Dear Catalog CEOs: Matchbacks

Dear Catalog CEOs,

During the past decade, matchback analytics have become an indispensable part of catalog marketing. Without matchback analytics, it is possible you would not have a catalog channel.

The age of the "matchback" changed our perception of marketing.

Do you remember the good 'ole days, like way back in 1994, before we had an e-commerce website, back in the stone age of catalog marketing?

Back in 1994, we cared a lot about the concept of "cannibalization". We executed a lot of exotic multi-variate tests to detect cannibalization. For instance, say we mailed two catalogs, one on September 1, one on October 1.

• September 1 Catalog = $5.00 per catalog.
• October 1 Catalog = $5.00 per catalog.
• Total Demand = $10.00 per customer.

And then, we got excited! If we could generate $5.00 per catalog, maybe we should add a third catalog. So in 1995, we added a third catalog, on September 15.

• September 1 Catalog = $4.00 per catalog.
• September 15 Catalog = $4.00 per catalog.
• October 1 Catalog = $4.00 per catalog.
• Total Demand = $12.00 per customer.

Remember what we thought? We were happy with the new catalog, but we were concerned with the performance of the two existing catalogs. It was almost like they weren't working well anymore! And in fact, if we ran a profit and loss statement, we found that $12.00 of customer demand across three catalogs was less profitable than $10.00 of customer demand across two catalogs.

We thought about cannibalization, a lot. We were genuinely concerned about how one marketing activity cannibalized another activity.

Then matchback analytics came along. The data enabled the catalog vendor community to change our minds about how we thought about marketing activities.

We stopped thinking about "cannibalization". Heck, these catalogs didn't cannibalize business. Nope, these catalogs "added" business, they "drove" business to other channels.

The industry script (supporting an additive model vs. cannibalization) benefits the catalog ecosystem. The co-ops and database vendors created reporting that illustrated how catalogs drove sales across all channels. Their reporting supported the notion that we should rent more names from the co-ops. In other words, matchback reporting fuels the financial success of the co-op and list industry. The paper industry supports the concept of matchbacks. Printers support the concept of the matchback. The USPS supports the concept of the matchback. Your favorite Catalog Consultancy that helps you with mailing plans benefits from the matchback. Even third-party opt-out services benefit from matchbacks ... without matchbacks, they serve fewer customers who are getting unwanted catalogs.

The industry script benefits the entire catalog ecosystem.

Now let's focus on you, the Catalog CEO. Do you benefit from matchback algorithms?

As we head into the Holiday season, I'd like to ask you to do our industry a favor:

• Randomly sample 5,000 or 10,000 customers from the universe you would mail your best-performing Holiday catalog to.
• DO NOT mail these customers your best Holiday catalog.
• Code these customers as a unique segment, and enter these customers into your matchback routine with your favorite matchback vendor. Remember, these customers were not mailed a catalog, so your matchback vendor should show that no orders are matched back to the catalog that you did not mail.

If your matchback vendor matches online orders back to a catalog that was not mailed, then you have an estimate for how much your matchback vendor is over-stating the results of your catalog mailings.

Catalog CEOs, this is a very important topic. If your matchback vendor is over-stating your catalog performance because your vendor fails to take cannibalization into account, then you are over-mailing your customer base, and in all likelihood, you are wasting marketing dollars, squandering profit.

Increasingly, I am hearing of big discrepancies between matchback results and real-world results obtained via holdout tests. One company told me that every phone order was paired with one online order matched-back in their matchback algorithm. And yet, when they executed a holdout group, they only saw a 5% drop in total demand --- almost no phone demand or online demand was lost when the catalog was not mailed.

In other words, cannibalization was so significant that the catalog was basically adding no incremental demand. This is an important concept --- cannibalization testing shows no additional demand, while matchback algorithms show that catalogs drive online business, forcing you to mail more catalogs.

Do you understand the distinction?

Catalog CEO's, please ask your marketing folks to give this test a try. The entire catalog industry ecosystem benefits from matchback algorithms, and they aren't supporting matchback to be evil ... it's the best available reporting folks have. I'm asking you to question your results, to execute a test and validate that the matchback algorithms are giving you honest results. I don't benefit from doing this test, I have no financial interest in positive or negative results. Only you will benefit if you find that orders are being mistakenly attributed to catalog mailings.

