Take a look at this graph. What do you see?
If you take it at face value, you initially see an inverse relationship between ad spend and revenue. Is it possible that increased marketing activity is actually having a negative impact on revenue? That doesn’t make sense, but data doesn’t lie, right?
While data doesn’t lie, it can certainly mislead. It’s not the data’s fault - we just have to make sure we’re interpreting it properly. This data is not showing us an inverse relationship between marketing spend and revenue; it’s showing latency between investment and return. For this Working Planet client, there is a ~4 week turnaround from the time someone initially engages with an ad to when they actually generate revenue. There is some significant and consistent lead time required to process materials for this business.
Given nothing but the in-week results from the graph, and a ROAS target (200%, in this case), we would be stuck on this perpetual performance rollercoaster: pushing ad spend when efficiency looks great, then slamming on the brakes when it drops significantly below that 200% threshold… never actually growing their business.
The solution: look at cohort performance. In this case, the revenue tracked in a given week has very little to do with what was spent in that week. Instead, we should look to a day ~4 weeks ago to understand the revenue we track today. This provides a much clearer picture of performance based on cause and effect. It allows us to break out of the performance roller coaster. Instead, we can build sustainable and profitable growth, tolerating short-term losses on in-week performance and focusing instead on the longer-term gains.
Key Takeaways:
Seek to understand how latency impacts your interpretation of performance
If something doesn’t make sense in your data, dig a little deeper
In-day/week/month performance is important for cash flow, but using cohort performance to measure success is the key to sustainable growth
Comments