Reverend Billy might suggest that the decision to discontinue live piano music in select Nordstrom stores is an indicator of the coming "shopocalypse".
For some, this is like McDonalds taking down the Golden Arches, or Nike removing the swoosh from their shoes. The decision erodes the story the consumer tells herself about why she chooses one retailer over another.
I detest it when bloggers indiscriminately rip big companies in the name of generating additional "subscribers" ... I mean, really, how different is that from when big companies make bad decisions in the name of shareholder value? Real people, good people, work at big companies, doing good things. I don't like to see these people harmed.
I also detest it when big companies indiscriminately humble the very consumers they depend upon for their existence, exploit low-wage workers to satisfy shareholders, or slowly deviate from the things that made them special, different.
Do we buy from a retailer because they hire a person to play piano? Do we buy from a retailer because they meet a need at an acceptable price? Do we buy from a retailer because they are nearby and convenient? Do we buy from a retailer because the employees treat you well?
Or do we buy because of the way these factors "interact" with each other?
Most likely, it is the latter. And nobody is smart enough to forecast how one factor interacts with another. We just wait and see.
November 30, 2007
Books You've Purchased
Don't know why I didn't do this sooner ... I recently joined the Amazon Associates program.
I get to see some interesting information. For instance, about eleven percent of you buy something when clicking-through one of my two book links. More than half of those purchases end up being something that I didn't author.
Here's what you end up purchasing:
I get to see some interesting information. For instance, about eleven percent of you buy something when clicking-through one of my two book links. More than half of those purchases end up being something that I didn't author.
Here's what you end up purchasing:
- Hillstrom's Multichannel Forensics (thank you).
- Hillstrom's Database Marketing (thank you again!).
- Super Crunchers by Ian Avres.
- Libey and Pickering on RFM and Beyond, by Don Libey and Christopher Pickering.
- Web Analytics An Hour A Day by Avinash Kaushik (ok, you're not buying this book, but you should be). It's the 2,441st most popular book on Amazon, written by one of our "Friends of MineThatData". I bought the book ... so far, so good. All proceeds are donated to charity.
- The Business of Database Marketing by Richard Tooker.
- Optimal Database Marketing by Ronald G. Drozdenko.
- Data Mining Techniques by Michael A. Berry.
- Competing on Analytics, by Thomas H. Davenport.
- Strategic Database Marketing by Arthur Middleton Hughes.
- Marketing Metrics by Paul W. Farris.
- Waiting For Your Cat To Bark by Bryan Eisenberg.
- The Long Tail by Chris Anderson.
- Know How, The Eight Skills That Separate People, by Ram Charan.
- Made To Stick, by Chip Heath.
November 29, 2007
Diminishing Returns, The Square Root Rule and Peak Profit
We've previously talked about diminishing returns in marketing.
Understanding how diminishing returns impact profit is something that anybody can do.
You don't have to be accountable for paid search or online marketing or catalog marketing to understand how effective your strategy is at a macro level.
And the level of science is irrelevant. Most statisticians would become paralyzed by all of the assumptions being violated here. You're not trying to be a perfect statistician. You're trying to approximately understand where marketing spend is optimized.
Here's Example #1:
Here's where you have a series of choices. You are generating $600,000 of profit at $1,000,000 of paid search spend. However, "peak profit" occurs at $600,000 of spend. If you want to achieve better profitability, you spend $400,000 less, and give up nearly a million dollars of demand.
Folks who measure lifetime value combine short-term and long-term profit, and probably end up spending more than a million dollars based on those findings. Notice how diminishing returns occur, especially after a million dollars of spend.
Also notice how few online marketers measure anything beyond "cost per order" --- there's an opportunity for online marketers to improve their "tool box" with this style of analysis.
Example #2:
Here, we see that "peak profit" occurs at $28,000,000 of catalog spend. However, the amount of profit difference between $28,000,000 of spend and $40,000,000 of spend is not significant. Most marketers would err on the side of spending more, in this instance, due to the obvious benefits of growing the customer file.