So see for yourself! Run a holdout test, code the customers as a segment, run them through your matchback algorithm, and see if there is a bias that is causing you to over-state your results.

Thank you for your consideration,

Kevin Hillstrom, President, MineThatData

The Most Important Catalog Marketing Metric: Organic Percentage

The most important metric in catalog marketing is called the "organic percentage".

The metric is defined as the percentage of demand, at a segment level, that would occur if no catalog mailings were delivered to a customer.

Most of the catalogers I speak with assume that the organic percentage is zero --- in other words, if catalogs were not mailed to a customer segment, the segment would not spend any money.

Of course, this assumption is false, perpetrated by biased matchback algorithms that incorrectly assign online orders to catalogs mailed to the customer, when in reality, the catalog had nothing to do with the generation of the order in question. You'll know that your matchback results are biased if the percentage of demand you add on to your acquisition segments (after matchback) is significantly lower than it is for housefile customers.

Catalogers who attempt contact strategy tests, say over a three month period of time, find relationships like this.

• Telephone - Only customers have an organic percentage around 10%.
• Telephone + Online customers have an organic percentage around 25%.
• Online - Only customers have an organic percentage around 40%.

In other words, if no catalogs are mailed to an online-only customer, the online customer will still spend 40% of the demand they would spend if they are mailed all of the catalogs during the quarter.

The organic percentage metric is critical, because it dramatically impacts your calculation of profit and loss. If you have a high organic percentage, then you are significantly overmailing customers, regardless of what your matchback analytics vendor tells you. If you have a low organic percentage, then you have no choice but to mail catalogs in order to generate demand.

The image at the beginning of this post shows the difference in profitability for the same segment of customers, comparing a 10% organic percentage to a 40% organic percentage. The ten percent level requires four mailings per quarter. The forty percent level maximizes profit at just one mailing per quarter. Think about what you could do with the expense from the three additional mailings?

If there were just one metric I'd ask catalogers to track at a segment level, during 2009, it would be the "organic percentage" metric. Knowing this metric fundamentally changes how you decide to contact different customer segments.

How important is this percentage? Take a brand like Nordstrom. This is an $8.5 billion dollar business that is luck to generate ten percent of that total from marketing activities. Therefore, the organic percentage is ninety percent. This brand generates ninety percent of sales without the aid of traditional marketing activities. That's a strong brand.

Think about Zappos. There's the volume they generate due to online marketing and search marketing, and then there's the volume they generate via word of mouth. I'd guess that half of their volume happens without the aid of marketing, plus or minus twenty percent.

And then think about a traditional cataloger. The traditional cataloger believes that the vast majority of demand happens becaue of catalog mailings. If mail/holdout tests validate this, then the cataloger is at the mercy of catalog marketing --- if customers are no longer responsive to this form of marketing, demand dries up.

The goal, of course, is to build a brand that has a high organic percentage, not needing advertising to drive sales and profit.

We can learn how much of customer demand is generated by advertising by executing thorough mail / holdout tests, in both catalog marketing and e-mail marketing.

Matchback Bias

You're probably partnering with your list organization, data warehouse vendor, or co-op on the never-ending scope of matchback analytics.

The goal, of course, is to prove that catalog marketing is a vital piece of the modern marketing puzzle. You're trying to truly understand the ROI of this activity. That's why you try so hard to attribute every online order back to one of the dozens of catalogs you mailed in the past year.

Now let me ask you this.

Do you go through the same effort to attribute every phone order back to the original online source?

You don't?

I met with a business that is doing just that. They combine their web analytics tool and their matchback analytics platform to attribute phone orders back to the online marketing activity (which is usually organic/natural search) responsible for driving the phone order.

Why is it that our industry is so bent on proving that catalog marketing drives online orders, but doesn't invest the energy to prove that online marketing drives phone (and store) orders?

Our view of the world is biased, folks. And that bias favors co-ops, printers, the USPS, the paper industry, and the list rental/exchange industry.

Your thoughts?

Who Benefits From Flawed Matchback Analyses?

This is my final discussion about matchback analyses for awhile, as I'm sure many of you are ready to read about other topics. But I got chewed on, I was told to stop talking and get on the multichannel bandwagon. That bandwagon may be financially hurting some catalogers. Somebody needs to talk about that.

