Managing the Risks of Advertising Campaigns by Philippe Barbe 17 Mar 2021

Part 19 of 20 in a series examining the interplay of Data Science, AI, the media and advertising.

Doing business is risky business! Many things can go wrong, which is why companies come and go.

The media, advertising and content creation industries are no exception. Antiquated business processes (detailed on part 10) make the future of the edited media industry unclear. The automation we’ve been discussing is absolutely essential for it to remain competitive.

But automation is not sufficient, as it only addresses the mechanical aspects of operating at a larger scale. A cognitive shift is needed as well, where actors think not only of objectives but of risk as well.

The combination of the current cumbersome business processes, low transaction volume and the high dollar value of transactions requires individual attention to manage the multiple uncertainties involved including audience forecasting, revenues, preemption, and execution.

One can deal with risks in several ways including more oversight and procedure which increases the cost of doing business, or spreading the risk through some form of diversification which is viable if the cost of doing business is insensitive to scale.

By enabling a larger scale operation, the Exchange proposed in part 12 offers a mechanism to hedge risk through diversification. The financial and marketing objectives of the players would be achieved or exceeded through a larger number of smaller transactions.

Most of these transactions would not need specific monitoring since their individual effects in a campaign or on the budget of a media outlet would be small. The collection of all transactions would be managed like a portfolio. Because of its scale and low latency, the Exchange would allow instantaneous correction to that portfolio for all parties involved.

Campaign Planning stage

At the campaign planning level, the Exchange would allow advertisers to see different campaign proposals with various prices that are based on the uncertainty involved with each proposal and how much protection against the uncertainties the advertiser needs.

These Exchange-generated proposals would be very different than those that exists today, in two ways:

1. Breadth. Since more media outlets would expose their inventory for sale, proposals could contain thousands of line orders drawn from a much more diverse inventory. Intelligent visualizations would allow planners to compare price and risks of proposals at any level, from a complete aggregation reflecting the entire campaign to the individual outlet and programs / section / content where ads would be placed. These would be the basis of negotiations if needed.

2. Transparency. These proposals would also help advertisers understand how the different risks impact prices and the needs to monitor execution. For instance, the cost of requesting high priority to limit preemption would be clear. And uncertainty of audience forecast would be tied to uncertainty on the price requested to achieve marketing objectives.

All in all, Exchange proposals and their associated visualizations would provide much better information to make rational, more predictable and more profitable decisions in an environment which is always uncertain.

In such a system, Marketing departments would keep their role as defining campaign objectives, albeit with better information while Planners would spend less time assembling a plan for execution but more time balancing risk and cost.

Campaign Execution

The straight-forward goal of an advertising campaign… reach a certain audience with a certain frequency… assumes that the rate of reach is constant over time, or should fluctuate according to some specification.

To make this point clear, consider the example of a 3-month campaign aiming to reach 2 million people, with each person being exposed to the ad 30 times. It might be possible to accomplish this in 3 days with a massive advertising blast but implicit to the 3 month timeline of the campaign is that exposure should be about 10 exposures per person per month or somewhere between 2 and 3 per week.

Because of this desire to reach the audience over time campaigns are constantly monitored for under-delivery. Thus, during the execution phase, the total campaign objectives are often broken down into weekly objectives. With that breakdown comes more risk if one focuses on each outlet’s delivery: an average calculated over a week is less stable than an average over 3 months.

With the Exchange, media buyers and agencies would still monitor campaigns as they currently do but in a different way as the performance report would also contain indications of:

• whether a fluctuation in delivery is within statistical errors to be expected,
• what is at risk for the remainder of the campaign,
• recommendations on how to mitigate that risk by shifting some of the budget from media outlets where the campaign is over-delivering to those where it is under-delivering.

Because of its scale and near-real time capability, the Exchange would offer advertisers a user experience and agility similar to that in the digital world with faster process flows involving many more transactions at smaller operating costs, hedging of risk by tapping larger inventory for diversification, and smart execution management and monitoring.

The Exchange concept doesn’t try to fight the complexity and fragmentation as the industry is doing now.

Instead, it aligns with the historical trends and makes it possible for the ecosystem to operate at a larger scale, in a cost effective way, and in a manner that allows management and reduction of the risks associated with advertising campaigns. But to be effective, it requires a more conceptual way of thinking from advertisers than their current view.

Optimizing inventory usage and meeting revenue targets

On the media operation side, the Exchange would track revenues by providing demand forecasts, and therefore revenue forecasts.

Similar to what it offers advertisers, it would offer media operators risk estimates on their expected revenue. Running a media operation would then be balancing act between having a steady, certain fixed income versus taking more risk with potentially higher rewards.

The Exchange would not give incentives to adopt any specific strategy but it would give media operators the possibility of defining and executing any strategy they decide upon. That strategy could be either at the outlet level or, in the case of media group, at the corporate level.

Because it would allow media operators to conduct their business at a more conceptual level and in a far more efficient manner, the Exchange would allow media outlets to leverage the technological evolution outlined in part 11. And, equipped with ways to buy more content as expressed in part 18, they would be able to manage far greater inventory with better workflow and reduce risk by diversification.

They could spend much less time on each transaction, offer more inventory, allow more audience segmentation, and dilute the risk on each individual transaction.

Higher profitability would result with replacement of the current low volume/high dollars model with its high costs, risk and attention to a high volume/low dollar model with lower cost, risk and attention.

Conclusion

Operating a global media/advertising ecosystem with millions of daily transactions and millions of concurrent users requires not focusing on each transaction or customer, but to think in statistical aggregate terms, with few possible exceptions, such as Super Bowl advertising, that need individual attention.

The unedited digital media colossus was built upon the ability to conceptualize a market size never dreamed of before and then use newly available technology to reach and offer that audience to advertisers.

The traditional edited media is a victim of its past, stuck in ways of thinking that do not align with present reality. The outsized focus on each single transaction and expensive use of resources, which works at the scale at which the business has historically operated, is ill-suited for its future.

To deliver performance and increase its market share, the industry needs to change its way of thinking, more adapted to scale, more statistical, supported by automated execution, with decisions made rationally in terms of risk and benefits.