How I'm got my Twitter CPC down to 0.0039 (i.e. less than half a cent)
FOLLOW UP: I realized that this was too good to be true and wrote a follow-up about my mistake: https://www.indiehackers.com/post/how-i-thought-i-was-getting-a-great-cpc-in-twitter-but-really-wasnt-acb5b8cb83
Just a brief note to help anyone doing Twitter paid ads.
I started using a strategy where I target the followers of my five main competitors, the ones with the audiences most likely to also get value from my screencasts about how to run and get customers for indie web-apps.
My ads are along the lines of "are you a fan of competitor? Then you might like my stuff".
At first I was getting a terrible CPC of €2.50 per click. But this was because I ticked the box letting Twitter automatically figure out my bid. It seems like you literally have to tick the "don't rip me off" box to get a good deal. (Or less cynically, maybe the algorithm just needs time to train and find you cheap clicks. IDK).
Anyway I went back in and limited my bid to €0.10 and sure enough my CPC went down €0.25 . Then I dropped it to €0.05. Twitter's algorithm did the rest.
Learnings:
- target your competitors followers
- don't do auto-bid
Things I'm unsure about:
- the number of clicks Twitter report is 4x or so what I'm seeing in Analytics. This may be because I'm targeting programmers and they are exceedingly likely to use blockers. If anyone has any other theories, I'm all ears.
You're targeting bots on the audience network
So I wrote a general follow-up (or rather: retraction) after digging into the data.
https://www.indiehackers.com/post/how-i-thought-i-was-getting-a-great-cpc-in-twitter-but-really-wasnt-acb5b8cb83
re: bots. Is there anything one can do to avoid them when advertising on Twitter?
I was able to explain about 50% of the discrepancy in Twitter clicks vs. on-site sessions after seeing the data in a bit.ly link I used for one of my ads. But I still can't rule out bots so curious to dig deeper.
Jack,
Two questions:
Incidentally do you do much Twitter advertising? And if so, is it cost-effective?
As far as I know (depends on jurisdiction) it's ok to talk about trademarked competitors as long as you make it abundantly clear that you are not that competitor. I understand trademark law's purpose to be to stop people from passing themselves off as someone they are not and tricking customers into transacting on that basis
Unfortunately disappointing, as I've just figured out. Basically I've retracted the post you read and wrote a follow-up with what I learned after digging into the Analytics: https://www.indiehackers.com/post/how-i-thought-i-was-getting-a-great-cpc-in-twitter-but-really-wasnt-acb5b8cb83
This is valuable experience mate. We’ve all done something similar. I thought Instagram was getting me insanely low cost acquisitions until I started to dive into other conversion events in the funnel and realised I was burning cash like crazy.
You live and you learn.
As for Twitter: it’s not really where my audience buys so I don’t do too much paid there. When I have, it has been to increase brand engagement around a specific event rather than for conversions.