Let's think about the industries that benefit from incorrectly executed matchback analyses. Mind you, I'm not picking on any specific individual --- I've observed many folks in these industries who counsel clients in a positive way.


The USPS: Ever wonder why it seems like there are more catalogs in your mailbox these days, compared with a decade ago? Here's a secret ... if you mail every one of your internet buyers a catalog, a matchback analysis might tell you that the catalogs drove all online orders within twelve weeks of the catalog mailing ... even if search and e-mail marketing and organic demand were truly responsible for the orders. The USPS (and now the good folks in the UK as well) commission studies that "prove" that catalog mailings drive online orders. I'm not saying catalog mailings don't drive online sales --- I'm just saying we are significantly overstating the importance of catalog mailings via flawed matchback algorithms.

The Co-Ops: Catalogers love co-ops these days. Catalogers get names that perform better, and get them at a lower cost than via list vendors. So co-ops have a financial incentive to promote flawed matchback algorithms (though some truly try their hardest to do a good job). By "proving" that catalogs drive online orders, catalog clients order more names from the co-op, driving co-op sales and profit. An even bigger conflict of interest occurs when co-ops actually execute the matchback algorithm for the client.

Merge-Purge Houses: The cataloger gets matchback results from the co-op, orders more names, names that are merged at the merge-purge house, driving increased sales and profit for the merge-purge house. Also, many merge-purge houses run matchback analyses for catalogers, earning $$$ for their efforts.

Printers: If catalogs are "proven" to drive 70% to 80% of online sales (which does occasionally happen, but not as often as we're being told), then printers benefit, too. The cataloger mails more catalogs than they normally would, which drives sales and profit for the printer. If the printer delivers catalogs deep into the mail system, then the printer can earn more $$$ too.

Paper Industry: Some of my feistier conversations have been with folks in the paper industry. More catalogs means more paper, which means more $$$ for those in the paper industry.

List Industry: I'm much less critical of the list industry, because by and large, these folks acted with integrity for the past decade, recommending that clients shift names from lists to the co-op industry, knowing all-too-well that it would result in the death of the list industry. But flawed matchback analyses help those in the list industry as much as they help the co-ops.

Trade Journals: We read about multichannel marketing and matchback analyses in trade journals. These publications depend upon the vendor community for advertising revenue. The vendor community depends upon the trade journal to "get the word out". This symbiotic relationship benefits from promotion of matchback analyses that may not accurately reflect the "truth".


So, let's look at the ecosystem that depends upon matchback analyses that are sometimes flawed.

Co-ops and merge-purge vendors do the matchback analysis, attributing too many online orders to the catalog channel. This causes the cataloger to order more names from co-ops and list vendors than they should, financially helping co-ops and list vendors. These names go into the merge-purge process, financially helping merge-purge vendors. Next, the names go to the printer. Paper reps financially benefit from over-mailing, as do printers. The printer delivers the catalogs deep into the mail system, where the USPS benefits by delivering too many catalogs to customers. Then trade journals tell us all about multichannel customer behavior, funded by the profits the vendor community get from matchback analyses.

It looks to me like the entire catalog ecosystem benefits from flawed multichannel matchback analyses. The only parties who don't benefit are customers, who may not want the catalogs, and catalogers who over-mail catalogs, causing harm to the profit and loss statement.

This is why I've been told to stop talking, to "get on the multichannel bandwagon". This is why I try hard to freely share information with catalogers and multichannel retailers.

Three Types Of Catalog Buyers, And Profitability

Catalogers face big challenges when evaluating the profit of catalog mailings. Given that matchback analyses have long over-stated profitability (at the benefit of the vendors providing these analyses, or the list/co-op industry, folks who depend upon inflated catalog results for improved financial success), we've trained a generation of catalog and online marketing experts to evaluate catalog profitability in a suspect manner.

Some catalogers are studying profitability by evaluating quarterly contact strategy tests. These catalogers purposely choose to not mail segments of customers for three months at a time. At the end of the test period, the difference in performance between the mailed and control group is evaluated.

There are three types of catalog buyers that are frequently evaluated.

First, let's evaluate customers who only shop via telephone. These customers are the easiest to measure, because they seldom buy online, meaning our old-school analytical techniques are still effective.

