Kevin Hillstrom: MineThatData

How Will You Give New Items Exposure?

It's hard for the customer to know you have new merchandise unless you tell them you have new merchandise. You do that through print (for some of you), you do that through email marketing, you do that through your home page, you do that via social media.

Now think a bit about the evolution of search. Your search-centric customers already skew to existing items ... just perform the analysis yourself, it's easy to do and it is revealing as heck. Algorithms push customers to items that algorithms know about.

AI likes/needs to be "trained". How do you train an algorithm on an item that you won't offer yet for two more months?

Email marketing is going to become your most important new merchandise advertising platform. It will be one of the few places you have control over the message. Start your experiments today, learn as much as you can, and be ready for the change that is coming.

The Organic Percentage Over Time

When I left Lands' End in 1995, one of the challenges the analytically-focused person faced was the fact that some in Management did not believe that if you didn't mail a catalog 30% of the demand would still happen (from surrounding catalogs). This "fact" (measured via mail/holdout tests) meant that what was marginally profitable was, in fact, unprofitable. Nobody wants to be tied to something that is unprofitable. Consequently, facts were not accepted.

At the time I thought "these people just aren't smart enough to understand simple math". I was wrong. People are smart enough to understand facts, be it in marketing or politics. People choose to ignore facts for perfectly good reasons.

The organic percentage, at 30% in 1996, would become larger, creating ever-bigger problems.

By decade, the organic percentage grew.

1996: 30%.
2006: 50%.
2016: 65%.
2026: 80%.

In 2026, the only customers that can be profitably mailed are good or great customers. That's it.

And I get it ... some of you are reading this and you go back to your matchback analytics and you say "I'm good". You run a p&l on each segment and you believe you are mailing 36 month customers profitably. If you believe this, execute the following experiment.

Assume your organic percentage isn't 80%, assume it is just 50%. Pretend like it is 2006.
Run a p&l for each segment assuming your organic percentage is an old-school 50%.
What does the analysis tell you?

It tells you an inconvenient truth.

If you do what the analysis says, your company will be more profitable. Much more profitable.
If you do what the analysis says, you will cost your company top-line sales/revenue/demand, and you have forces within your company who will never let you do that.

This is why so many catalog brands ultimately went bankrupt. They couldn't let the top-line contract, so they just kept doing what they were doing until unprofitable became "too unprofitable" while customer response also contracted.

When the AI bubble pops (and it most assuredly will pop), new ecommerce models fueled by what AI is supposed to become will emerge. Print will be forced to be leveraged only by brands with the GDP of Bolivia or will be a tactic to leverage with best customers with high iROAS. It will not be a channel to center an entire business around (and honestly, it hasn't been for 20 years). Also - the same forces that came for catalog marketing will come for old-school ecommerce in a few years. It's how capitalism works.

An Interesting Evaluation Of Marketing Effectiveness

A link was offered to download a "white paper" demonstrating that print has fabulous ROI.

The "good" from the PostPilot report?

S-Tier Analysis: Conducted via mail/holdout groups. Yeah, they went there. Kudos!
They Didn't Lie: Their results "make sense". They aligned with reality.

The "not so good" that made me #sigh?

Results derived from highly responsive customers.

You can cover a whole bunch of sins by executing marketing against the most responsive customers. Just ask the loyalty industry for verification. Here is an example of the kind of direction they're leading you in.

There is nothing nefarious happening here - the analysis is measured properly, the organic percentage is common/reasonable, and iROAS (incremental demand / ad spend) is incremental.

Is there a flaw in the analysis?

No.

Is the analysis misleading?

Yes.

Why?

The analysis was executed against highly responsive customers ... ones about to buy or ones who just purchased. In either case, the customer is not representative of the customer base. EVERY marketing effort "is likely" to work when measured against highly responsive customers. Including print. Especially print!

The logic falls apart when you take the same conditions and apply them to, say, a customer who has not purchased in 18 months.

Everything is the same here - same incremental lift, same ad cost, same organic percentage. However, the customer is simply not responsive. As a consequence, the marketing channel is not effective, with a horrifying iROAS and a loss.

That's how you make something appear to be effective - execute it against responsive customers.

But when you try to extend the audience? Not so good.

Email and Social are the ways you speak with non-responsive audiences. Ad costs are so minimal that the numbers work.

You Get This Newsletter Every Saturday, Right?

If you want to see what a mix of professionalism, free information, and personal expression looks like, get on this email subscription (I've been here for more than a decade, long enough to see his kids grow up and move into college and begin careers).

Saturday's email was about the resurgence of brand marketing ... which is interesting because on many recent video conferences there have been robust discussions about the death of catalog marketing and the importance of email marketing as one of the few ways to have an actual "relationship" with a customer.

25 Years

It's almost unfathomable that a full generation has passed since 9/11. Twenty-five years. More than a quarter of the people who experienced that day are no longer with us, including 3 out of 8 of the adults who experienced that day.

Pressure causes people to behave in interesting ways. On 9/11, I recall a peer VP sobbing. An Executive looked at me and said "Why is she so sensitive?". He then locked myself and my Marketing Director in a conference room from 2:30pm - 7:00pm because, as he said, "we'll get an opportunity to fix the business and nobody will be here to bother us". I sent my employees home, then failed to fix the business during the 4.5 hour session. Obviously, the Leader was responding to pressure in a manner inconsistent with logic. It's easy to feel empathy for the Leader today. I felt something different that afternoon.

A woman I've played pickleball with described what it was like to try to get out of the city on 9/11. Her ordeal was obviously more harrowing than a 4.5 hour session in a stale conference room.

It's like the events of that day sent us down a different reality path ... not unlike the one in Back to the Future where our heroes end up in a timeline controlled by Biff Tannen.

No matter what happens, we move forward. We are resilient.