For about a year, I worked as a writer alongside specialists in performance-based advertising. Those were quite passionate advocates of gathering data on user actions to get more out of ad campaigns.
Down below, I’d like to give you a sort of interview. A summary of thoughts of one ad specialist named Eugene.
Eugene started working with digital advertising 7 years ago. Back then, performance-based ad campaigns were a new thing for most business owners. Some still require its explanantion as of now.
What Makes Performance-Based Advertising
The whole performance-based advertising is built on the idea of tracking every result of each action.
Before analytical and advertising systems have developed to reach their current state, the dominant share of advertising was brand-based.
Brands used to pour lots of money into making themselves visible. They just wanted for everyone to know them and hoped to get more income through building up their fame.
Then, the trend has shifted. The current data tracking capabilities allow us to track everything and make informed decisions. Moreover, you can set a threshold and show ads only to people who spend more than the specified sum.
How Performance-Based Advertisers Use Data
The data on all potential business clients is accumulated and depersonalized. Ad specialists, of course, do not track data on each separate person.
Data gets combined for groups of people who share characteristics.
If a person fits into a group, we can anticipate future actions of such person based on the past actions of people within the same audience group.
So here’s the final answer as to how data is used in performance-based advertising.
The data on past actions lets you predict the future behaviour up to very specific details. You can track your current customer flow and model all future interactions of potential clients with a business.
What you get in the end is your ability to make better decisions.
This is what happens on the side of advertisers. How do ad systems use data?
Some Words About Optimization
Advertisers make predictions on the most profitable outcomes. So does an advertising system.
Complex ad systems like Facebook and AdWords each have their own optimization script. Facebook Ad Manager uses its Facebook Pixel.
A script registers user actions at the source where you install it. Having gathered the data from the past, a system can start showing your ads to users like those who’ve already performed the required actions.
This a win-win-win situation.
Advertisers reach the relevant audience, users see the relevant ads, and the system gets its margin. Everybody’s happy and everything works smoothly.
One thing is crucial here: you have to set up optimization in advance.
Ad delivery can be optimized so that a certain action happens more frequently. In such a case, you need to make it clear for the system what is this action you need.
Moreover, optimization bases on the actions that have already happened.
Let’s say that you want to optimize your ad delivery to receive more conversions and select conversions as your optimization objective. For this to actually work, the system needs the past data on conversions.
If the script you’ve installed hasn’t yet registered enough past actions, your ad delivery won’t be optimized.
But once you have the data for the optimization, you can step aside from guessing characteristics of the audience you want to reach. You don’t have to pick interests and narrow demographics. You can use the actual data on real people and scale your ad activity to the infinity and beyond.
With the data from the optimization script, you must build no more hypotheses. That’s when you start testing ad offers instead of drafting user personas.
At this point, you can just keep on scaling. Who knows where’s your limit?