ACCM 2009, Multichannel Forensics, Zip Code Forensics

Are you coming to my session at ACCM? If you are, and you want to include your data in the Zip Code Forensics database, bring a year's worth of data (one row per zip code, each column represents total annual sales in that zip code by channel), and I'll be sure to include your data in the database. Please contact me for additional details.

For those of you who are scheduling time with me to talk about your Multichannel Forensics for Catalog Marketers projects, review the slide presentation below. This explains what a typical Multichannel Forensics project for Catalog Marketers is, and illustrates the cost of a project. The projects are great for reducing catalog expense while minimizing demand loss. Let me know if you want to schedule time to meet at ACCM 2009.

Multichannel Forensics Projects For Catalogers: 2009

View more presentations from MineThatData.

ACCM 2009: Attend My Session And Get A Speaker Discount For The Conference

I've heard from many of you who plan on attending my talk at ACCM 2009 in New Orleans. Now how about the rest of you??

My session is on Wednesday, May 6, from 11:15am - 12:15pm, and is titled "Multichannel Customer Behavior: Profit By Understanding The Role Of Geography". We'll do a deep dive into the magical world of Zip Code Forensics, the leading (and free) geographical segmentation tool for direct marketers.

There's still time to register, and if you use this form, you can receive a 25% discount because you are a loyal member of MineThatData Nation.

I'll be around on Wednesday (5/6) if you wish to get together to talk about your Multichannel Forensics project, or you wish to discuss an upcoming Multichannel Forensics project.

And I might even bring along a couple of Multichannel Forensics books to share with loyal blog readers.

So please attend my session!!!!

Housefile Modeling & Co-Ops

In customer acquisition, you may have a mutually beneficial relationship with the co-ops. You get enough value to grow your business, they generate enough revenue to be profitable.

In housefile modeling, however, the story changes. Clients using Zip Code Forensics, a free tool that outlines the most productive zip codes, tell me that they are achieving 80% to 85% of the performance improvement seen via co-op housefile modeling. Here, however, the difference in cost is enough to turn the tables.

	Co-Ops	Zip Code Fr.	Base Perf.
Circulation	25,000	25,000	100,000
Buyers	313	303	1,100
Response Rate	1.25%	1.21%	1.10%
Average Order Value	$135.00	$135.00	$135.00
Demand	$42,188	$40,838	$148,500
Contribution	$14,766	$14,293	$51,975
Less Book Cost	$13,750	$13,750	$55,000
Less Co-Op Expense	$625	$0	$0
Variable Profit	$391	$543	($3,025)
Profit Per Buyer	$1.25	$1.80	($2.75)

In this example, Zip Code Forensics only gets us 73% of the improvement (you tell me that average number is about 82%) yielded by the co-op housefile model. However, at $0.025 per housefile name, you don't generate enough business to offset the benefit of the freely available Zip Code Forensics segmentation tool.

Unless your co-op is tossing in a bunch of goodies, you're better off using Zip Code Forensics to overlay against lapsed buyers. You went for the lower cost solution when you abandoned the list industry for the co-ops. Why not use a free tool instead of a low-cost solution?

New England vs. Silicon Valley

There are fundamental differences in the way our customers shop, based solely on geography.

New England is home to a veritable plethora of viable catalog brands, and we see a lot of orange on the map, don't we? A catalog in the mailbox is a welcome part of the day in New England.

Meanwhile, 3,000 miles across the country we find Silicon Valley. This is the hub of technological innovation! Notice that almost the entire metro Bay Area is light green or dark green --- an e-commerce hotbed.

It should not come as a surprise that the most popular third-party catalog opt-out service is located in California. For many in this region, a catalog in the mailbox is an unwelcome intrusion, symbolic of the destruction of the planet, or a relic of a bygone marketing era, at minimum.

The catalog marketer of the 2010s will have a three-pronged marketing strategy, based on customer differences and geographical variations.