Quarterly Test Results
Audience = Telephone - Only Buyers
	Phone	Online	Stores	Total
Mailed Group	$15.00	$2.00	$2.00	$19.00
Not Mailed Group	$0.00	$1.00	$1.50	$2.50
Increment	$15.00	$1.00	$0.50	$16.50
Demand			$16.50
Net Sales	80.0%		$13.20
Gross Margin	50.0%		$6.60
Less Book Cost			$3.00
Less Pick/Pack/Ship	11.0%		$1.45
Variable Profit			$2.15

This analysis is straightforward. The mailing strategy generated $16.50 demand and $2.15 profit per customer. Matchback analyses are typically accurate among this audience, due to limited spend in the online or retail channels. As long as online/retail spend is minimal, matchback analyses are accurate.


The second segment of customers provide more of a challenge. In the past twelve months, these customers shopped via telephone, and shopped via the internet. Here is what the results can look like within this audience/segment.

Quarterly Test Results
Audience = Telephone + Online Buyers
	Phone	Online	Stores	Total
Mailed Group	$7.00	$8.00	$2.00	$17.00
Not Mailed Group	$0.00	$4.00	$1.50	$5.50
Increment	$7.00	$4.00	$0.50	$11.50
Demand			$11.50
Net Sales	80.0%		$9.20
Gross Margin	50.0%		$4.60
Less Book Cost			$3.00
Less Pick/Pack/Ship	11.0%		$1.01
Variable Profit			$0.59

This is where matchback algorithms begin to fail. The matchback algorithm will take credit for all $8.00 per customer spent online, allocating that revenue to the catalogs that were mailed. Mail/holdout tests tell us the true story, however. Had catalogs not been mailed, $4.00 would have happened online anyway.

Your matchback vendor tells you that you got $7.00 over the phone, and $8.00 online, so all is good! In reality, you got $7.00 over the phone, and $4.00 online --- profit isn't nearly as good.


The third audience includes customers who only shop online. Multichannel pundits strongly believe that catalog mailings drive these customers online. Here's what one might observe, after a quarterly contact strategy test.

Quarterly Test Results
Audience = Online - Only Buyers
	Phone	Online	Stores	Total
Mailed Group	$2.00	$13.00	$2.00	$17.00
Not Mailed Group	$0.00	$9.00	$1.50	$10.50
Increment	$2.00	$4.00	$0.50	$6.50
Demand			$6.50
Net Sales	80.0%		$5.20
Gross Margin	50.0%		$2.60
Less Book Cost			$3.00
Less Pick/Pack/Ship	11.0%		$0.57
Variable Profit			($0.97)

This audience is treated incorrectly by matchback algorithms. Your matchback vendor will tell you that you got $2.00 via the phone, and $13.00 online, yielding $15.00 total. Your matchback vendor will tell you that this is good!!

However, your mail/holdout test results tell you something different. Had you not mailed catalogs, you still would have gotten $9.00 of the $13.00 online. Therefore, when you run the incremental profitability calculation, you find that catalog marketing is unprofitable in this audience.

The reality is that natural search, paid search, e-mail marketing, affiliate marketing, portal advertising, shopping comparison marketing, word-of-mouth, and brand recognition all contribute to the $9.00 of volume you achieve if you don't mail catalogs to this customer.


This type of analysis is sorely missing in modern catalog planning. Some matchback vendors understand these issues, and genuinely try to help us. Sometimes, the thought leadership simply isn't there --- and it is costing catalog marketers millions of dollars of profit.

My level of frustration on this topic continues to grow. Recently, I was told by a vendor-based industry leader to stop talking, and "get on the multichannel bandwagon".

I have no problem with multichannel marketing. I do have problems with industry leaders that mislead (maybe not purposely) catalogers in a way that harms catalogers, but helps the very vendor industry that depends upon catalogers for success.

United Kingdom: Catalog Driving 70-80% Of Online Orders?

Jim Fulton points us to this article, authored by the Catalogue Exchange in the UK. The article suggests that catalogers, using traditional matchback algorithms, are driving 70% to 80% of their online orders via catalog mailings.

The UK online market is a lot like the US online market five years ago. We read a lot of these articles in the US five years ago.

Early in the maturity of the online channel, catalogs are going to be the primary driver of online volume. Most likely, the study is directionally accurate.

Things get very interesting once the customer is trained to shop online, and decides she no longer needs the catalog to shop online. For some brands, this change in behavior never happens. For others, this happens very quickly. I lived through this change at Nordstrom --- folks with a traditional catalog mindset, very bright folks, don't understand the phenomenon when I describe it to them.