Again, you don't have to be responsible for online marketing, catalog marketing, or e-mail marketing to do this type of analysis. Take the initiative to get your hands on some data, partner with your finance team if necessary, and analyze where "peak profit" occurs in your marketing efforts. Find out where diminishing returns take a bite out of profit.
Understanding how diminishing returns impact profit is something that anybody can do.
You don't have to be accountable for paid search or online marketing or catalog marketing to understand how effective your strategy is at a macro level.
And the level of science is irrelevant. Most statisticians would become paralyzed by all of the assumptions being violated here. You're not trying to be a perfect statistician. You're trying to approximately understand where marketing spend is optimized.
Here's Example #1:
- Your paid search marketing budget is forecast to be $1,000,000 this year. You also expect to generate $4,000,000 demand. You convert demand to profit at a rate of 40%.
- Profit = $4,000,000 * 0.40 - $1,000,000 = $600,000.
- Using the square root rule, we can measure where "peak profit" occurs.
| Square | |||
| Spend Levels | Root | Demand | Profit |
| $500,000 | 0.707 | $2,828,427 | $631,371 |
| $600,000 | 0.775 | $3,098,387 | $639,355 |
| $700,000 | 0.837 | $3,346,640 | $638,656 |
| $800,000 | 0.894 | $3,577,709 | $631,084 |
| $900,000 | 0.949 | $3,794,733 | $617,893 |
| $1,000,000 | 1.000 | $4,000,000 | $600,000 |
| $1,100,000 | 1.049 | $4,195,235 | $578,094 |
| $1,200,000 | 1.095 | $4,381,780 | $552,712 |
| $1,300,000 | 1.140 | $4,560,702 | $524,281 |
| $1,400,000 | 1.183 | $4,732,864 | $493,146 |
| $1,500,000 | 1.225 | $4,898,979 | $459,592 |
Here's where you have a series of choices. You are generating $600,000 of profit at $1,000,000 of paid search spend. However, "peak profit" occurs at $600,000 of spend. If you want to achieve better profitability, you spend $400,000 less, and give up nearly a million dollars of demand.
Folks who measure lifetime value combine short-term and long-term profit, and probably end up spending more than a million dollars based on those findings. Notice how diminishing returns occur, especially after a million dollars of spend.
Also notice how few online marketers measure anything beyond "cost per order" --- there's an opportunity for online marketers to improve their "tool box" with this style of analysis.
Example #2:
- You spend $40,000,000 on catalog marketing. Although your productivity has been in decline for several years, your catalogs are still profitable. You generate $170,000,000 of demand across channels, with 40% converted to profit.
- Where does "peak profit" occur?
| Square | |||
| Spend Levels | Root | Demand | Profit |
| $20,000,000 | 0.707 | $120,208,153 | $28,083,261 |
| $24,000,000 | 0.775 | $131,681,434 | $28,672,574 |
| $28,000,000 | 0.837 | $142,232,205 | $28,892,882 |
| $32,000,000 | 0.894 | $152,052,622 | $28,821,049 |
| $36,000,000 | 0.949 | $161,276,161 | $28,510,464 |
| $40,000,000 | 1.000 | $170,000,000 | $28,000,000 |
| $44,000,000 | 1.049 | $178,297,504 | $27,319,002 |
| $48,000,000 | 1.095 | $186,225,670 | $26,490,268 |
| $52,000,000 | 1.140 | $193,829,822 | $25,531,929 |
| $56,000,000 | 1.183 | $201,146,713 | $24,458,685 |
| $60,000,000 | 1.225 | $208,206,628 | $23,282,651 |
Here, we see that "peak profit" occurs at $28,000,000 of catalog spend. However, the amount of profit difference between $28,000,000 of spend and $40,000,000 of spend is not significant. Most marketers would err on the side of spending more, in this instance, due to the obvious benefits of growing the customer file.
Again, you don't have to be responsible for online marketing, catalog marketing, or e-mail marketing to do this type of analysis. Take the initiative to get your hands on some data, partner with your finance team if necessary, and analyze where "peak profit" occurs in your marketing efforts. Find out where diminishing returns take a bite out of profit.