1. A catalog marketing strategy for the 50+, rural, Mountain and New England based catalog shopper.
2. An e-commerce marketing strategy for the middle-aged suburban customer.
3. A social shopping strategy for the under-40, suburban/urban customer who fully integrates technology with lifestyle.

This won't be easy, but it is essential to the profitability of the businesses we manage.

Zip Code Forensics: Version 2 Is Free And Now Available

I welcome all of the companies that volunteered anonymous zip code sales data during the past three months. Your contributions result in an improved version of B2C Zip Code Forensics!

If you are interested in participating in Zip Code Forensics, at no cost, please download the following documents.

• A Graphical Representation Of Version 2 Of Zip Code Forensics.
• Spreadsheet With The Actual Segment Assignments For The State Of Washington.

And by the way, what do you notice about urban/suburban/rural zips in Washington? Most of the Seattle/Tacoma and Spokane area is green (e-commerce). Most of the zips in the Cascades and Rockies (rural) are orange (catalog focused). Keep thinking about this urban/suburban/rural thing, and the place for catalogs in the future of direct marketing.

Test results from our participants indicate that if a cataloger only mails the Catalog Crazies and Online Bliss segments, the quantity in the RFM segment or outside list will be reduced by about 75%, with improvements in performance of between 7% and 20%.

There aren't many folks offering you the ability to improve your marketing efforts at no cost --- and the algorithm suggests that the more participants we get, the better the algorithm performs!

So join us ... contact me for information on how to participate, for free!

And B2B marketers --- we're very close to having a version of Zip Code Forensics that works in a reliable way for your channel. We need a few more participants to make this work well!

Here are the six segments:

1. Catalog Crazies: The most productive traditional direct marketing zip codes.
2. Catalog Fans: Average performing traditional direct marketing zips.
3. Catalog Preference: Marginal performing traditional direct marketing zips.
4. Online Preference: Marginal performing e-commerce zips.
5. Online Spend: Average performing e-commerce zips.
6. Online Bliss: The most productive e-commerce zips.

Geographic Differences

This is a map of the Northern Great Plains and Northern Rocky Mountain states.

Orange zip codes represent customers who are responsive to traditional direct marketing. Green zip codes represent customers who shop via e-commerce.

Notice the big difference between the Great Plains states and the Rocky Mountain states!

We have a lot of marketing challenges in the future, given that so much of marketing is becoming "pull" marketing (the customer pulls the information they want, using the channels they wish to use). One of the advantages we have is data. We control what we push, and we don't have to push something to an individual who doesn't want information pushed to them.

If we want to test a new merchandise line in an e-mail campaign, there's no reason we couldn't select the "green zip codes" in the map above, coupled with responsive e-mail subscribers who have a propensity for the merchandise we're sharing with the customer. We don't have to annoy the customer in North Dakota.

Take advantage of geography --- it is going to be a part of our multchannel, multi-platform future.

Sprint Cell Phone Coverage And Zip Code Forensics

There are many applications of Zip Code Forensics in marketing (free to you if you contribute anonymous annual zip code sales).

One of the more interesting applications is in understanding channel preference.

This map illustrates Northeast Minnesota and Northwest Wisconsin. The orange zip codes represent "Catalog Crazies" and "Catalog Fans", zip codes that spend a lot of money because catalogs are sent to these customers.

Now take a look at the next image, one from the Sprint website, illustrating cell phone coverage on the Sprint network.




Green is good on the map ... it means that your phone will actually work!

Now compare the areas on the Sprint map that are white (no coverage) with the Zip Code Forensics map. Areas that are white on the Sprint map are often orange on the Zip Code Forensics map (orange represents old-school catalog responsiveness).

There is a correlation (not causation, mind you) --- no mobile marketing channel correlates with a robust old-school catalog marketing channel.

This is where we, as marketers, repeatedly fail.