If you're reading this in the UK, here's a checklist for you:

• This study was published by a group of individuals with a vested interest in promoting catalog marketing. Just keep that fact in mind.

• Multichannel Forensics represent another way of understanding the long-term impact of catalog advertising, across all channels.
  

• Matchback analyses are highly biased. Want to learn how biased they are?
  
• Identify who you were going to mail your next catalog to.
  
• Take a sample of the audience.
  
• Split that sample in half.
  
• One half receives your next catalog.
  
• One half does not receive the next catalog.
• After twelve weeks, do a matchback analysis on the sample that received the catalog.
• After twelve weeks, do a matchback analysis on the sample that did not receive a catalog --- just don't tell your matchback vendor that they didn't receive the catalog.
• In this group, zero (0) customers ordered online because of the catalog --- they didn't receive the catalog!! If you matchback vendor tells you that online orders happened because the catalog was mailed, you've identified the magnitude of your matchback bias.
• Subtract the difference between matchback results in the mailed and not-mailed group. This is the true number of orders your catalog drove to the online channel.

Your Questions About Multichannel Forensics, Matchback Analyses, and Web Analytics

Tell people you have a book coming out, and you get a few questions.


Question: How are Multichannel Forensics different than Web Analytics?

• Multichannel Forensics and Web Analytics are very different disciplines. Web Analytics looks "back" in time, and tells you what customers did during a specific visit to a website. Multichannel Forensics look back in time to understand online customer behavior, then forecast over the next five years how online, retail and catalog sales will evolve. Notice that the methodology views online customer behavior, then illustrates what happens in all channels in the next five years.
  

Question: How are Multichannel Forensics different than Catalog Matchback Analyses?

• Matchback Analyses are very trendy these days, with dozens of catalog and analytical vendors vying for the right to explain to you how e-mail marketing, online marketing, and catalog/direct marketing work together to cause purchases to happen. Matchback Analyses typically "guess" at what caused customers to purchase, using business rules. Matchback Analyses only look back in time. Multichannel Forensics seldom try to quantify the impact of an individual e-mail marketing campaign, online marketing campaign, or catalog campaign. Instead, Multichannel Forensics quantify the annual impact of an e-mail marketing program, online marketing program, or catalog marketing program on telephone, online and retail channels. This is a fundamental and important difference. Multichannel Forensics is a strategic tool used by CEOs/EVPs to determine the best future strategy for various marketing platforms. Matchback Analyses are tactical tools that can be very beneficial for determining how to allocate a budget.

Question: What is the best use of Multichannel Forensics?

• There are four areas that I really like applying the technology.

• Understanding the long term impact of eliminating a product, brand or channel.

• Measuring the long-term impact of various forms of advertising.

• Identifying the reason why one specific channel is not growing the way management expects the channel to grow.

• Graphically illustrating how customers interact with advertising, products, brands and channels interact with each other.

New Book: Hillstrom's Multichannel Forensics

Well, the cat is out of the bag!

This morning, I received an e-mail from Amazon.

The e-mail said because I previously purchased a book from a different database marketing author, I might be interested in a book being released on September 30, called "Hillstrom's Multichannel Forensics".

I am interested in the book. I AM THE AUTHOR OF THE BOOK!!!!

I'm guessing I'm not the only person Amazon sent this e-mail to. So it is time to share with you what this new book is all about.

The text is the culmination of twelve years of research into how customers behave in a multichannel environment. Twelve years ago, e-commerce began to impact, then influence, then usurp catalog marketing. Peers at competing retailers and catalogers shared their frustration with me about understanding customer behavior in a multichannel environment.

Vendors and Research Organizations shared multichannel customer facts and figures that were impressive. They seldom told us meaningful information that helped us understand our customers. And if they had meaningful information, you can be sure the information would be highly monetized!

I felt compelled to create a methodology, a framework, for understanding and explaining multichannel customer behavior (in a b2b or a b2c environment).

I strongly believed the methodology should do two different things.

First, the methodology had to explain how customers interacted with advertising, products, brands and channels ... in a way that a CEO or Executive Vice President could understand.

Second, the methodology had to illustrate how sales within products, brands or channels will evolve over the next five years. This allows the CEO or EVP to make decisions today that limit future business challenges.