November 28, 2007
Gift Wrapping At Amazon.com
Here's a thorny challenge.
Maybe you're like me (most likely you're not like me, but that's a topic for another day). Being 1,700 miles from family, you visit Amazon to purchase a gift for your nephew for Christmas.
You select the gift, then you select gift wrapping. Except, the item isn't eligible for gift wrapping. The item comes from one of Amazon's partners.
In this case, you run into a unique challenge in multichannel marketing. Amazon uses a vendor that allows for a different customer experience than the experience typically promoted by Amazon.
If you were CEO of a brand that had business partners with different customer service capabilities than your brand, how would you deal with this challenge?
Maybe you're like me (most likely you're not like me, but that's a topic for another day). Being 1,700 miles from family, you visit Amazon to purchase a gift for your nephew for Christmas.
You select the gift, then you select gift wrapping. Except, the item isn't eligible for gift wrapping. The item comes from one of Amazon's partners.
In this case, you run into a unique challenge in multichannel marketing. Amazon uses a vendor that allows for a different customer experience than the experience typically promoted by Amazon.
If you were CEO of a brand that had business partners with different customer service capabilities than your brand, how would you deal with this challenge?
November 27, 2007
A Record Cyber Monday!
At least it was a record "Cyber Monday" at The MineThatData Blog.
I'm willing to bet that most online retailers could make this claim for at least 325 different days during 2007 ... for instance, on October 22, 2007, online sales outperformed those of October 22, 2006, and outperformed those of October 22, 2005, suggesting record performance on October 22 of this year.
And you didn't need to offer special promotions to make magic happen on October 22.
And that's the sad part of all of this. By creating an online holiday that never really existed, business leaders believed the hype, and discounted their brand with %-off and free shipping promotions in an effort to "remain competitive". What a shame.
Update 11/28/2007, 6:00am PST: Another press release from Shop.org suggests that 72,000,000 Americans were planning to shop on Cyber Monday. Shop.org also estimated that $700,000,000 was spent online on Cyber Monday. Let's do some math ... $700,000,000 spent online divided by the 72,000,000 customers "planning" to shop online = between $9 and $10 spent per planned online shopper. We know that number isn't mathematically possible (average order sizes are often +/- $100), suggesting the online research is highly flawed. In reality, it is more likely that about 7,000,000 customers were planning to shop online on Cyber Monday. Ugh, endless hype.
- Site visitors were up 150% over last year's Cyber Monday bonanza.
- RSS subscribers were up 700% over last year.
- There were no incentives, doorbusters, or "free" promotions of any kind required to drive this kind of traffic.
I'm willing to bet that most online retailers could make this claim for at least 325 different days during 2007 ... for instance, on October 22, 2007, online sales outperformed those of October 22, 2006, and outperformed those of October 22, 2005, suggesting record performance on October 22 of this year.
And you didn't need to offer special promotions to make magic happen on October 22.
And that's the sad part of all of this. By creating an online holiday that never really existed, business leaders believed the hype, and discounted their brand with %-off and free shipping promotions in an effort to "remain competitive". What a shame.
Update 11/28/2007, 6:00am PST: Another press release from Shop.org suggests that 72,000,000 Americans were planning to shop on Cyber Monday. Shop.org also estimated that $700,000,000 was spent online on Cyber Monday. Let's do some math ... $700,000,000 spent online divided by the 72,000,000 customers "planning" to shop online = between $9 and $10 spent per planned online shopper. We know that number isn't mathematically possible (average order sizes are often +/- $100), suggesting the online research is highly flawed. In reality, it is more likely that about 7,000,000 customers were planning to shop online on Cyber Monday. Ugh, endless hype.
November 26, 2007
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.
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.
November 25, 2007
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.
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.
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.
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.
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.
More RSS Feeds Posted On The MineThatData Blog
When I asked for feedback about this blog, I learned that this blog was used as a homepage (by the way, I've awarded the free book, thanks for all of the feedback!).
