Kevin Hillstrom: MineThatData

Conflicting Data

As you've been told, I write a Weather Newsletter. I also write a Pickleball Newsletter, but that's a topic for another day.

About three weeks ago an epic heatwave gripped the Western third of the country. I received two themes of feedback.

This is what happens when climate change is amplified. The future is worse than this.
This weather pattern is a normal weather pattern with an extreme outcome and has nothing to do with climate change.

Both sides are very passionate about their argument. They're quick to educate, they're in some ways overconfident ... they surely believe they are "right".

Forty years of data analysis suggest that humans struggle with conflicting data. We struggle with feedback loops. We struggle with interactions.

It happens in ecommerce. Remember our Category Impact table.

Category 17 is the biggest sales driver (sales are not shared in this table). It is also the biggest source of new customers. It also negatively impacts Category 6. This means when Category 17 does a good job, Category 17 hurts Category 6.

Assume you are the CEO of this brand. The merchant in charge of Category 17 looks great. She's responsible for a high-volume category that brings in new customers. And yet ... every time she does a good job, the poor merchant in charge of Category 6 looks bad. His sales decrease.

Should the CEO fire the guy running Category 6 if sales decline in Category 6? Or is the guy running Category 6 at the mercy of the woman running Category 17?

It's a nuanced situation, isn't it?

Just like weather / climate is a nuanced situation.

Obviously, you need the right data to even consider a reasonable response.

Almost none of you have the data to answer this question.

Selling Chickens

If you think you're performing ecommerce magic, you probably are, but you aren't doing what Sears did in the 1930s and 1940s when they (checks notes) sold live chickens via catalog.

Categorizing Categories

Based on the table below, we can categories our categories into three groups.

Givers.
Neutral.
Takers.

You want Givers/Neutral categories. You want to de-emphasize Takers.

Here's the table.

There are ten obvious Giver categories.

2 / 3 / 6 / 7 / 8 / 9 / 12 / 13 / 14 / 17.

There are a handful of Neutral categories.

0 / 4 / 10 / 14.

There are a few Taker categories.

1 / 5 / 11 / 16 with 11 being an egregious Taker!

The email marketer tells stories about anything not in the Taker category.

The social marketer can do "something" with Taker categories, but not much. If you have to promote Taker categories, do so with channels that have essentially no variable cost ... like oraganic social.

The search marketer adjusts spend based on Giver / Neutral / Taker ... spending more on Giver categories, spending less on Taker categories.

The catalog marketer? You won't do this, but you cannot put Neutral / Taker categories in your catalog. Maybe just the winning items in those categories, that's enough. Those pages are TOO DARN EXPENSIVE to cause harm to your existing customer base. If you disagree (and you will disagree), talk to the paper / printing / postage / boutique agency folks and demand they align the expense structure of the discipline back to 2006 levels.

Alright, I'm off the soapbox.

Next week, we'll put all of this category stuff into perspective.

A Taker

We've all met takers in our personal lives. It's a horrifying experience. A narcissistic politician, an angry mother-in-law, a friend who always wants everything his way. All of it exhausting.

It isn't much different in ecommerce.

You undoubtedly have a merchandise category that is a TAKER. It takes away from other categories. It doesn't develop other categories, instead it borrows customers who then spend less in other categories and more in the TAKER category.

Here's our table. Take a look at category xx.

Read across the category labeled M11.

What a train wreck.

The category only adds value to categories 8/11/12.

The category TAKES value away from a whopping ten (10) categories next year.

This is a category that you shouldn't advertise. You can offer the category if it is profitable, sure (though that is questionable given this table), but your ROAS-obsessed analytics folks are making rampant mistakes giving this category equal treatment in any marketing channel. Catalog marketers should not feature this category anywhere ... bury it in the back of the catalog! Only feature winners here. Email marketers needs to not feature these items ... make it easy for the customer to "get" to the items online but most certainly don't feature this stuff ... those items are TAKERS harming your business long-term.

If you are going to advertise TAKER categories, do so on Organic Social. It costs you nothing. Do not waste precious marketing dollars on categories that don't play well with other categories.

A Giver

What you really want is to offer merchandise that causes customers to want to buy EVERYTHING you sell. That's how you know you're doing a good job.

Remember our table from yesterday? This brand has seventeen (17) merchandise categories. Reading across the rows, we see how much money a customer will spend next year in each category (columns) based on $1 spent in the category (row) previously. I weight historical dollars (100% past year, 60% 1-2 years ago, 35% 2-3 years ago, 20% 3-4 years ago) to add a "recency" component to historical purchases.

Here's the table.

Read across the row labeled "M09". This is Merchandise Category 9. This category is a GIVER. When a customer buys from this category, the customer spends more money next year EVERYWHERE!.

$0.05 in Category 1.
$0.21 in Category 2.
$0.37 in Category 3.
$0.23 in Category 4.
$0.20 in Category 5.
$0.22 in Category 6.
$0.24 in Category 7.
$0.11 in Category 8.
$0.20 in Category 9.
$0.29 in Category 10.
$0.48 in Category 11.
$0.33 in Category 12.
$0.18 in Category 13.
$0.33 in Category 14.
$0.10 in Category 15.
$0.26 in Category 16.
$0.22 in Category 17.

This category is a GIVER. It delivers added value to every category. It either causes customers to "need" what is offered in other categories (i.e. an iPhone buyer purchasing an Apple Watch) or adds a halo to the brand experience.

This is the category that gets extra attention on your home page, in your email marketing campaigns, in social. You pay extra in Google Ads for items in this category. Terms like ROAS have no real meaning, because ROAS is what average marketers use to judge success ... you have a category that adds value to every category in the future ... you're willing to pay more to have more success.

And for the catalog marketers in the audience? Your catalog pages are too expensive now, you can't afford to feature stuff like you used to. Feature the categories that cause customers to buy everything in the future. Use your catalog dollars in a smart manner, ok? (and yes, I realize that won't happen but it should happen).