We try to integrate everything, to make everything the same, as if the same customer is using mobile marketing and social media and e-mail marketing and television marketing and radio marketing and newspaper marketing and direct mail and catalog marketing to make the crucial decision to purchase a $19 shirt.

The maps suggest an opportunity. Why not work with the natural constraints that customers live with? In other words, in areas where mobile marketing isn't possible, why not fully capitalize on catalog marketing? And in Minneapolis / St. Paul (heavy e-commerce --- with ubiquitous 3G mobile broadband available), why not take full advantage of communicating with users who are updating Facebook pages on their 3G phones?

The data suggest we could do this.

And why not contribute your anonymous annual sales data at a zip code level, so that you can have access to the insights available in Zip Code Forensics?

Alaska: An Online Marketing Hotbed

There are nice catalog marketing zip codes in Alaska as well.

Buy online / pickup in stores doesn't work so well in Alaska. This is one state that would benefit from a bit of customer understanding and personalization!


Hillstrom's Zip Code Forensics is free for all direct marketers contributing anonymous, annual sales totals by physical channel (mail, phone, online, retail) at a zip code level.

To-date, participants are seeing an improvement in direct marketing activities against the six segments of between 5% and 20%, based on actual tests and backtests.

I'm also looking for a few more B2B volunteers --- I am creating a separate model for businesses selling to businesses --- the dynamics are simply different than the B2C model that has been created. Please contact me if you want to participate, for free.

The Best Catalog Zip Codes

Modern cataloging disproportionately skews to various rural areas in New England.

This map illustrates a hyper-responsive cluster of zip codes in New York, Vermont, and New Hampshire.

Urban areas (Boston) love e-commerce. Rural areas shop via catalogs, and shop disproportionately over the telephone.

Our job is to recognize the fundamental differences in customer behavior represented by geography. Mobile marketing and social media and a "Twitter-storm" are not as relevant in New Hampshire as a well-merchandised 124 page catalog that drives the customer to a call center, where a pleasant voiced sales associate walks a customer through an order.

The fact that almost nobody in our industry talks about this stuff is a bit surprising, isn't it? After all, you can improve the profitability of marginal segments by between 5% and 20% by applying Zip Code Forensics, based on actual tests and backtests.

Consulting Project Focus Is Changing

The Multichannel Forensics projects you're asking me to work on have taken a turn.

Two years ago, you asked me to explain how customers interacted with channels.

One year ago, you asked me to explain how customers interacted with both channels and merchandise divisions, with an eye toward forecasting the future.

Then you saw what the future held, and it wasn't pretty.

Today, you ask me to use Multichannel Forensics to identify customers who will keep purchasing if advertising is significantly reduced.

The framework isn't significantly different than Multichannel Forensics projects from a few year ago. I still measure how customers interact with products, brands, and channels. And I still forecast the long-term trajectory of your business by product, brand or channel.

The mechanics of the project, however, have changed. We use whatever data is available to understand how customers move along the continuum above --- organic, social, algorithm, advertising, and begging. We attempt to identify where the customer resides on this continuum.

Customers who respond to begging (discounts, promotions, free-shipping, GWPs) are at the bottom of the ladder. We'll need to market to them, and we'll need to give them a reason to purchase. These may be profitable customers, but we'll have to work hard at creating gimmicks to encourage them to purchase. This is the realm of the marketer, especially in Fall 2008. In so many ways, we ruined e-commerce with our obsession of begging customers to purchase.

Traditional direct marketing focused on customers who respond to advertising. This is a segment of the customer file that is decreasing in size. We look for attributes that suggest a customer must be advertised to, in order to purchase. Customers who order over the telephone, customers who give catalog key-codes when shopping online, customers who click-through e-mail campaigns, customers living in zip codes classified as "Catalog Crazies". These customers are unlikely to buy in the future unless they are marketed to.

Then we have customers who use algorithms to purchase. Yup, these are the customers who use tools like paid search to purchase merchandise. These customers are different. They don't always respond to future advertising, and when they do respond, they combine advertising and algorithms to make decisions. This is where your Net Google Score comes into play. Catalog brands really struggle with algorithm customers, and online marketers struggle with e-mail marketing programs for algorithm customers.