So I've expanded the number of RSS feeds featured on the right side of the page. Each day, you'll be able to enjoy content from many different sources in the direct marketing industry, including the following:
You might be wondering about some of the comments folks offered for this blog. Here's a sampling of comments:
So I've expanded the number of RSS feeds featured on the right side of the page. Each day, you'll be able to enjoy content from many different sources in the direct marketing industry, including the following:
- DMNews, Multichannel Merchant, Catalog Success, Direct Magazine, Seth Godin, Marketing Sherpa, Forrester Research, Shop.org, Marketing Profs Daily Fix, Avinash Kaushik, Alan Rimm-Kaufman, Jim Novo, David Raab, Tamara Gielen, Juice Analytics, LunaMetrics, Becky Carroll, James Taylor, Adelino de Almeida, Sandro Saitta, Jeff Larche, Suzanne Obermire, Chris Baggott, Alison Bolen, Customer U Blog, Michael Fassnacht, Sandeep Giri, Rick Whittington, Anthony Power, Gretchen Scheiman, Pat LaPointe, Dean Abbott, Sarah Clelland / Snow Patrol, Mark Brownlow, Mack Collier, Grant Johnson / Johnson Direct, Manuel Lima / Visual Complexity, Ted Grigg, Harry Joiner, F. Curtis Barry, Bob Bly, Charlene Li, John Hagel, Varien's e-Commerce Cache Blog.
You might be wondering about some of the comments folks offered for this blog. Here's a sampling of comments:
- More discussions about retailers.
- Shorter posts.
- More discussions about analytics, and show the actual solutions.
- More free spreadsheets (mentioned often).
- Re-define what a "catalog customer" is.
- Impact of price of oil on direct merchants.
- Diminishing returns of increased advertising.
- More on time series modeling.
- More on business applications of data mining (mentioned often).
- Make big ideas implementable for small businesses.
- More B2B ideas and discussions.
- Talk more about measuring campaigns across channels.
- Keep the list of posts broad, not narrow in scope.
- Talk more about ROI.
November 24, 2007
Organic Demand
One of the least understood concepts in cataloging is the concept of "organic demand".
A long time ago (aka 1994), catalogers controlled every aspect of their business. Sales were not generated unless a catalog was mailed.
Conversely, retailers thrived almost entirely on the concept of "organic demand". In other words, because there was an Ann Taylor store at the Galleria, sales were going to be generated. Ann Taylor could do absolutely no advertising whatsoever, and yet, the store would have loyal customers shopping every few months.
Back in 1994, you analyzed all customers who purchased in 1993. For catalogers, the relationship looked like this:
Starting in 1995, e-commerce ruined each discipline.
See, e-commerce is a true hybrid of cataloging and retailing. Today, a cataloger might have the following relationship:
And this is where we, as direct marketers, fail miserably. We want to "attribute" every single order to one of our marketing strategies. We are systematically frustrated by the customer who simply types http://cuddledown.com and buys something.
So we do a matchback analysis, and claim that a catalog mailed eight weeks ago must have been responsible for this order.
If you work for a cataloger, and struggle with the concept of "organic demand", demand that occurs without any advertising, set up an appointment to meet with a non-competitive retailer. If you work at Cuddledown of Maine (an example, not a critique of this wonderful brand), why not call the folks at a non-competitor like Best Buy, and arrange for a two day field trip? Learn how retailers cope with the concept of organic demand by spending time with folks who deal with this concept, a concept that requires a fundamentally different style of measurement.
A long time ago (aka 1994), catalogers controlled every aspect of their business. Sales were not generated unless a catalog was mailed.
Conversely, retailers thrived almost entirely on the concept of "organic demand". In other words, because there was an Ann Taylor store at the Galleria, sales were going to be generated. Ann Taylor could do absolutely no advertising whatsoever, and yet, the store would have loyal customers shopping every few months.
Back in 1994, you analyzed all customers who purchased in 1993. For catalogers, the relationship looked like this:
- Normal Mail Stream = $150 spent in 1994.
- Do Not Mail Catalogs = $0 spent in 1994.
- Normal Advertising Campaign = $150 spent in 1994.
