Kevin Hillstrom: MineThatData

Case Study: A Marketing Problem is Identified

We're busy analyzing "Beans ... The Internet's Only Variety Store!". Yesterday I ended the post with reference to a table from my Elite Program runs, called the "Comp Segment" table. There are two sub-tables within the table ... one analyzing new/reactivated buyers by month ... one analyzing how customers with exactly two purchases in the past year spend money in the following month. Here we go.

The top portion of the table shows us three years of comp new/reactivated buyer counts. If a customer purchases in a month, and I don't see a purchase from the customer for the twelve prior months, the customer is "new/reactivated".

Tell me what you see in the top half of the table?

Here's what I see. In 2024, comp new/reactivated buyers increased by 0.5% on an annual basis. In 2025, comp new/reactivated buyers decreased by 22.7%. This is likely a marketing problem. New/Reactivated buyers come from well-executed marketing tactics. We see modest decreases in November/December 2024, then consistent decreases through most of 2025.

Pay attention to November/December counts. What do you observe?

Counts are very high in November/December.

In other words, this business is disproportionately skewed to November/December, which means that "The Internet's Only Variety Store!" is really a Christmas-based Gift-Giving store. All of the risk of the year is back-loaded into November/December. It is VERY HARD to run a November/December centric business.

Look at November 2025 ... counts are down 42.7%. This is a MARKETING problem. It likely means the CFO told the CMO to not spend money, without understanding how the business actually works. I could be wrong ... but that's the pattern I repeatedly see when I analyze Elite Program clients. When a business has an annual rebuy rate < 40%, the business must focus the majority of all marketing efforts on new/reactivated buyers ... and when the business fails to achieve this mission, the business struggles.

The bottom half of the table shows how much more/less customers who purchased exactly two times in the year spent in the following month. If we have five customers, four do not purchase ($0) and one does ($40), the average is $8.00 per customer. That's what is being reported above.

What do you observe in the table?

From April 2024 to August 2025, comparable customers spent less, consistently, vs. the prior year. This "can" happen when marketing dollars are reduced. This is "more likely" to happen when something changes with the merchandising strategy of a company.

Over the past two years, comparable customers are spending 22.9% less than they previously spent. That's unsustainable.

The two tables within a table tell a story.

It is obvious that "somebody" decided that marketing should spend less money in 2025.
It is obvious that there is a merchandising problem causing customers to spend less year-over-year.

Footnote: When I woke up on Monday morning, I was greeted with three unsubs. All three were industry consultants (not people actually working for a brand like "Beans"). Two of the three were merchandising consultants. In other words, it was clear that I offended them. The same thing happened back in 2014 when I introduced Merchandise Forensics ... the most popular analysis project I ever sold. It is fashionable and easy to blame marketers for missteps. It causes uncomfortable feelings when you suggest that a merchant isn't perfect, but instead, is just like the rest of us.

Case Study: Interacting Challenges

Category performance comes later ... early in a project, I look for high level trends that help guide me to solutions.

I liberally borrow analytics from my Elite Program runs. The rolling twelve-month view of customer behavior is always illuminating.

There's all sorts of goodies in the table.

Customer counts fall away during 2025.

Annual orders per buyer have declined from about 1.17 per year to 1.14 per year ... in other words, customers are only buying once per year. More on that in a future post.

Items per order collapsed in November 2025.

Price per item purchased accelerated in November 2025. It is usually not a coincidence when items per order decrease and price per item purchased increases ... customers have a budget.

AOV decreased for years, but in the last three months of the year AOV increased at a healthy rate.

I mentioned yesterday the problem with new items. This table clearly illustrates the challenge, and the challenge is a two-year challenge that accelerated late in 2025.

The last two columns illustrate trends among items selling at/above their historical average price point (i.e. a $30 item selling for $30) and items selling below their historical average price point (i.e. a $30 item selling for $20). In the past year, items selling below their historical average price point declined faster than items selling at/above their historical average price point. Though demand collapsed, demand collapsed faster among items selling below their historical average price point. The merchandising team was trying to sell at full price. Good job!

Another important Elite Program table is the Comp Segment table. Here is the table.

