Over the past few months we’ve come up with a system that lets you reliably attribute conversions and make better spend decisions.
It involves 3 separate data points that you most likely already have.
1. MER - Marketing Efficiency Ratio
This is your ultimate source of truth. As a business you will have a target MER - for e-commerce businesses this is usually somewhere between 3-5x of spend.
It needs to be monitored weekly as you scale spend up and down. If you are under your target MER then you either need to shift budget between channels or change the spend distribution.
If you are above target you can increase budgets, you just need to know where.
Which is where the second data point comes in.
2. Post-purchase Attribution
As channel level performance data has become much less reliable, having 1st party data is crucial. Asking customers a simple question is really powerful: Where did you first hear about us? This will give you an insight into how people think they discovered you.
With top of funnel channels like Meta and TikTok you usually find that self reported attribution is much higher than in-channel attribution.
It may tell you for instance that Meta drives 50% of your conversions. Now you can overlay this onto your new customer revenue figures and find your true revenue. And consequently your true ROAS by overlaying spend.
The goal is to align the % of conversions your channels are driving to the % of spend you are allocating to them.
If you find that Meta is driving 50% of revenue but only 30% of budget is allocated you have an opportunity to increase spend.
You can monitor this data weekly or fortnightly depending on volumes.
3. GA Revenue Proxy
The final KPI you can look at to make sense of your data is GA. Last-click is notoriously undervaluing channels such as Paid Social. So this needs to be treated with caution.
To make this data more useful you can come up with a proxy. From our post-purchase data we know that Meta might drive 50% of total conversions. In GA it may only report 20% of conversions as coming from Meta.
This gives you a proxy: if you 2.5x your GA numbers you have a rough revenue figure you can trust.
It can be helpful to look at overall campaign performance. How do these campaigns stack up against each other when you 2.5x their performance and overlay spend?
Putting it all together
How does this look in practice? The first step is always to understand overall marketing efficiency. If your MER is above target, great! Now you just need to figure out which channel is driving it.
By checking the post-purchase data you know Meta is driving x% of spend and y% of conversions and you can figure out your true ROAS. If it beats your target you can increase spend.
Finally to sense check in-channel performance you can look at your GA data and overlay your proxy to understand which campaigns are driving the most efficiency.
Et Voila! Now you can have attribution model that directionally tells you what is working and where to allocate more and less spend. Unless you are spending 500k+ a month on ads having an attribution tool is very likely not necessary.
Happy scaling.
Comments