- No Advertising At All = $130 spent in 1994.
Starting in 1995, e-commerce ruined each discipline.
See, e-commerce is a true hybrid of cataloging and retailing. Today, a cataloger might have the following relationship:
- Normal Advertising Strategy = $150 spent in 2007.
- No Advertising At All = $70 spent in 2007.
And this is where we, as direct marketers, fail miserably. We want to "attribute" every single order to one of our marketing strategies. We are systematically frustrated by the customer who simply types http://cuddledown.com and buys something.
So we do a matchback analysis, and claim that a catalog mailed eight weeks ago must have been responsible for this order.
If you work for a cataloger, and struggle with the concept of "organic demand", demand that occurs without any advertising, set up an appointment to meet with a non-competitive retailer. If you work at Cuddledown of Maine (an example, not a critique of this wonderful brand), why not call the folks at a non-competitor like Best Buy, and arrange for a two day field trip? Learn how retailers cope with the concept of organic demand by spending time with folks who deal with this concept, a concept that requires a fundamentally different style of measurement.
November 23, 2007
A Veritable Harvest Of Information
For your weekend reading pleasure:
Chad White covers the e-mail campaigns of the top one hundred retailers, so that you don't have to. Here is a summary from the past few days.
Mark Brownlow shares places to get good ideas for e-mail campaigns.
From Johnson Direct, a brief discussion about handwritten thank you cards.
Marketing Sherpa ... how to get your business book published!
Forrester Research continues to dive head-first into Facebook issues and privacy.
Tamara Gielen talks about competition in E-Newsletter Ads. Her e-mail blog is one of my favorites for reasonably unbiased information.
David Raab knows how to review software. Here's a discussion about SAS real-time decisioning.
Avinash Kaushik shares six recommendations for measuring blog success. By the way, his book on Web Analytics is a HUGE success, with proceeds donated to charity.
The Rimm-Kaufman Group has an excellent post on paid search for companies just starting out in the world of paid search.
Here's a story about the Seven Perils of Segmentation from the esteemed Jim Novo.
An oldie but a goodie from Jim Fulton on hiring, training and nurturing Database Marketers.
The folks at Juice Analytics tell us how to optimize a low traffic site.
Ron Shevlin vents about the content at conferences.
Enterprise decision management cult icon James Taylor talks about the revival of personalization.
Adelino de Almeida shares his thoughts about Starbucks and their television marketing strategy.
Sandro Saitta reviews the book, Web Dragons, a book about search marketing.
Targeted B2B marketing is one of the topics that Jeff Larche discusses this month.
RRW Consulting shares a retention case study that's worth reading.
Alison Bolen shares a few comments from a recent interview that SAS founder Jim Goodnight gave.
Customer U talks about no-cost customer research tools.
Michael Fassnacht writes about a presentation he watched on the "Read/Write" culture (I also watched this presentation, it's a good use of 20 minutes).
Sandeep Giri talks about the polarization of business intelligence solutions.
Jonathan Starets co-created a campaign that was a top three finisher for DMA Marketer of the Year.
Rick Whittington has a series on Holiday Shopping Tips. Here's tip number one, put best selling items on the homepage.
Anthony Power mentions that last month's DMA conference lacked buzz.
Gretchen Scheiman links to an article that discusses the history of e-mail.
Pat LaPointe doesn't think Google is about to dominate marketing dashboards.
UK-based Catalogue/E-Business allows readers to ask questions of direct marketing experts.
On the F. Curtis Barry blog, there's a discussion about how contact centers handle e-mail.
Harry Joiner talks about resume spam.
Ann Handley at Marketing Profs now offers snack-sized newsletters.
Mack Collier offers his readers a veritable plethora of link love to his Marketing Top 25.
Damien Francois writes about intelligent machines over at his blog.
Don Libey shares his thoughts about mergers and acquisitions with DMNews.
Chad White covers the e-mail campaigns of the top one hundred retailers, so that you don't have to. Here is a summary from the past few days.
Mark Brownlow shares places to get good ideas for e-mail campaigns.