Increasingly, we find ourselves managing social customers. If you're Crutchfield, you have customers who buy merchandise, customers who write reviews, and customers who are referred from blogs to your site. The latter two groups represent "social customers". Social customers are different than are typical catalog customers, and are different than typical e-commerce customers. Catalogers are way behind the curve when it comes to managing social customers. In fact, almost everybody is behind the curve regarding social customers. Hint: Social customers don't necessarily embrace catalogs, and sometimes get really angry when they are stuffed in the mailbox.

Finally, we get to the most valuable customer on the planet, the organic customer! I receive a lot of criticism about my assertion that there are customers who do not need to be advertised to. Why? I don't know. Many of you think customers only buy something if they are advertised to. Amazon.com gets a lot of organic business. Now it is true that maybe Amazon sent an e-mail at one time, and you bought because customers like you purchased certain texts. But that doesn't explain the fact that you see "Outliers" discussed on a blog, so you go and buy the book on Amazon (that makes you a social customer!). Or maybe you read about the book in New York Magazine, then buy it on Amazon (that makes you an organic customer). Organic demand is the most important kind of demand to generate, because it comes without advertising cost. Retailers have thrived for centuries via organic demand. E-commerce is a hybrid of retailing (organic demand) and cataloging (advertising demand).

So how did project work change?

These days, I score customers across each of the five dimensions listed above. If the customer generates organic demand, the customer gets an "A", if not, the customer gets an "F". The same process happens for Social Customers, Algorithm Customers, Advertising Customers, and for Customers Who Respond To Begging.

Once customers are graded, we monitor migration. Does the "Begging" customer migrate to "Organic" status? If so, then discounts and promotions work! Does the "Algorithm" customer slide down to "Begging"? If so, then Google isn't doing us any favors. Is the "Advertising" customer married to advertising? If so, then we have to keep streaming the catalogs at this customer. We apply the migration patterns, understanding the long-term trajectory of your business. Finally, we identify the customers who we can stop marketing to, without a significant dip in business.

Online pureplays are using this methodology, too ... they want to understand who should receive e-mail marketing, and they want to understand how deep they should dive into paid search.

Retailers ask me to do this, so that they can identify retail shoppers who are unresponsive to direct marketing, customers who have a high organic percentage.

Catalog Choice should love this (especially given the slowdown in user growth in recent months), because the end result of the project is the discontinuation of catalog marketing to customers who no longer respond to advertising, while protecting the catalog relationship among highly responsive customers.

That, my dear readers, is a description of the type of project I am being asked to work on by online marketers, retailers, and catalog brands. And it is big-time fun!

Salt Lake City

Compared to Medford, Salt Lake City is an internet hotbed.

Zip Code Forensics suggests that, east of town, you have a plethora of high-spending skiers who love to shop online.

Ok, marketers, tell me how your strategy for the customer in Medford compares with the strategy for the customer east of Salt Lake City?

Medford

I'd guess that most of you haven't been to Medford, or haven't taken that curvy two lane highway down to Crescent City.

Southern Oregon is catalog country, as illustrated by Zip Code Forensics. You're looking at dial-up internet access, folks. The only "tweets" you hear are from birds.

Social media pundits struggle here, don't they? --- the average person isn't carrying a 3G phone with fancy apps. The social media pundit would dismiss this person as being a technophobe ... even though the 3G network doesn't reach all people in this area ... in fact, the old-school network doesn't reach all residents in Southern Oregon.

Retailers can be snobby, too. You won't find $30,000 gifts from Neiman Marcus in any shopping center in Ashland. But they do sell a lot of t-shirts at the sprint car races in Medford.

You're a marketer, or you wouldn't be reading this. What is your strategy for the customer living in Southern Oregon? Or, do you just disregard this customer? Or do you treat this customer the same as you treat the customer living in Manhattan?

