3 ms·
One key here - eCPC dramatically ramps as a campaign goes on, given the same target demographic. If he was targeting a demographic like women in SE Asia, 18-35
by colinplamondon 13y ago
One key here - eCPC dramatically ramps as a campaign goes on, given the same target demographic.
If he was targeting a demographic like women in SE Asia, 18-35, then his test campaign would have first seen cheap clicks. Facebook's algorithms would find the cheapest users within that demographic, who are the least in demand by other advertisers, for whatever reason. CPC's would be low right off the bat, with a $200-300/day campaign.
Then, when he just suddenly increased to $100,000/day, he would have absolutely tore through the cheap available impressions within a matter of minutes, leaving only the more expensive users, with much higher CPC's.
As soon as the CPC skyrocketed, he should have pulled the plug on the campaign. But, it sounds like, there was a bad mix of both him being inexperienced with online advertising (needing training in their office) and FB sales reps encouraging him to let it ride.
If he had slowly increased budgets, from $300/day to $500/day to $1,000/day, he probably would have started to see a spike in CPC as soon as he moved it up to $1,000/day, and a vertical "holy crap KILL THIS CAMPAIGN" spike in CPC when he hit $1,500/day - $2,000/day.
It's really hard to manage campaigns at that volume - and there's entire teams of people needed to manage a campaign at the level of $100,000/day.
This reminds me of those Groupon horror stories. The small businesses were wrong to not running the numbers on deals that would clearly absolutely destroy them. But, the sales people at Groupon were even more in the wrong by taking advantage of those small businesses to turn a quick profit.
- downandout 13y agoBut you specify the maximum that you are willing to pay per click. Presumably he knew what he could pay per click for average users while maintaining profitability, and set the maximum accordingly. He doesn't seem to have a problem with his cost per click. The point of the post is that the "users" became increasingly less valuable as the volume of clicks he purchased rose. It shows that you cannot obtain large numbers of legitimate clicks/likes through FB ads. As the reach of the campaign grows, the percentage of authentic users clicking on ads appears to rise dramatically. That is a huge problem. My personal belief is that the percentage of fake accounts and bot activity is far in excess of Facebook's estimates, and there is little that they can do about it.
- colinplamondon 13y agoAgreed with that - I know that in my campaigns, I prefer having a bunch of smaller campaigns, which I can track the performance on at a very incremental level. One twist is also that a lot of advertisers don't do CPC bids with FB. I know for Mobile Installs, the default is optimized CPM, without a specific CPM bid. Just throwing it to the FB algorithm wolves, and seeing where it lands. I've found that has better results than CPC, CPM, or CPI bids, but, because it's entirely in the hands of Facebook, I have to watch it like a hawk. If I look away for a few days, the costs can get out of control.
- downandout 13y agoThat was meant to say the percentage of authentic users goes down, not up. Too late to edit the comment unfortunately.