Facebook wants to let you know when they think you’re targeting too broadly or too narrowly.
Check out this speedometer gauge on the audience counts.
In this case, we’re targeting the 90,000 people in the United States who list themselves as working at Facebook.
And here’s what happens when we narrow this down to Facebook employees who are 13-14 years old. It should be nearly zero, but shows 1,040 because of all the kids who can list their employment as whatever they want.
Notice the red coloring as a warning.
And earlier this year, they made their estimates rougher — to the nearest 100,000 for large interests and showing “under 1,000″ if the true count was 999 or less.
This makes it that much more important to trim your audiences to the proper size, independent of Facebook’s speedometer gauge.
Some of the “under 1,000″ audiences are nearly zero, while others are closer to 1,000. That can be the difference between spending nothing and $10 a day on that one target, so the only way to really know is to test.
Do you see bias in Facebook trying to get advertisers to spend more by encouraging larger audience sizes? Or do you believe that their optimization is getting so good that they can seek out the right users within a broad target?
We believe both are true.
In the last 6 months, we’ve seen ads that target an entire country for certain fashion companies still drive impressive retail sales. And if you’re tracking to an offsite conversion pixel, it’s even better.
Combining interest targets into fewer boxes, having fewer ad units available, and creating multiple ad variations from a single ad are examples of this.
Readers, what do you see as next in improvements to the ad interface and back-end tuning? Do you like it?