Kevin Hillstrom: MineThatData

Why Does MRV Matter?

I've told the story before, but 32 years ago at Lands' End I was invited to a Catalog Review meeting. On the walls of the room were each spread in the catalog, colored GOLD / GREEN / BLUE / RED based on the profitability of the spread (a spread is a two-page combination in the catalog ... like pages 14-15, for instance).

The story of the catalog became apparent the moment you walked into the room ... you could see that the first twenty pages of the catalog STUNK and consequently the catalog didn't perform. You could see that somebody decided to put Home merchandise on pages 16-23, and woo-boy did that kill the momentum of the catalog (including killing the items on pages 24-25).

In other words, you learned why you business didn't perform to expectations. Yes, what you sold mattered. HOW you chose to present it to the customer also mattered ... a +/- 15% impact ... which is the difference between being profitable and not being profitable.

A great sadness of the rest of my career is the complete lack of attention to detail that followed by the companies I either worked with subsequently or observed. As the late Lori Liddle once said, "there's money just lying on the floor, waiting for somebody to pick it up, why aren't you picking it up?" Good question!

The emergence of e-commerce is largely responsible for the lack of attention to detail.

I know, here come the unsubs.

You know that 1,438 visits came from Pinterest. You know nothing else about those visits (oh, I get it, you know that the conversion rate was 1.33% and the AOV was $171). You generally don't know what those customers bought, and you most certainly don't know if those customers bought items that caused the customers to become more loyal in the future.

This is why MRV (Merchandise Residual Value) is so important. It's a modern day proxy for the value of an item. Would you rather have an item that generates $500,000 and causes customers to be 24% more loyal in the future or an item that generates $600,000 and causes customers to be 40% less loyal in the future? You'd prefer the former, not the latter.

Say you are Macy's and you feature this dress in an email campaign (as they did last week).

Do you think that Macy's has any idea whatsoever if that dress increases customer loyalty among the customers who purchase the dress?

You already know the answer.

Now, it's perfectly fine to advertise/feature items that don't have great MRV, especially if they generate sales today. Every item matters. But it is far smarter to give "more real estate" to items that cause customers to become more loyal, right? It just makes common sense.

Common sense.

A Simple X1 / X2 / Y Model Gives Us MRV

As mentioned yesterday, MRV (Merchandise Residual Value) is a function of two independent variables and one dependent variable (yes, I'm about to tell you how to calculate the metric yourself ... something your favorite agency absolutely will not do).

Analysis Period: All items sold in the past year.

First Independent Variable: Time ... the average amount of time that has passed (months) since the average customer bought that item. A value of 1.4 means the average customer purchased that item 1.4 months ago.

Second Independent Variable: Order ... the average order frequency for the customer buying the item. If the customer places her fifth life-to-date order buying that item, the value is five (5). A secondary benefit to calculating this metric is that not only do you get to derive MRV, you also get to know if your best customers buy an item or if new customers buy a specific item. You'd want to know the items that attract new customers, right?

Dependent Variable: Rebuy ... do customers repurchase after buying the item being analyzed?

Select all items generating at least "$x" in the past year ... in the example I'm using, it's at least $10,000 sold in the past year.

Regression: Time and Order variables used to predict Rebuy.

MRV: Rebuy / Predicted Rebuy. You could also use Rebuy - Predicted Rebuy, the outcome is the same, I now find the ratio to be more important.

So, there. Go have your analyst run the analysis for you, Go have your analytics agency run the analysis for you. Seriously, what stops you from doing this right now?

Or, contact me right now (kevinh@minethatdata.com) and let's get busy (click here).

MRV

I call the metric "Merchandise Residual Value", and when you hire me for your S-Tier Analysis you get to see how MRV varies at an item level.

What is Merchandise Residual Value (MRV)?

(The Probability of a Customer Who Bought "Item X" Repurchasing Again) / (Expected Probability of a Customer Who Bought "Item X" Repurchasing Again).

It sounds nerdy, sure. But it sure is effective.