From Johnson Direct, a brief discussion about handwritten thank you cards.
Marketing Sherpa ... how to get your business book published!
Forrester Research continues to dive head-first into Facebook issues and privacy.
Tamara Gielen talks about competition in E-Newsletter Ads. Her e-mail blog is one of my favorites for reasonably unbiased information.
David Raab knows how to review software. Here's a discussion about SAS real-time decisioning.
Avinash Kaushik shares six recommendations for measuring blog success. By the way, his book on Web Analytics is a HUGE success, with proceeds donated to charity.
The Rimm-Kaufman Group has an excellent post on paid search for companies just starting out in the world of paid search.
Here's a story about the Seven Perils of Segmentation from the esteemed Jim Novo.
An oldie but a goodie from Jim Fulton on hiring, training and nurturing Database Marketers.
The folks at Juice Analytics tell us how to optimize a low traffic site.
Ron Shevlin vents about the content at conferences.
Enterprise decision management cult icon James Taylor talks about the revival of personalization.
Adelino de Almeida shares his thoughts about Starbucks and their television marketing strategy.
Sandro Saitta reviews the book, Web Dragons, a book about search marketing.
Targeted B2B marketing is one of the topics that Jeff Larche discusses this month.
RRW Consulting shares a retention case study that's worth reading.
Alison Bolen shares a few comments from a recent interview that SAS founder Jim Goodnight gave.
Customer U talks about no-cost customer research tools.
Michael Fassnacht writes about a presentation he watched on the "Read/Write" culture (I also watched this presentation, it's a good use of 20 minutes).
Sandeep Giri talks about the polarization of business intelligence solutions.
Jonathan Starets co-created a campaign that was a top three finisher for DMA Marketer of the Year.
Rick Whittington has a series on Holiday Shopping Tips. Here's tip number one, put best selling items on the homepage.
Anthony Power mentions that last month's DMA conference lacked buzz.
Gretchen Scheiman links to an article that discusses the history of e-mail.
Pat LaPointe doesn't think Google is about to dominate marketing dashboards.
UK-based Catalogue/E-Business allows readers to ask questions of direct marketing experts.
On the F. Curtis Barry blog, there's a discussion about how contact centers handle e-mail.
Harry Joiner talks about resume spam.
Ann Handley at Marketing Profs now offers snack-sized newsletters.
Mack Collier offers his readers a veritable plethora of link love to his Marketing Top 25.
Damien Francois writes about intelligent machines over at his blog.
Don Libey shares his thoughts about mergers and acquisitions with DMNews.
November 22, 2007
Cyber Monday: Whoopie!
Having digested my turkey dinner, it is time to focus on where commerce is headed over the next few weeks.
Grizzled direct marketing veterans spent portions of the past ten decades forecasting sales volume between Thanksgiving and Christmas. It was an important job, because you had to have just the right staffing level to handle the call volume on "peak day". Every cataloger had to predict the day that would be the "peak day" between Thanksgiving and Christmas.
And then the internet channel came along. Suddenly, everything needed to be "re-discovered". You didn't walk down the hall to talk to the catalog forecasting manager to identify which day would be "peak day".
Nope, there would be a new term, called "Cyber Monday". It would be the Monday after "Black Friday", with Black Friday being the day when you woke at 2:00am so that you could stand in line for the opening of the Kohls store at 4:00am so that you could enjoy the myriad of benefits associated with a vintage "doorbuster" promotion.
Some people tried to do good with the term.
But for many others, it represents an opportunity to get attention. "Cyber Monday is the First Monday in December", or "Cyber Monday is the last Monday before Christmas", or "There are actually four Cyber Mondays", or "You better have discounts or promotions or you'll be left behind on Cyber Monday", or "You'll be dead without an e-mail strategy for Cyber Monday".
Most assuredly, you still have a peak day, dependent upon consumer demand and your catalog/e-mail/paid search/organic traffic/free shipping/last day for expedited shipping strategy.
Somebody in your company knows what day this is. Go ask that person what day really matters in your company.
