We (along with the entire industry as far as I know) have seen Facebook's performance decline since Apple's introduction of ATT. Over the last 12 months, we've made channel exploration and expansion a core focus. One channel where we've seen surprisingly great performance is Nextdoor! In this post, I'll do a quick overview of the Nextdoor UI and a few callouts about creatives. We've also pulled actual ROI results for one of our clients that we've indexed alongside their other major digital channels (Facebook, search, etc).
1) They have a self-serve ad platform that follows the typical Campaign > Ad Groups > Ads convention. It's more or less modeled on Facebook's UI.
2) They have a typical JavaScript pixel that supports both "Standard" and "Custom" events.
3) Targeting is more or less what you'd expect. You can go down to the zip code level and up to the state/country level. They have a custom audience solution alongside the standard device, interest, and demographic targeting options.
4) Bidding is a bit of a throwback. All bids are set on a CPM basis and they still support a traditional frequency cap.
5) Nextdoor has a few placement options which they outline in detail here. We've generally seen the best performance targeting their Newsfeed placement which is essentially the default option.
They do provide this somewhat complex budget allocation/bid adjustment tool but, as noted above, we usually max out the newsfeed placement 100% and call it a day.
6) We recommend testing 8-10 creatives at launch and grouping them ~3 per ad group. We also recommend periodically killing low CTR creatives (say under .45%) to help force spend on newer creatives.
7) We strongly recommend focusing on image/static assets (which is a bit unusual!).
This lead gen client spent just over $4.6m via digital channels in Q4 of 2022. They do last-click attribution on lead submit based on UTM parameters and measure cost/revenue ("cost of marketing") for each cohort of leads.
In Q4 2022, Nextdoor comprised a very respectable ~5.5% of total spend.
I wanted to look at cost per lead, cost per sale, AOV, and cost of marketing. To do so (while keeping the data anonymous) I indexed everything relative to the FB performance. So, for example, a 1.0x CPL is equal to the CPL we saw on Facebook, and a 1.5x CPL is 50% higher than the CPL we saw on Facebook for this period.
You can see above that the Nextdoor cost per lead was significantly higher than that of FB (1.74x), but the conversion from lead to sale was also significantly higher (insofar as the cost per sale was pretty close to Facebook at 1.10x). That translated to Nextdoor being a bit more expensive (1.23x) than Facebook on a cost-of-marketing basis but still very much in the same neighborhood.
PS: This post was written in close collaboration/direction with Mason Matney. Mason + team are our resident Nextdoor experts at Thesis.