In the dataset I'm analyzing (actual data), the brand has a new item. The item has been sold for about 1.364 months, and has been purchased by customers placing (on average) their 2.27th order.

Based on this data (1.364 months, 2.27th order), we "expect" a rebuy rate of just 8.4% after buying this item.
In reality, the rebuy rate is actually a whopping 36.4%.
MRV, or Merchandise Residual Value, is (0.364) / (0.084) = 4.343.

In other words, customers buying this item are 4.343 times more likely to repurchase than customers buying an average item.

Wouldn't you want to know that?

Shouldn't you HAVE to know that?

Every item you sold in the past year has a MRV. You need to give extra marketing love to items with a high MRV.

Your S-Tier Analysis

I've been telling you about this all week ... here is your S-Tier Analysis Project!

What do you get?

I analyze all of your new/reactivated customers in recent years to determine if the "mix" of who you are acquiring is shifting ... shifting in a good way or bad way.

I will score all new customers for the next year, either on a monthly or quarterly basis, so we can understand if you are making progress regarding "who" you acquire.

I will measure the life cycle the customer possesses. Does the customer migrate to different channels as the customer matures? Does the customer purchase different merchandise as the customer becomes loyal? Does the customer ever become loyal? Are there key inflection points early in the customer lifecycle when you should apply Action Streams to convert the customer? Are there seasonal differences that need to be taken into consideration (i.e. Cyber Monday or Spring or Summer)?

I will analyze your merchandise assortment ... identifying items that skew to new/reactivated customers, while measuring the items that lead to high Merchandise Residual Value (MRV) ... items that cause customers to become more loyal. You'll know which portion of your assortment is valuable to acquiring new customers.

I will measure the effectiveness of new customers generated via Marketplaces. This has been a really big weak spot across clients post-COVID.

I will produce a five-year forecast that identifies S/A/B/C/D/F Tier new customers, showing the long-term impact of generating too many D/F Tier new customers.

Given all of the shifts in your business tactics post-COVID (especially shifts into Marketplaces), it's more important than ever to understand if you are making choices that harm your customer base over time. Most important - home grown prospects are the path to success. Growing your Instagram / YouTube / Email subscriber platforms means you don't have to ultimately pay somebody else for a one-and-done customer. One of the biggest problems in my post-COVID analyses are the acquisition of too many one-and-done customers.

Cost: $24,000 ... half due up-front, half due within fifteen days of project completion.

Special Subscriber Opportunity: As always with my subscribers, I give you an opportunity to help break in my code. If you respond by 11:59pm EDT on September 16, you can help break in my code for just $17,000. If you are a prior paying client, it's $15,000!

Contact me right now (kevinh@minethatdata.com) to get busy!

P.S.: Need adaptations to this project for your unique business issue? We'll modify the scope, no worries.

It's Different Today

Not quite sure what AI had in mind as it blended a bunch of stuff in this image.

Maybe a decade ago I visited a retail brand - the CEO was ANGRY - we've all been in the meetings. HIs anger was obviously misplaced. His stores performed admirably! His share of e-commerce volume was low. He didn't like that. At all.

Ten years ago the pundits demanded full integration of all operations. If you sold something in a store, you HAD to sell it online at the same price and the same promotions. What a paralyzing thesis.

We learned that the customer could have cared less.

In 2025, it's different.

We've learned.

You do what is best for the customer, regardless of "channel".

Sometimes there are popular items that you want to make available all the time, in every channel. One problem. The customer who buys those items doesn't come back and buy again. It's a bad way to acquire a customer.

This means you have to have a measurement method to know what items new customers purchase ... items that lead to lousy downstream performance.

This analysis will be part of what I announce to you, shortly. If you want additional details, let me know (kevinh@minethatdata.com).

Kevin Hillstrom: MineThatData

Why Does MRV Matter?

A Simple X1 / X2 / Y Model Gives Us MRV

MRV

Your S-Tier Analysis

It's Different Today

More Recent Articles