Your choice.

How About A Few More B2B Volunteers For Zip Code Forensics?!

We're still looking for a few more volunteers for the B2B version of Zip Code Forensics.

Thanks to all of you who volunteered data during the past week, I hope you enjoy using your own custom-built model in the weeks prior to the development of the official B2B model.

We also received word from one of our volunteers that the model performed 20% better than a control group, in actual testing --- and performed essentially the same as co-op models (better in terms of profit due to the fact that Zip Code Forensics is free).

For your viewing pleasure, I present you with the B2C version of Zip Code Forensics, illustrating the Charlotte, NC area. This is an e-commerce dominated area, with significant pockets of high-potential customers in the Lake Norman / I-77 corridor, and south of the city. Unlike most cities, Charlotte has good performing e-commerce zips in the central city area.

B2B Brands: A Customized Zip Code Forensics File Just For You!

The research strongly suggests that B2B brands have a different geographic customer profile than do B2C brands.

Therefore, it makes sense to have separate models --- one segmentation strategy for B2C, one for B2B.

So, I'm looking for a half-dozen B2B volunteers who wish to contribute data. You send data with one row per zip code, columns including zip code, 12-month phone sales, 12-month online sales, 12-month retail sales for that zip code.

I build a custom model for B2B brands, and everybody benefits.

Of course, there's no cost to you. E-Mail me your questions, I look forward to hearing from you.

B2C brands --- Zip Code Forensics is free for those who contribute data, $5,000 a year for everybody else.

Indianapolis

Going into 2009, we have an opportunity to view marketing differently.

This is Indianapolis. It's a "green" market, meaning the good folks of this area prefer e-commerce, especially those north of the city.

Industry experts suggest that customers want a seamless experience across channels. And that might be the best thing to do.

It might be worth considering what you do with a retail customer living in on the southwest side of town. This is an area that is not responsive to direct marketing. If you're marketing to a small audience on this side of town, direct marketing might not be the most effective way to get a message out to the customer.

That's where the secret sauce comes in. Our future is bright when we look at each channel as a customized tool, tailored to the individual.

In other words, we can offer six marketing channels to one customer, or we can offer a unique marketing approach to six customers. Long-term, the latter makes sense, an approach we'll adapt to over time.


Hillstrom's Zip Code Forensics: Free For Brands Contributing Anonymous Zip Code Sales By Channel, $5,000/year For Everybody Else.

Housing Market Deflation Correlates With Zip Code Forensics

Don Libey forwards us this map of the United States as illustrated in the NY Times (see via this link, related article here), illustrating states with high "debt to value" ratios.

His question ... does this data correlate with Zip Code Forensics?

The answer ... yes.

We can match state-level data to the zip-level data in Zip Code Forensics. Statisticians would point out the flaws associated with doing this. Point taken.


Catalog Crazies, the most productive zip codes in the United States, have an average debt-to-value ratio of 63.0%.

Online Bliss, among the most productive zip codes in the United States, have an average debt-to-value ratio of 63.5%.

Catalog Fans, zip codes with average productivity, have an average debt-to-value ratio of 63.8%.

Online Spend, zip codes with average productivity, have an average debt-to-value ratio of 66.2% (oh oh, a trend is beginning).

Catalog Preference, zip codes that perform well below average, have an average debt-to-value ratio of 67.8%.

Online Preference, zip codes that perform well below average, have an average debt-to-value ratio of 69.0%.

In other words, Zip Code Forensics identifies regions that have customers who, on average, have an additional four to six percent equity in their homes. These customers are less impacted by the mortgage/housing crisis, and therefore, have more money to spend on other things.

And you can have access to this data for free! Contribute your anonymous annual sales data by zip code by channel to the database, and you'll receive this segmentation system. If you do not wish to contribute your data, the cost for an annual license is $5,000.

Download the paper to learn more.

