat and toil of a more-talented-than-usual journalist. If your article gets the clicks, then it must be because you are so good at what you do.

Unfortunately the world is not so simple.

We might choose to build our organisations around this sort of idea (and indeed, BuzzFeed et al work this way) but it tends to foster a short term and overly transactional view of work that ignores a lot of the value that workers, or a community of workers might create.

The first problem is the obsessive focus on outputs, on the assumption that the worker is responsible for all the value created. Outputs depend on inputs, and not just the worker’s skills. You can’t make a silk purse out of a sow’s ear, as the saying goes.

While the worker might be skilled, their work is also dependant on the quality of the materials they have to work with. Take the journalism example. A manager somewhere is splitting up the work, either by handing out the story ideas or by allocating topics to individuals. Not all ideas or topics are equal. It’s possible for someone to come from outside this system by finding a new approach—as Nate Silver did with a data-drivern approach—but that’s the exception rather than the rule. It’s more typical for the quality of the value of the outputs to be bound by the quality of the inputs, not the effort of the individual.

We see something similar in sales. It’s easy to sell in a rising market, and a booming market will see many sales people getting large commissions for no reason other than turning up. In a down market, though, its a different story, and we punish some of our best people for working hard just to bring anything into the business.

If we want to reward individuals based on their contribution then we need to quantify the amount of value they added, rather than the amount of value they lucked into. If we don’t then we’ll create a feeding frenzy for the juicy bits of work, while other less attractive (but possibly no less important in the overall scheme of things) get ignored.

Unfortunately it’s surprisingly difficult to measure value-add for many workers as it can be challenging to gauge the quality of the materials that they have to work with. A good example of this are the efforts in the US to measure teachers on the value they add in the class room, efforts which are struggling as it seems nearly impossible to objectively measure the quality of the students that they have to work with. There’s just too many variables.

Second is the problem of cumulative advantage. Success typically brings more success for no other reason than you were successful. Consider the opportunities created when you win an Oscar. The Oscars are an annual competition, so they’re awarded even if the year’s releases aren’t particularly good (such as if there’s a writers strike during most of the past year).

It doesn’t matter how you win the Oscar—either by creating great art and a big box office success, or simply be being the best of a bad lot—the attention that the Oscars garners you brings you to the attention of the world and the opportunities start flowing in. This improves the quality of the materials you can choose to work with. You might break the VW emissions story due to dumb luck, but it results in more story ideas flowing your way. You might not be the best journalist, you might not even be the journalist best positioned to make the most of the idea, but the idea is yours none the less.

Entire careers are built on the back of a lucky break followed by cumulative advantage. While this is good for the few lucky individuals, it’s not so good for the firm as it means that the firm might not be making the most of the materials at its command (though picking winners does make it easier for management). Nor is it much good for the equally talented individuals who weren’t quite so lucky.

Third is the problem of context. It’s rare, these days, to work in isolation. The context we’re in provides us with resources and connections that we couldn’t get elsewhere, or even just a boss that we can work with. While we might thrive in one environment, we struggle in others. One good example is star analysts, who often struggle when they leave the firm where they built their reputation. Some of that value in the outputs created might be the result of a productive work culture or effective management structure and team, factors that are the result of the everyone’s contributions, and not just the contributions the individual creating the deliverable.

Mr Cowen’s problem is that he has mistaken ease for cost. It’s cheaper than ever to measure all sorts of factors associated with work. At the same time, work has evolved making it hard know what to measure. While it might be cheap to generate all sorts of stats on worker activity, it’s not easy to tie these back to productivity.1)Aside, that is, for work situations which are explicitly configured as piece work, such as Uber drivers.

The root cause of this a recent shift (possibly sometime around 2005) from value being defined by the producer, to being defined by the consumer. The emergence of the consumer internet put the consumer in control as it enabled the consumer to have more information on a product than the merchant or producer, and the ability to source the product from any merchant around the globe. This was followed by the more recent emergence of social media, enabling consumers to turn to their peer, rather then brands.

Value used to be defined in terms of product features and functions, and we could measure a worker’s productivity by their contribution to creating these features and functions. Frederick Taylor started the trend by measuring how long it took for a man to unload a cart. The modern version is the basis of Mr Cowen’s article: counting the number and reach of articles carrying a byline, or worker surveillance where everything a worker types at a computer, everything they do is logged, recorded, and measured.

Value today is defined by a customer’s relationship to a product. Value is relative and shifting because it is a function of an expanding choice space for consumers. While all your workers contribute to creating this value, it’s not always obvious how to quantify their contribution.Their contribution might also be different for each customer, as relative value means that each customer could possibly conceive value differently.

Any retailer who heads down the omnichannel path, for example, needs to deal with the challenging of aligning a salesforce measured on their sales with a strategy that has sales skipping across multiple channels and contact points as the customer learns about the firm, develops their own understanding of what value is created, and winds their way to a decision. When you consider this it’s not surprising the Apple’s stores (some of the most profitable in the world) are not measured on sales, and fall under the marketing budget.

In the mean time we have many firms racing to quantify and optimise individual tasks that their workers undertake. This might drive improvements in a short term and overly instrumentalist definition of productivity, and result in a few lucky individuals receiving large pay checks. In the longer term the same strategy is destroying the value created for the customer, and possibly taking the firm’s future with it.

Image: Lainey Powell

References   [ + ]

1. Aside, that is, for work situations which are explicitly configured as piece work, such as Uber drivers.