Opportunity cost and diminishing returns in user acquisition

In my column last week about the shrinking mobile user acquisition team, I argued that algorithmic campaign management tools from Facebook and Google have alleviated the need for large, matrixed advertising operations across different geographies, device types, channels, etc. A common approach to Facebook VO campaigns is to simply target as broadly as possible with a minimum ROAS target and let Facebook’s algorithm do the heavy lifting — if such a strategy can profitably deploy millions of dollars in marketing spend per month, how large of a media buying army does an advertiser need?

Of course there is nuance that is missing in the above assessment: each product’s audience is unique and successful marketing strategies need to be tailored to the needs of a particular business. I have certainly seen Facebook VO specifically not work for many types of apps, namely subscription apps (there are no subscription whales and so the magnitude of spend is the same for everyone, rendering VO somewhat impotent), hypercasual games that are monetized exclusively with ads, and low-priced D2C products. And it’s common, especially at conferences, to hear of cases where Channel X is the best performing channel for App Y, where X is some obscure traffic source with limited potential for scale.

I don’t doubt these people when I hear these claims, but I always question whether they fully appreciate the paramount, authoritative measure of success in mobile advertising: profitable scale. Nothing aside from profitable scale really matters, independently, in a user acquisition operation: not channel diversity, not number of ad formats, and not the qualitative virtue of ad creative. A user acquisition team’s first and only goal is to deliver scale within the parameters of some business objective (which might or might not be profitable unit economics). When teams utilize channels explicitly for the sake of not having their budget be concentrated, they are ignoring that imperative.

It’s common for me to meet teams that divide employee time across a wide swath of traffic sources even though some of those channels only represent a single-digit percent of spend. There can be a significant amount of opportunity cost incurred by having a team member focus on a channel that isn’t Facebook or Google just because it isn’t Facebook or Google: that cost is often profitable spend on Facebook or Google. Every channel brings with it some administrative overhead: dealing with billing, install attribution setup, creative management / re-sizing, cost attribution, reporting, etc.

If an advertiser is spending profitably on a small portfolio of the largest advertising channels already — which almost all are with some significant percentage of total budget — and wants to consider a new channel, they have to determine whether that same amount of incremental spend would deteriorate the performance of their existing campaigns, and whether that degradation would be worth the added cost of managing a new channel.

Put another way: if I’m spending $100k on Facebook and Google and have an addition $20k of budget to spend, I need to answer two questions:

1. Would I get better performance from a new channel on this additional budget than I would get from adding it to Facebook / Google? eg. after 120 days, Facebook / Google would deliver $25k from that $20k in spend and New Channel would deliver $30k;
2. If yes to #1, is that delta in yield on the incremental spend larger than the administrative expense of managing a new channel? (Also, if the answer to #1 is yes, the existing budget allocation should be examined)

The point here is that an indifference curve between Facebook / Google needs to accommodate that opportunity cost of a user acquisition manager’s time (if they are spending $100k / month on Channel X, could they be spending $500k on Facebook / Google?) as well as the overhead of interfacing with an additional channel.

Until the very largest channels by ad spend showcase negative marginal returns — that is, until the next dollar spent is unprofitable — it often doesn’t make sense to diversify outside of those channels (more on this idea in The “Quality vs. Volume” fallacy in mobile user acquisition). As uncomfortable as it might seem, most user acquisition teams would be better off operating only on the largest channels — concentrating their spend, developing channel specialization, and minimizing their analytical overhead — versus having their team’s attention and time spread across a wider array of channels with a longer-tailed distribution of ad spend and revenue.