Hillstrom's Zip Code Forensics: The Beta Roster Is Complete!

Pending data from two brands, the roster for Hillstrom's Zip Code Forensics is complete ... and the roster is a spectacular blend of leading multichannel brands spanning a wide array of merchandise presentations!

Thank you so much to those who volunteered your data! It is now my job to create something that will increase sales and profit. I so look forward to doing that!

The dataset has outstanding coverage ... the full roster of brands generated over a billion dollars of sales over the past twelve months. That provides plenty of coverage to make good, reliable insights.

Most interesting is the fact that the data already debunk many of the leading hypotheses about multichannel marketing. In fact, the data strongly suggest there is a compelling geographic component to customer behavior, a geographic component that lends itself to the practice of micro-channel marketing. Multichannel marketing as we know it today struggles to fit within the framework harvested by actual customer data.

The beta roster will benefit from free updates and use of the information for a minimum of one year. Kudos to all of you willing to take a risk on a new product, I am very appreciative!

Zip Code Forensics And Regional Differences

Geography continues to play a significant role in our marketing efforts. It's the reason for introducing Hillstrom's Zip Code Forensics.

In the social media world, folks are beginning to take notice.

In the multichannel marketing side of the world, zip code models have been in existence for a very long time, as evidenced in this article from 2000.

We haven't done a good job of drawing intelligence from our models. It's one thing to see if we can improve the effectiveness of marketing activities by treating customers in certain geographic areas differently.

It's quite another thing to change how we practice marketing, based on what the information tells us.

Looking at the map above, we notice a lot of orange in Southern Iowa and Northern Missouri. These zip codes tend to be less productive than average, and skew toward traditional marketing techniques.

In other words, we shouldn't expect these areas to be highly responsive to tactics like e-mail marketing or social media. This doesn't mean individual customers won't be responsive. But on average, we need to think differently about these customers.

How do we merchandise a catalog for the customer who lives in a rural area? How do we merchandise a catalog for the customer who has 1,443 retail opportunities within two miles of her home? How do we execute an e-mail marketing campaign for the customer in Northern Missouri? How do we execute an e-mail marketing campaign for the customer living in Silicon Valley? What version of a home page or landing page could be created for the rural customer, or the urban customer?

We're going to get better at using the tools already available to us. We will change how we present ourselves to unique customer segments.

When Zip Code Forensics Meet Multichannel Forensics

In a world where customers receive discounted offers for computers offered on Twitter, it may seem unusual that some customers still prefer to mail an order form to the brand they wish to place an order from.

Those of you with 1.21 gigawatt cable broadband speed can get up off the floor now and go use FriendFeed to notify the world of this shocking development.

Customers who mail their orders to brands are different than other customers. This is where Zip Code Forensics and Multichannel Forensics play a role in managing the customer relationship.

Here are repurchase indices for a brand that has a lot of customers who place orders via the mail. We appropriate the repurchase index across the four Zip Code Forensics segments. Take a look at zip codes that prefer catalogs, and have low sales potential. The Online channel has a very low repurchase index.

In other words, customers who live in low potential zip codes that prefer catalog marketing (i.e. rural areas) are unlikely to use the online channel, while all other customers who mail their orders have a reasonable chance of using the online channel in the future.

Eventually, we're going to see marketers divide their audience into sub-segments. Some customers simply want to have a traditional relationship with a brand. Other customers want to have a digital relationship with a brand. In catalog and e-mail marketing, we'll execute different versions of our marketing efforts for the traditional audience. This is the audience that can stomach 16 catalogs a year, twelve of which are remailed catalogs. The digital audience will require a different, more personalized strategy.

And we'll use Zip Code Forensics and Multichannel Forensics to help us identify the tactics we need to employ.

Marketers: Spaces are filling up quickly for the beta test version of Hillstrom's Zip Code Forensics. The first ten marketers get to participate in the product for free. Contact me immediately if you wish to participate in the free version of the product.