The combination of these factors became "Multichannel Forensics".

The perfect laboratory for testing this methodology came during my tenure at Nordstrom. In 2005, we eliminated our catalog marketing program. I used this framework, this methodology, to illustrate what would happen to online and retail net sales growth if catalogs were no longer there to support those channels.

The methodology worked!

Since beginning my own consulting practice in March, I've completed multichannel forensics projects for thirteen brands/titles. I'm continually pleased with the way CEOs and EVPs react to the methodology. I'm proud of the way the methodology forecasts what is likely to happen in the future. I love giving business leaders tools they can use to quell challenges from the Board of Directors, or Ownership Team.

There are three ways you can learn about multichannel forensics.

• You can read this blog. I will continue to give examples of how the methodology can be used in real-life settings.

• You can read the white paper, which goes into some level of detail on the topic.

• You can buy the book, and learn the nuts and bolts of the methodology. I want for you to be able to do these projects yourself!
  

• Of course, you can hire me to do a project for you.

I spend considerable time in the book outlining three-channel situations (i.e. catalog/online/retail or telemarketing/catalog/online, as examples). Two-channel and three-channel situations are very common, hence the focus on these topics.

I do not go into the complex simulation algorithms that I use to understand the interaction between, say, ten merchandise divisions and three channels. The concepts in the book are illustrated at a level that allows the reader to build the simulations, if desired.

How might you benefit from this book?

• CEOs and EVPs will learn the current trajectory of the business they manage, and will learn how they might mitigate negative trends.
  

• E-Mail Marketers will learn how their oft-criticized discipline builds long-term sales growth.

• Catalog Marketers will precisely learn the value of catalog marketing in a multichannel environment, in a way that matchback analyses cannot possibly explain.

• Online Marketers will finally be able to show retail leadership how online customers become retail customers, demonstrating how comp store sales growth is influenced by the online channel.

• Web Analytics practitioners and Business Intelligence analysts will be able to see how customer behavior can be analyzed longitudinally (over time), providing value that goes well beyond individual session metrics.

• Multichannel Vendors will be able to identify ways they can provide additional value to the clients they support, value that is targeted at the CEO/EVP level.

Amazon is allowing pre-orders of the book in anticipation of a September 30 release date. In the next two weeks, the book will be available on my publisher's website ... www.forbetterbooks.com. Obviously, Amazon is not part of the purchase process if you purchase the book from my publisher's website, Amazon's profit is reallocated among those who invested time, energy and money in creation of the book.

I invite you to learn more about Multichannel Forensics. Do so at no cost by downloading the white paper, or purchase the book! My thanks go to Don Libey for shepherding this book and my first book (Hillstrom's Database Marketing), for giving the little guy a chance!

Coremetrics, Abacus, Experian and Catalog Matchback

Notice the last paragraph of this article about Hanover Direct working with Coremetrics on Catalog Matchback algorithms.

My e-mail inbox and phone have been buzzing lately with queries from online analytics organizations looking to take market share away from established catalog attribution folks like Abacus and Experian.

It will be interesting to watch this trend unfold as executive leadership transitions to those with online marketing experience. These individuals have relationships with online vendors, not catalog vendors.

Explaining The Matchback Mistake

Now that I've frustrated many of you by not aligning with catalog industry best practices (i.e. the right way to implement results from a "matchback" analysis), allow me to explain the philosophical issues surrounding the methodology.

Catalogers like to look at a "segment" of customers, folks with similar behavior, folks with consistent future performance.

For instance, assume it costs a cataloger one dollar to mail a catalog. Also assume that thirty-five percent of all demand flows-through the p&l, resulting in "contribution" or "variable operating profit".

We mail a catalog to this segment of 10,000 customers, folks who last purchased within the past three months, and have spent $250 - $499 in their lifetime with the company.

By measuring source codes, we learn that this segment spent $2.00 per customer over the telephone. We run a profit and loss statement, and observe the following:

Households	10,000
Demand	$20,000
Flow-Through	$7,000
Book Cost	$10,000
Contribution	($3,000)

In other words, we lost money mailing this segment of customers.

This is where the matchback analysis comes in. Savvy catalog marketers partnered with list processing and compiled list vendors to "match" all customers who received a catalog, but ordered online instead, "back" to the catalog mailed to the customer. Typically, the most recent catalog mailed gets credit (and we can address all the flaws with that methodology another day).