Grizzled direct marketing veterans spent portions of the past ten decades forecasting sales volume between Thanksgiving and Christmas. It was an important job, because you had to have just the right staffing level to handle the call volume on "peak day". Every cataloger had to predict the day that would be the "peak day" between Thanksgiving and Christmas.
And then the internet channel came along. Suddenly, everything needed to be "re-discovered". You didn't walk down the hall to talk to the catalog forecasting manager to identify which day would be "peak day".
Nope, there would be a new term, called "Cyber Monday". It would be the Monday after "Black Friday", with Black Friday being the day when you woke at 2:00am so that you could stand in line for the opening of the Kohls store at 4:00am so that you could enjoy the myriad of benefits associated with a vintage "doorbuster" promotion.
Some people tried to do good with the term.
But for many others, it represents an opportunity to get attention. "Cyber Monday is the First Monday in December", or "Cyber Monday is the last Monday before Christmas", or "There are actually four Cyber Mondays", or "You better have discounts or promotions or you'll be left behind on Cyber Monday", or "You'll be dead without an e-mail strategy for Cyber Monday".
Most assuredly, you still have a peak day, dependent upon consumer demand and your catalog/e-mail/paid search/organic traffic/free shipping/last day for expedited shipping strategy.
Somebody in your company knows what day this is. Go ask that person what day really matters in your company.
November 21, 2007
Theoretical Catalog Circulation And E-Mail Question
Catalog circulation is a beautiful blend of art and science.
When you send a catalog that costs $0.80, you must generate a suitable return on investment. In other words, short-term plus long-term profit must exceed this huge cost threshold, or you cannot afford to mail the catalog. Out of necessity, the catalog circulation manager must be scientific, must be rigorous, must focus on every tiny detail.
E-Mail marketers never had to deal with this challenge, given that the discipline, on a variable cost basis, is close to free. The cost structure encourages vastly different behavior. The e-mail marketer can be much less rigorous, and yet be far more successful than the catalog circulation expert.
Take the example of an e-mail opt-in subscriber who failed to interact with any of the past forty weekly e-mail campaigns your company chose to send to this customer.
E-mail marketers, please make a case for whether this customer should or should not receive next week's e-mail campaign (which has essentially no variable cost associated with it), given that the customer chose to ignore the past forty campaigns, and that the customer gave you no inclination that the customer does not want to see future campaigns.
When you send a catalog that costs $0.80, you must generate a suitable return on investment. In other words, short-term plus long-term profit must exceed this huge cost threshold, or you cannot afford to mail the catalog. Out of necessity, the catalog circulation manager must be scientific, must be rigorous, must focus on every tiny detail.
E-Mail marketers never had to deal with this challenge, given that the discipline, on a variable cost basis, is close to free. The cost structure encourages vastly different behavior. The e-mail marketer can be much less rigorous, and yet be far more successful than the catalog circulation expert.
Take the example of an e-mail opt-in subscriber who failed to interact with any of the past forty weekly e-mail campaigns your company chose to send to this customer.
E-mail marketers, please make a case for whether this customer should or should not receive next week's e-mail campaign (which has essentially no variable cost associated with it), given that the customer chose to ignore the past forty campaigns, and that the customer gave you no inclination that the customer does not want to see future campaigns.
November 20, 2007
Fast Forwarding
A decade ago, I spent considerable time reading about what the folks at the Santa Fe Institute were doing.
These folks take advanced mathematics to a whole new level (compared to mere mortals like you or I).
I became acquainted with a pair of individuals at the Santa Fe Institute who were working on retail simulations. They developed a computer model that illustrated how customers walked through a store, with the trip culminating in a purchase, or the customer leaving the store. The simulation depicted dozens or even hundreds of simulated shoppers walking a simulated store.
Within the simulation, you could see what happened if the line at the checkout was too long, or if merchandise was sold-out in a department, or if the store was too crowded.
I asked these individuals to visit us at Eddie Bauer. I invited a team of peers and leaders to the presentation. The simulation developers shared the tool with the folks I invited to the meeting, using verbal language associated with the simulation software tool.
