Advertising platforms can find patterns quickly, but marketers still need to guide goals, data quality, creative and lead quality
Online advertising is moving steadily towards automated targeting becoming the only option. For most advertisers, that can work well: the PPC platforms have more data than we do. But automated targeting is just a machine for optimising towards the instructions it has been given. If these are wrong, the machine gets on with the job regardless. And we may not even know.
If we define success as a form being submitted, the system will try to get more forms submitted. If we define success as a download, it will try to get more downloads. If we define success as a sales meeting, it may produce more sales meetings. But are they the right ones?
The platform can see conversions, but they might include poor-fit companies, students, competitors, irrelevant geographies or other people who were never likely to buy. The automated system may not know the difference. So the quality of the input to the targeting system matters more than ever.
Good conversion tracking is no longer just important for reporting. If the data is weak, the campaign learns from weak signals. Worse, if initial data points in one direction, the system may keep pushing in that direction, even though that might not be the best one.
In addition, the incentives of the advertising platforms are not the same as ours. They aim to spend smoothly, learn quickly, and scale. We want profitable customers. So automated campaigns still need watching with a view of the business outcome beyond the headline numbers. We need to ask the right questions. Are the leads useful? Are certain types of enquiry increasing but going nowhere?
If the machines are to manage the audiences, the human role needs to move to defining good outcomes, writing better ads and making sure the algorithms are being pointed at something worth finding.