In this instance, the "matchback" analysis shows that customers mailed this catalog also spent $2.00 online during the life of this catalog. This changes the profit and loss statement, illustrated below:

Households	10,000
Demand	$40,000
Flow-Through	$14,000
Book Cost	$10,000
Contribution	$4,000

Now all is good in the world! The catalog drove online volume, the profit and loss statement works. Catalog list processing vendors, compiled list vendors, paper vendors, and list rental vendors rejoiced because the catalog becomes a viable marketing vehicle responsible for the majority of the online volume harvested by a business.

This strategy works well when the online channel is incapable of generating its own volume. In 2007, this is often an incorrect and dangerous assumption. This is where mail/holdout testing comes into play.

Simply put, mail/holdout testing shows you how much online volume occurs if a catalog isn't mailed. The methodology points out the fundamental flaw in a matchback analysis.

For many catalogers (certainly not all, maybe not even half), half of the online demand will happen anyway if a catalog is not mailed. In these instances, the mail/holdout testing clearly illustrate this finding (see the last article, business model number three).

In the case of our profit and loss statement, adding in one dollar per customer instead of two dollars per customer changes the profit and loss statement, illustrated below:

Households	10,000
Demand	$30,000
Flow-Through	$10,500
Book Cost	$10,000
Contribution	$500

In this case, the segment is above break-even, so depending upon your profitability criteria, the segment can be mailed next year.

It is this last profit and loss statement that catalogers need to be evaluating.

Almost all catalogers are mailing too many catalogs due to flaws in the implementation of the matchback analysis. This isn't the fault of your list processing or compiled list vendor. It is our fault, we failed to adequately understand customer behavior.

At Nordstrom, when we killed our catalog division, our online division actually continued to grow sales, year-over-year. Matchback analysis suggested that killing the catalog would create a catastrophe. Our inventory management team nearly fainted, thinking the implosion would be epic!

Mail/holdout testing accurately forecast a subtle sales hit that would largely be offset by organic growth in the online channel. Within a month of killing the catalog, we observed that mail/holdout testing was right, that matchback analyses were highly flawed.

Another flaw in the implementation of matchback analysis is attributing online orders to the original source (which in most cases, is catalog).

In other words, the catalog marketer gives the online channel credit for taking the order, but says that the order could never have happened had catalog marketing not been responsible for originally acquiring the customer. This analytical technique assures that catalogs will always gain too much credit --- in these cases, I've seen orders generated by paid or natural search (i.e. Google) attributed to catalogs, because the customer was acquired via a catalog twelve years earlier. I'd stay away from this popular method of attribution.

I realize what I am saying is utter heresy to most in the catalog industry, as evidenced by the feedback I receive from you! As leaders, we have a responsibility to maximize sales and profit in the business models we support. Let's measure the evolution of our business in a fair manner. We need to take our catalog silo hat off, and put our brand hat on. We'll still find that catalogs are an important part of the marketing mix used to educate customers about our merchandise offering.

Expanding Upon Multichannel Business Models

We really lit up the readership meter yesterday when we discussed Multichannel Business Models. Monday was one of the top five traffic days in the history of the blog, that post was the most read post of the day by a wide margin.

I'll take that as affirmation that multichannel business models are of interest to you, the loyal MineThatData reader. Let's expand upon yesterday's discussion.

A common question I hear is "How do I, with the data I have available to me, determine which business model my brand is classified in?" Good question! Let's explore each business model, and some of the things you're likely to see. We'll explore each business model by looking at results from mail/holdout tests, comparing dollar per customer metrics.


Model #1 = Simple Online Presence

Incremental Value Of Catalog Marketing
		Other
	Catalog	Catalog	Online	Retail	Total
	Demand	Demand	Demand	Volume	Volume
Mailed Segment	$3.00	$7.00	$0.25	$0.00	$10.25
Holdout Segment	$0.00	$8.10	$0.05	$0.00	$8.15
Increment	$3.00	($1.10)	$0.20	$0.00	$2.10
Incremental Results:		$2.10 / $3.00 =		70.0%
Matchback Analysis:		$3.00 + $0.25 =		$3.25

Notice that almost no online demand is generated by the mailing of the catalog. Also, if the catalog is not mailed, virtually no online sales occur. This clearly tells you that the website is just "there", customers are not really using it to order merchandise.


