Weird Science: A Lesson In Failed Experiments

By Lauren Rosner | @lkrosner | Sr. Production Specialist

Much of what we do day in and day out in PPC is stare at numbers. Beautiful, finite numbers. We know numbers to be objective. If we spend $12,622 and that spend garners 31 conversions, we know the average cost per conversion is $407.16129. If our campaign is limited by budget and we raise our budget, we will no longer be limited by budget. So why, day after day, do I come to work feeling like a mad scientist, maniacally trying to keep accounts performing well?


Recently, after monitoring contextual keyword campaign performance for a month, I decided to make a few optimizations. Because there were a handful of top-performing placements in this campaign, I made the decision as I have done many times before in other accounts, to take the placements and bid on them directly in a different campaign. My logical train of thought went something like this: When contextual keywords match to this page, and we show on this page, we perform extremely well. If we can bid on this placement to ensure we are specifically targeting this placement, we should also do well and this new method of targeting should be a top performer.


It was obvious after a day of running this strategy in the account that it wasn’t going to work. It was DOA. I gave it another 48 hours, watching and waiting for something magical to happen. When nothing improved, I applied Newton’s third law. If by implementing this new strategy performance plummeted, then by reverting the strategy, performance will bounce back.


I reverted the changes I made in the account and waited for traffic to improve. When it did not improve, I was puzzled. I started to peel back the layers, get really granular and do all of those other PPC buzz-word things. Everything in the account was exactly as it was before my experiment. I was stumped. I wanted concrete answers and I had none. I had to turn to my client with mounds of data that didn’t add up. I felt defeated.


I turned to my Associate Director of Services, Diane Anselmo. She reminded me: “If PPC were an exact science, we wouldn’t have jobs and it wouldn’t be any fun.” Diane, you Jedi master! How easy for me to forget that nothing we do—no lever we pull—is done in a black hole. The landscape is always changing. Sure, there are patterns and trends to indicate performance and drive strategy. But there are competitors and current events and weather and any other seemingly random influencer of PPC. We all need to take five giant steps back from the numbers and look at the world around our account.


After three years of learning and growing to be an expert of PPC, Diane’s simple observation knocked me down a few notches. Sometimes you work hard and your creation comes to life and changes performance for the better. Other times you work hard and no matter what you do, your cold and clammy strategy remains lifeless.


So here is my strategy for when a strategy just doesn’t work:

  • Don’t flip out. It is really easy to get flustered when you do projections and research a strategy before launching and you think you know, for the most part, how things will pan out and performance just tanks. Deep breaths. Go to your happy place.
  • Triple check your settings. Not going to sugar coat it: Bing and Google can be sneaky. When creating new ad groups or campaigns the default settings most likely are not the settings you need. Ensure you armed your strategy for success by making sure the settings match your goals.
  • Contact the help desk for the platform you are using. A second pair of eyes is rarely a bad thing and they could help uncover an issue you missed.
  • Go back to the drawing board. You have two options here: keep trying to tweak the new initiative to garner positive results or revert your strategy and start fresh. If at first you don’t succeed, try, try again.
  • Don’t write off the strategy as a failure. Put it in your back pocket. If it has worked for other accounts, it still could work in the current account. Maybe revisit it down the road.


At the end of the day, we are all PPC heroes trying to do the right thing at the right moment for our accounts. It isn’t an exact science but it isn’t guesswork either. It is a beautiful and brutal landscape of digital marketing. There is no one best way to do anything. One hack does not fit all. Follow the data, follow your gut, make mistakes, and try again. And most importantly, keep having fun doing your job.