Kevin Hillstrom: MineThatData

At 10,000 Feet?

We established that there are four categories that the Merchandising Team believes in ... and seven categories that are being purposely contracted.

Does a contracted category impact customer response?

Yes.

Let's look at how many customers bought from the four primary categories and all other categories (both) in the past four years.

3 Years Ago: 180,286.
2 Years Ago: 171,306.
1 Year Ago: 169,072.
Today: 110,850.

Yeah, that's a problem.

How many customers purchased just from the four primary categories?

3 Years Ago: 168,952.
2 Years Ago: 156,288.
1 Year Ago: 174,994.
Today: 184,286.

Counts are holding mostly steady.

How many customers purchased just from the other seven categories?

3 Years Ago: 217,656.
2 Years Ago: 243,028.
1 Year Ago: 218,436.
Today: 130,866.

Yeah, counts are crumbling.

Which customer do you want? Here is the amount of spend each customer segment generated in the past year based on how the customer was segmented a year prior.

Both Groups = $11.46.
Primary Categories Only = $6.44.
Secondary Categories Only = $4.60.

In other words, when you lose customers from Secondary Categories, you aren't losing highly productive customers. When you lose customers from Secondary Categories who also buy from the four Primary Categories, you lose $11.46 per customer, which is too much.

Worse, when a Primary+Secondary customer becomes a Primary-Only customer, value drops from $11.46 to $6.44. That's a stunning drop.

As the Merchandising Team at "Beans: The Internet's Only Variety Store" contracts the majority of categories, the Merchandising Team weakens the customer file - shifting customers from high-value segments to low-value segments. No Bueno! They are weakening the customer file, which lowers demand in all categories next year, which requires more of a marketing investment in new customers, which harms the p&l (in the short term).

So yes, before you start messing around with your merchandise assortment, think about what you are messing with! If you are lowering the value of customers for no good reason, well, that means you are harming the business.

The 15,000 Foot View

Yesterday we took a step back, looking at the business from a 30,000 foot view. We clearly observed that there was a new merchandise problem over the past two years.

I drilled down one level ... there are four categories that the Merchandising Team appear to have faith in.

Apparel Tops / Apparel Bottoms / Outside / Decorations were less than half the business three years ago ... today they're more than 60% of the business. Demand isn't necessarily stable, but it isn't bad. Existing items are stable, new items are down about $900,000 from two years ago.

Everything else?

Wow.

New items are down 45% from just two years ago. New item demand is down a whopping 70% from two years ago!!!!! In the past year, existing item demand is down a million dollars.

Tomorrow we'll apply this 15,000 foot level view of the business to customer response. It's obvious the Merchandising Team is doing something on purpose ... are customers impacted (hint - yes, they're impacted).

Click here for more information and file layouts.

The 30,000 Foot Level

It's common to run across a Marketing Team that is persecuted. "They don't know what they're doing!!" And yes, there are a ton of marketing teams that have absolutely no clue what they're doing ... that's life.

But there are marketing teams that are being persecuted ... they're blamed for problems that they have nothing to do with. Maybe the website has a 2.8% conversion rate and it was 3.3% last year ... "the marketing team is sending us terrible traffic". The Marketing Team blames Google & Facebook. Everybody is pointing at somebody.

Somebody should be pointing at the data. Here's the high level view of Beans ... at 30,000 feet, it is easy to see this isn't a marketing problem.

The business was at $27.4 million two years ago ... it is at $20.8 million today. Yes, there is a problem.

A simple cut of the data by new item sales vs. existing item sales shows us there is a merchandising problem.

New Items: Decreased by $6.5 million.
Existing Items: Decreased by $0.1 million.

Marketing problems result in equal declines.

Merchandising problems result in uneven and chaotic outcomes.

Here's a worse problem ... if you have a new merchandise problem today, you'll have an existing merchandise problem tomorrow. New items become existing items, and if you are down 40% in new items demand you'll be down a bunch in existing item demand in a year or two.

One simple table ... the one illustrated above ... it's not hard to run. If you are in marketing you need to know what the table illustrates. You need to know "what" specific problem you are solving, and it's frequently not the problem others in your company think you need to solve.

P.S.: Click here for information about your own "Categories" analysis.

Another Way To Depict The Challenge

If a business discontinues too many existing items (or fails to introduce sufficient quantities of new items), we should be able to see the problem manifest via sales declines. Right?

Here is a scatter plot of number of items sold each year (>= $1,000) on the x-axis ... the y-axis shows us annual sales per item.

The three outliers on the bottom left of the chart? Those represent the past three months when this brand also decided to not spend money acquiring new customers, compounding their own challenges.

Otherwise, there's a clear relationship. Every additional item offered yields an ever-decreasing rate of demand/sales.

Is this good or bad? It almost looks bad.

Let's perform a multiplication.

"Beans" generates $9,800 per item when it sells 2,700 items = $26.5 million.
"Beans" generates $12,750 per item when it sells 1,900 items = $24.2 million.

In other words, fewer items = less demand/sales.

Pair it with a customer acquisition catastrophe in the past three months (largely caused by reductions in the marketing budget) and "Beans" has problems.

A group of Professionals appear to have harmed the business, all independently, on their own.

It will take a team of Professionals, working together, to rebuild "Beans".

Ready for your own customized "Categories" project? Contact me now (kevinh@minethatdata.com) ... click here for additional information.

A Problem, Presented Visually

The following bar chart clearly illustrates a problem.

I depicted the number of styles (i.e. t-shirt, not white t-shirt, not XL white t-shirt, just the style level) with annual sales >= $1,000.

There are two dates that are problematic in the image.

End of December 2023.
End of December 2024.

Because each date represents twelve-months of purchase history, it likely means that many items were discontinued at the End of December 2022 and the End of December 2023. It could also mean that new items weren't introduced right after that whereas they were introduced at comparable times the year prior.

This is such an easy query to run.

So run it ... does it say anything about why your business is thriving or struggling?

Contact me for your own Categories analysis (click here). Blog subscribers and prior clients save $$$.