Model #2 = Online Order Form: Check out the differences in this table:

Incremental Value Of Catalog Marketing
		Other
	Catalog	Catalog	Online	Retail	Total
	Demand	Demand	Demand	Volume	Volume
Mailed Segment	$2.00	$7.00	$2.00	$0.00	$11.00
Holdout Segment	$0.00	$8.10	$0.10	$0.00	$8.20
Increment	$2.00	($1.10)	$1.90	$0.00	$2.80
Incremental Results:		$2.80 / $2.00 =		140.0%
Matchback Analysis:		$2.00 + $2.00 =		$4.00

Notice how different this table looks. In this business model, demand is driven to the online channel when the catalog is mailed. Notice that almost no online demand occurs in this scenario. So, the catalog drives orders online, but the online channel is not yet capable of generating its own incremental volume. The online channel is a glorified order form.


Model #3 = True Catalog Multichannel Model

Incremental Value Of Catalog Marketing
		Other
	Catalog	Catalog	Online	Retail	Total
	Demand	Demand	Demand	Volume	Volume
Mailed Segment	$3.00	$7.00	$3.00	$0.00	$13.00
Holdout Segment	$0.00	$8.10	$1.50	$0.00	$9.60
Increment	$3.00	($1.10)	$1.50	$0.00	$3.40
Incremental Results:		$3.40 / $3.00 =		113.3%
Matchback Analysis:		$3.00 + $3.00 =		$6.00

Notice the significant differences in this business model. If the catalog is not mailed, half of the online demand occurs anyway. This is a view that many catalogers are missing these days, due to an over-dependence upon matchback analyses. In this case, $3.40 of demand per customer were generated. However, the matchback analysis indicates that $6.00 of demand per customer were harvested. If the cataloger goes with the latter, the executive team will significantly over-circulate catalogs, causing profit to be sub-optimized. This is probably the most significant analytical error happening in our industry these days --- our list processing, compiled list vendors, industry experts and and paper representatives have unknowingly pushed us down this path, and we let it happen. Nobody is to blame, it's simply our responsibility to do a better job of analyzing the business models we manage.


Model #4 = Retail Business, Catalog Heritage

Incremental Value Of Catalog Marketing
		Other
	Catalog	Catalog	Online	Retail	Total
	Demand	Demand	Demand	Volume	Volume
Mailed Segment	$3.00	$6.00	$3.00	$3.00	$15.00
Holdout Segment	$0.00	$7.00	$2.00	$2.00	$11.00
Increment	$3.00	($1.00)	$1.00	$1.00	$4.00
Incremental Results:		$4.00 / $3.00 =		133.3%
Matchback Analysis:		$3.00+$3.00+$3.00		$9.00

These results are interesting. In a true multichannel version of a catalog business model, volume is spread across other catalogs, the website, and retail stores. Typically, the catalog will drive modest amounts of volume online, and to stores. Notice that online and store channels still get a ton of volume, even if the catalog is not mailed. In these cases, matchback analyses are flat-out wrong --- much care needs to be taken to accurately read matchback analyses in a retail environment of this nature.


Model #5 = Online Business, Retail Heritage

Incremental Value Of Catalog Marketing
		Other
	Catalog	Catalog	Online	Retail	Total
	Demand	Demand	Demand	Volume	Volume
Mailed Segment	$1.00	$4.00	$5.00	$20.00	$30.00
Holdout Segment	$0.00	$5.00	$4.00	$19.00	$28.00
Increment	$1.00	($1.00)	$1.00	$1.00	$2.00
Incremental Results:		$2.00 / $1.00 =		200.0%
Matchback Analysis:		$1.00+$5.00+$20.00		$26.00

These business models are also fascinating. Notice that catalog advertising plays a very small role in influencing business results. Online demand and retail volume are barely moved by the mailing of a catalog. Yet, in total, the catalog is twice as effective as source code reporting would indicate.


There's no need to talk about online pureplays, as catalog dynamics are not part of that equation.

Given what has been shared over the past two days, what are your thoughts? Does this framework make sense? What are you seeing in the business models you manage? Do you agree that matchback analyses are frequently in error, sometimes significantly in error, when measuring the incremental value of a catalog?