On most days this December, our party game, Ransom Notes, was the #1 board game bestseller on Amazon (occasionally #2 — damn you, Connect 4.) Getting to #1 was directly tied to a very effective Facebook ad campaign, but an unusual one: we’ve found the most effective way to sell games on Amazon is to send 100% of ad traffic to our own website.
We stumbled into this “Facebook Ads halo effect” in late 2020. Despite then being a two-person company staffed by at least one moron (it rotates), we’d inadvertently equipped our little company to notice and measure this phenomenon. We’d launched a handful of party board games on Kickstarter, at platform that in 2020 had ad tracking tools roughly on par technologically with the compass. So, we did our ad efficiency tracking the old fashioned way: a spreadsheet that took daily advertising spend, divided it by games sold, and spat out a daily “Cost per Acquisition” — the ad dollars we spent to sell each game. If that number was $7 or below, great — turn up ad spend. If it was above $12, we might be losing money.
Once we started selling on Amazon, we just redirected our Facebook ads to our Amazon listing, simply because it converted higher than our own website. Amazon’s version of ad tracking is putting a bag over your head and dropping you in the woods, so our spreadsheet stayed our one source of truth — advertising spend in, daily units sold out, and trying to measure anything deeper was for people smarter than us.
Fast-forward some (montage of us stumbling through the woods, bags still on our heads, Jeff Bezos cackling), and we decided try out Facebook’s fancy “conversion-optimized ads.” Since these aren’t compatible with Amazon, we had to run them to our own dinky website. But at least a feedback loop could now happen — FB’s algorithm could learn who was actually making purchases, not just clicking, and optimize itself to show ads to similar folks.
From looking at the Facebook Ads dashboard, these new ads were a disaster. Our Cost per Acquisition, as measured by Facebook, was $35 — our games cost $29.99. Since this is the advertising equivalent of your airplane’s altitude readout being “underwater”, we rushed to kill the experiment. Luckily, our manual entry, stone age-era spreadsheet let us stick our head out the window, and we saw a very different story.
From an overall “money in, units out” perspective, our Cost per Acquisition was actually below $7 — quite healthy. Our spreadsheet didn’t differentiate between games sold on our website versus Amazon, and when you mashed it all together, the effect was clear — despite the ads pointing to just our website, they were having a dramatic effect on our Amazon sales that Facebook’s dashboard couldn’t see.
Being the coldhearted fun-suckers that you’d expect party game inventors to be, we were immediately skeptical — people were clicking our ads, ending up on our website, but then deciding to search for and buy the game on Amazon instead? Our spreadsheet didn’t lie, though, and even after we ran tests of turning ad spend way up and way down, Amazon sales responded accordingly. We even made a fancy correlation graph, once, that I was too lazy to dig up.
Once we recognized this halo effect, the Facebook ads dashboard became useful again. A $35 CPA there roughly translated to a true $7 CPA once you factored in Amazon sales, so even though our Facebook-side metrics looked awful, they weren’t — Facebook just wasn’t measuring all the people that decided to buy on Amazon instead.
Armed with this new knowledge (yet still with bags over our heads, a dangerous combination), we put out a new game called Ransom Notes. Ransom Notes turned out to be a particularly good fit for Facebook ads — easy to understand, idiotically amusing — and the halo effect held up. We now do 65-70% of our sales on Amazon, despite zero Facebook ad spend directed there.
This holiday season, we’ve just continued to follow our trusty spreadsheet and look at sales holistically — from the Facebook dashboard, we’re madmen willing to lose thousands per day on ads that aren’t working. From Amazon’s side, we’re #1.