So I learned a lesson about online advertising.
Last week I gloated about how I got my CPC down to less than half a cent on Twitter.
After posting it here, I got some insightful questions about conversion rates, so I investigated the data in Analytics to answer them and discovered a few disappointing things:
So, narrowly I learned this lesson: Always set country in online advertising.
And more broadly, I learned to dig deep into your Analytics and split conversion rates up by country, device, landing page, etc. and figure out what's performing best before embarking on the next iteration of paid ads.
Always slice it up to customers LTV, so you know if that segment is profitable. At times is really hard to tell why one group is better/worse than another. Even when on the surface all is the same. One extreme example is that Facebook click start with the best and cliff jump in quality, at least was my experience a long while ago 🤷.
Great point. The only thing that ultimately matters is LTV. In my primary business, Oxbridge Notes, a typical Adwords first purchase might be worth 50 bucks. But (if I remember correctly... which is a big if...) the average customer rebuys three times over the ensuing years, therefore LTV is more like 200 bucks (and therefore it makes sense to bid higher than the 50 of the first sale).
You do much advertising? What's your main channel?
Had past experience across the breath of internet marketing community, but was out of the game for several years, currently building a developer tool..
Cool!!! What's it do?
I've been indie for 10 years and my main channels are SEO and paid. I spent too much of my life fighting horrible ad interfaces. Had a random thoughT today: I wish there was Terraform for Analytics/AdWords/Fb ads etc. -- some recipes to set up everything nicely.
https://www.indiehackers.com/post/programmers-buddy-f3c3f5a77e
https://www.indiehackers.com/post/tool-for-developers-interest-capture-page-unnamed-coder-bot-3d57ad4690
Recently found Dependabot by @greysteil, a more generic service of similar concept.
In my experience, Ad providers (e.g. Facebook and Google) will always encourage you to go the path of purchasing their least desirable inventory.
The best way to know NOT to do something in those platforms is when an option has the "(recommended)" text next to it.
Also, click fraud is a huge issue. We have detected headless bots coming from Facebook ads, which is suspicious since Facebook itself is the only one with the incentive to do this.