October 06, 2022

Attribution

Catalog agencies use analytical techniques designed to encourage you to mail more ... and for good reason! They make more money when you mail more often.

If you execute mail/holdout tests, you know the truth. It's usually a startling truth, because the truth is so different from the story told by our trusted partners.

For a retail brand, here are results of a mail/holdout test run against RETAIL ONLY customers.


Matchback analytics lie the worst. Just because a customer bought from your favorite retail brand on a Tuesday two weeks after receiving a postcard doesn't mean the customer bought because of the postcard. Taking the postcard list and matching against every customer who bought within "x" days and then assigning every order to the postcard is malpractice. Think about it this way. You need to buy groceries each week. You have to. If Safeway sent you coupons for Doritos and you purchased Doritos, should the entire $275 purchase be matched back to the coupon distribution? No, never. The coupon has nothing to do with the $275 purchase.

Attribution analytics are not a lie, but they are a guess, and that's just as bad. Your attribution vendor parses the $275 grocery order and allocates shares of the $275 to your marketing efforts, as if they were inside the mind of your customer. Sure, they're using "advanced AI or machine language", but that is meaningless if they aren't using mail/holdout tests as the basis for their analytical genius.

In the image above, we see that the segment generated $2.69 profit via matchback analytics, $1.13 profit (per customer) via attribution algorithms ... but the segment of customers lost $0.18 when measured incrementally via mail/holdout tests.

Can you see why your agency wants you to believe their matchback analytics or attribution analytics?

The mail/holdout results above commonly happen for retail brands mailing retail-only customers. The results commonly happen for e-commerce brands mailing lapsed buyers. The results commonly happen for catalog brands with 90%+ e-commerce sales penetration.

Think about how differently you approach marketing investment if you choose to prioritize the truth over the stories told by matchback/attribution analytics? It literally changes how you think about marketing. By the way, the same concepts apply to email marketing. And display. And geographic tests for paid search. And paid social. 

Now imagine how your marketing efforts change once you know that what is displayed above happens across all of your marketing efforts?

Think about it.


P.S.:  Yes, you guessed right ... I'm going to get hate mail from two sources ... agencies ... and hardened 55-year-old-plus professionals who chose to believe agency stories (matchback/attribution) because it allows them to keep doing what they love doing. This blog has been active for nearly seventeen years, and the same cast of characters always berate me after I write this content. It's almost like politics ... if you tell lies long enough your viewers/customers believe the lies and run with them. There are "true believers" who adore matchbacks and flawed attribution analytics ... they will berate me and they are not the people who this post is written for. I'm writing this for you, the open-minded professional who wants to do the right thing for your company.

P.P.S.:  Here's a tweet from yesterday (click here). Yeah, he's correct.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.

Price is Not the Entire Story Here

We studied one category yesterday. Here is the relationship between change in price and change in total margin dollars, by category. Tell me...