Amazing Subtleties

Here's what the analysis looked like for an e-commerce brand. This brand claimed that customers acquired via Paid Social were "lousy". Any analysis, even after controlling for customer quality, seemed to show that Paid Social customers were "lousy".

Dig into the data a bit deeper, and the suggestion that Paid Social buyers were "lousy" was "inaccurate".

I categorized each item the brand sold based on the channel that the item sold best in. If the item sold best in email marketing, it was given a "PREFERENCE = EMAIL" indicator. If the item sold best in paid social, it was given a "PREFERENCE = PAID SOCIAL" indicator.

Again ... every item is given a PREFERENCE based on actual sales performance across channels.

From there, I could analyze customer behavior controlling for customer quality, the channel PREFERENCE of each item sold, and the channel the customer purchased from. For instance, this e-commerce brand saw the following.

• Customer buys PAID SOCIAL PREFERENCE ITEMS via EMAIL = $100.00 Future Sales.
• Customer buys PAID SOCIAL PREFERENCE ITEMS via PAID SOCIAL = $80.00 Future Sales.
• Customer buys EMAIL PREFEFENCE ITEMS via EMAIL = $120.00 Future Sales.
• Customer buys EMAIL PREFERENCE ITEMS via PAID SOCIAL = $100.00 Future Sales.

In this example, marketing channels are only half the story.

• The paid social channel delivers customers worth $20.00 less in the future (sales).
• PAID SOCIAL PREFERENCE items deliver customers worth $20.00 less in the future (sales) regardless whether purchased via paid social or email.

In other words, in this example, the merchandise is HALF the problem. It is what sells via Paid Social that is half the problem, with Paid Social as a channel being HALF the problem.

Extend this analysis to all of your marketing channels, and get ready to be surprised. Yeah, what you sell matters a lot more than you think - and if you purposely push certain items within certain channels, you are creating interesting customer feedback loops.

There are amazing subtleties in your data, if you choose to look at your customer data in ways that Shopify or Google haven't pre-designed for you.