There are many tidbits here that shape where the analysis is headed. I'll explain those tomorrow. Interestingly ... there IS a marketing challenge that the table above identifies. Can you see it?

So far ...

A clear merchandising problem with new items.
Customers only purchase one time per year.
Customer counts are decreasing throughout all of 2025.
Prices increased late in 2025.
More full-priced selling in 2025.

Case Study: First Steps

When you work on a new consulting project, you don't ask silly questions like "Why is your company named 'Beans'?" You wonder, you ponder, but you don't ask.

Instead you do two things.

You listen.
You run client data through the Elite Program Code.

Remember, Beans ... The Internet's Only Variety Store ... was a $31 million dollar business that is now a $20 million dollar business. That's an unacceptable drop, and a drop of that nature is generally self-inflicted. Businesses can collapse on their own, they can collapse because they are mismanaged, and they can collapse for both reasons. Typically ... somebody mismanaged something or didn't understand something.

Somebody at Beans mismanaged something or didn't understand something.

I like to run new data through my Elite Program framework. I want to see how a client compares to all other clients. Within a client, there are many points of view ... there are fables, there are stories employees tell themselves. There are also facts about actual customer behavior and actual merchandise categories.

Let's look at a simple comparison in the data. Depicted below are two graphs, one for annual demand from new items, one for annual demand from existing items. I elected to graph each metric with the same scale on the y-axis because doing so is illustrative of the challenges Beans faces.

What do you observe?

Existing item demand takes a hit in the last months of 2025, sure. Now look at the trajectory of new item annual demand. From fourth quarter 2024 through fourth quarter 2025, we have a catastrophe.

I'm literally 30 minutes into the project, and I already have a potential culprit - it's the Chief Merchandising Officer. In my mind, I prepare myself for the fact that future discussions are not going to go well. If this was a marketing mistake, both graphs would decline at the same rate. If this is a merchandising mistake (and it most certainly is), one graph doesn't look too bad while the other graph looks ... bad!

Yup, we're off to a good start.

I'll typically send a tidbit to the project sponsor and anybody the project sponsor wants me to include. I'll clearly communicate that my first view of the data suggests a merchandising challenge, not a marketing challenge. Then I'll duck, because somebody is likely to be angry with my assessment.

P.S.: If you like the comparison above, consider becoming part of my Elite Program, with runs in June / October / February of every year. Cost is $1,800 for the first run, then $1,000 each thereafter. June is just around the corner, friends!

Case Study

Let's spend some time talking about a case study. I will analyze a business called "Beans". The data is real, the company and scenario is fictional.

The business is owned by Paisley Ingram. She calls Beans "The Internet's Only Variety Store". If you ask Paisley, she'll tell you ... "I sell whatever I want to sell. Always have. My customers love me."

Annual Demand/Sales for the past five years.

Demand 4 Years Ago: $31.0 million.
Demand 3 Years Ago: $28.4 million.
Demand 2 Years Ago: $27.4 million.
Demand 1 Year Ago: $25.6 million.
Demand as of 1/1/2026: $20.8 million.

Woo boy.

Let's dive in and see what a study of merchandise and categories can tell us about the problems at The World's Only Variety Store.

Your Mission

Years ago I worked with a CEO who had a Mission for his team.

He firmly believed in retaining customers, in loyalty programs. He believed that he could "force" a customer to become loyal. Seriously. He talked about it. Just get the right offer in front of the right customer at the right time. Throw some points at the customer. Let the customer redeem points when he told the customer it was the right time to redeem points. "It works for BJ's Restaurant and Brewhouse!" Good for him ... free pizookies all around.

The Mission? "We're going to thrive by creating a large number of highly loyal customers, and we have the tools to do this outside of the merchandise we sell."

I mostly disagree with the guy, but he had a clear Mission for his team.

(IDEAS) * (TALENT + LEADERSHP + STORYTELLING) = MISSIONS

His team moved in lock step with his Mission. The guy did his job as a Leader. His Team was talented. His Leadership was good. His ideas were poor. His storytelling was non-existent. This meant his Mission, while clearly articulated, had minimal value.

He needed new customers.

It is possible to have a good Mission but a bad outcome.