#everything-marketing
Thread
What's everyone's source of truth for advertising? Facebook's 1v/7c is very aggressive and says many of our ads are >3:1 ROAS. Triple Whale Attribution often says like <1 even as bad as <0.5. We ran a 30% off + Free Gift Sale this weekend, which is a mid-level promotion for us (we are a discount focused fashion brand) and I turned the ads off after $1k spend at our TW Roas was ZERO and our FB Roas (including views was only 0.5). I've been running ads for over a decade for my brand with over 5M in spend on FB and I've never seen ROAS like this!
I rarely look at FB post click numbers. Who knows what their “statistical modeling” is actually doing!
The only truth is the P&L.
Everything else no matter how snazzy is an educated guess.
Anything with views is going to be very generous. If you want the most accurate to reality set everything to 1DC only. Which will make your results look much worse, but may match up with your daily P&L a lot better.
A lot of brands had low results this weekend, whether its a Meta issue or a burned out from Labor Day Customer--I would see how this weekend goes.
How were your new customer sales this last weekend compared to how much you spent? That’s your truth.
I agree with @Peter Quadrel here. Your P&L should be a true indicator of the incrementality.
Check for trends. Has the discrepancy always been that way between TW and Meta? Have there been other channels in the mix? Any changes to the audience, targeting, etc.
P&L and attribution aren’t really related! Strong attribution understanding allows you to invest in the channels that are actually working Vs those that aren’t. If you’re a single channel advertiser then sure use the PNL. When I first started there wasn’t even a FB pixel. But if you’re spending on all the social channels and search platforms + creators and affiliate you need to know where you’re getting true incremental revenue.
True incremental revenue is hard to measure as the buyer journey is complex, you can def measure which had more impact by creating separate tracking (via coupons, landers....) to test what ultimately nudge buyers to complete purchase
To measure incrementality holdout test are the best I have found
If you want to rely on attribution look at data driven 1DC on TW. Or for your product price point: last click 1DC, that should get you closest to reality. Do you know your average sales cycle length from first touch point? Whatever that is should be your window, but if that’s not clear 1DC.