At the end of the nineteenth century, electricity was starting to have a profound effect on the world. As dramatized in the excellent novel The Last Days of Night, Thomas Edison battled with George Westinghouse (the latter aided by Croatian genius/madman Nikola Tesla) for control over the burgeoning market for electricity generation and supply. The popular symbol of the electrical revolution is of course Edison’s famous light bulb, but almost more important was the humble electric motor.
The electric motor was so important because it revolutionized manufacturing, enabling factories to create assembly lines and realize huge efficiency dividends. The Ball Brothers Glass Manufacturing Company, for example, replaced 36 workers with a single electric crane for moving heavy loads across the factory where they made their famous Mason jars.
But for all the benefits of electric motors, many factories were slow to embrace the new technology. As this article explains, by 1900, almost twenty years after Thomas Edison started selling electricity from his generation plants in Manhattan and London, only 5% of factories had switched from steam to electric power. Powering a factory with a steam engine was costly, complicated, and dangerous. So why the reluctance to move to electricity?
The reason lies in the way those factories were organized to take advantage of steam power generation. A typical nineteenth century factory looked like the image above. Mechanical power was generated by a single large steam engine which ran continuously, and was transferred to individual machines (such as looms or lathes) via a series of drive shafts, gears and drive belts. Because the power was being transferred mechanically, the machines were packed closely together. This, combined with the constant spinning of the drive shafts, made these factories very dangerous to work in; in 1900, over half a million people in the US were maimed in factory accidents.
Simply replacing the central steam engine with an electric motor did not deliver significant benefits – the drive shafts and belts to the machines still broke down, factories were still crowded, inefficient and dangerous, and the central motor (now powered by comparatively expensive electricity) still had to be kept running constantly.
To truly capitalize on electrification, factories had to reinvent themselves, replacing all their individual machines with versions that were powered by their own electric motors, with power transferred to them via wires rather than spinning drive shafts. In turn this meant that machines did not need to be so tightly packed together; factories could be reorganized to be more spacious and facilitate the flow of items, paving the way for the production line and improving factory conditions and safety.
Ultimately, it was the transformation in the way things were made which was electrification’s biggest benefit.
Reorganizing the marketing factory
The story of electrification in the first decades of the twentieth century provides an interesting parallel to the impact of data and AI on the marketing industry in the first decades of the twenty-first.
Today, many marketing organizations have adopted data the same way factories first adopted electricity: by applying it to existing processes and ways of working. In direct marketing, the core processes of list-generation and campaign delivery have not changed fundamentally in fifty years – marketers build a target audience list, map messages to this list, deliver those messages, and then measure the response. The sophistication and complexity of all these steps has changed dramatically, but the process itself is still the same.
However, as electricity led to the development of new kinds of manufacturing machines, so data is leading the development of new kinds of marketing machines, powered by AI. These new systems promise to transform the way that digital marketing is done. But just as before, getting there won’t be easy, and will require marketing leaders to embrace disruptive change.
The current ‘factory layout’ for many marketing organizations is based around individual teams that have responsibility for different channels, such as web, search, email, mobile and so on. These teams coordinate on key marketing calendar activities, such as holiday campaigns or new product launches, but manage their own book of work as a sequence of discrete activities. Some progress has been made in the last few years on bringing these teams together, and supporting them with a common set of customer data and common marketing automation tooling. But individual campaigns are still largely hand-crafted.
AI-driven marketing systems use a wide range of attributes at the customer level, combined with a continuous testing/learning approach, to discover which of a range of creative and messaging should be executed next, for which customers, in which channels. They break down the traditional campaign-centric model of customer communications and replace it with a customer-centric, ‘always on’ program of continuous nurture. For these systems to work well, they need a detailed picture of the customer, including their exposure and response to previous communications, and they need a wide range of actions that they can take, including the ability to choose which channel to communicate in for a given message and audience.
A marketing organization that is looking to evaluate the potential of AI-driven marketing will, prudently, lean towards trying the technology in a relatively limited pilot environment, likely choosing just one campaign or program in a single channel for their test. This makes sense – few companies can easily try out new technology across multiple channels, for both technical reasons (i.e. wiring the thing up) and organizational reasons (getting multiple teams to work together).
But this approach is a bit like a 1900’s factory owner deciding to replace just a single machine in the factory with an electric version. Dedicated (and expensive) wiring would have to be laid to power the machine; it would still be crammed in with all the others, so its size and design would be constrained; and it would likely need a dedicated operator. In this environment, it would be unlikely that the single machine would be so transformatively efficient that the factory owner would rush out to buy twenty more.
And so it is with AI-driven marketing. A test within a single channel, on a single campaign, will likely generate modest results, because the machine’s view of the customer will be limited to their experience with that brand in that channel; its message choices will also be limited, since it can only communicate within the single channel. These problems are exacerbated by the expense of laying dedicated data ‘lines’ to the new system, and of building many creative variants, to give the system enough message choice within a single
All or nothing
What’s needed is for AI-based optimization to be applied as an enabling capability in all marketing campaigns, across multiple channels and products. This requires significant investment in data and channel integration; but even more importantly it requires marketers, and marketing organizations, to operate differently. Digital advertising, CRM and e-commerce teams (and their budgets) need to be brought together; instead of marketers creating many discrete campaigns, marketers need to create more evergreen programs that can be continuously optimized over time.
The marketing factory needs to be organized around the customer, not the product or channel.
This kind of model represents very disruptive change for today’s marketing organizations, as it did for yesterday’s factory owners. In the end, much of the rise of electrified factories a hundred years ago was due to the efforts of newcomers to the field such as Henry Ford, who jumped straight to an electrified production line in the production of his Model T. Today’s marketing chiefs would do well to heed this lesson from history, as disruptors like Amazon, Tesla and Stitch Fix use process innovation to create streamlined, customer-centric marketing functions that are poised to exploit the transformative technology of AI.