Eddie Bauer employees stared at the tool. One person said something like "How would you apply this tool to our business?" The meeting ended. There were no questions. I thanked the folks from the Santa Fe Institute for their time.
We never mentioned the meeting, or software tool, again.
When we watch television on our DVR, we skip commercials. We seldom skip all the parts between the beginning of a show, and the end of a show. What happens in-between is important to the conclusion of the show.
Similarly in business, we cannot fast forward our co-workers from where they are to where we want for them to be. Our job gets even harder when we have to fast forward ten employees, or a thousand employees, or fifty thousand employees. Sometimes we expedite the process, we "trump" individuals by having a leader point the organization in a certain direction. But at a grass roots level, fast forwarding is hard work.
Do you have examples of concepts you tried to evangelize in your business, only to find that the business wasn't ready to embrace your ideas? What did you do when you ran into obstacles?
These folks take advanced mathematics to a whole new level (compared to mere mortals like you or I).
I became acquainted with a pair of individuals at the Santa Fe Institute who were working on retail simulations. They developed a computer model that illustrated how customers walked through a store, with the trip culminating in a purchase, or the customer leaving the store. The simulation depicted dozens or even hundreds of simulated shoppers walking a simulated store.
Within the simulation, you could see what happened if the line at the checkout was too long, or if merchandise was sold-out in a department, or if the store was too crowded.
I asked these individuals to visit us at Eddie Bauer. I invited a team of peers and leaders to the presentation. The simulation developers shared the tool with the folks I invited to the meeting, using verbal language associated with the simulation software tool.
Eddie Bauer employees stared at the tool. One person said something like "How would you apply this tool to our business?" The meeting ended. There were no questions. I thanked the folks from the Santa Fe Institute for their time.
We never mentioned the meeting, or software tool, again.
When we watch television on our DVR, we skip commercials. We seldom skip all the parts between the beginning of a show, and the end of a show. What happens in-between is important to the conclusion of the show.
Similarly in business, we cannot fast forward our co-workers from where they are to where we want for them to be. Our job gets even harder when we have to fast forward ten employees, or a thousand employees, or fifty thousand employees. Sometimes we expedite the process, we "trump" individuals by having a leader point the organization in a certain direction. But at a grass roots level, fast forwarding is hard work.
Do you have examples of concepts you tried to evangelize in your business, only to find that the business wasn't ready to embrace your ideas? What did you do when you ran into obstacles?
November 19, 2007
Offer Your Suggestions, And You Might Win A Free Book On Multichannel Forensics
More than a thousand of you read this blog on a daily basis.
Why?
I'm looking for a little feedback.
If you're willing to send me an e-mail or leave a comment, share why you read this blog, your company category (client or vendor), industry (catalog/multichannel, retail/multichannel, online pureplay, e-mail, search, web analytics, business intelligence), and what you'd like to see me talk about in the future. Criticisms are welcomed, feel free to tell me what you're sick of!
You can leave an anonymous comment.
If you choose to identify yourself (probably best done via e-mail), you'll be entered into a drawing for a free copy of my new book, "Hillstrom's Multichannel Forensics". That's a $95 value, folks!
Thank you in advance for taking a moment to help me understand how to provide you with a better experience.
Thanks,
Kevin
Why?
I'm looking for a little feedback.
If you're willing to send me an e-mail or leave a comment, share why you read this blog, your company category (client or vendor), industry (catalog/multichannel, retail/multichannel, online pureplay, e-mail, search, web analytics, business intelligence), and what you'd like to see me talk about in the future. Criticisms are welcomed, feel free to tell me what you're sick of!
You can leave an anonymous comment.
If you choose to identify yourself (probably best done via e-mail), you'll be entered into a drawing for a free copy of my new book, "Hillstrom's Multichannel Forensics". That's a $95 value, folks!
Thank you in advance for taking a moment to help me understand how to provide you with a better experience.
Thanks,
Kevin
Hologram Marketing On The Horizon
We're evolving toward Hologram Marketing at Darwinian speed:
- Brookstone 3-D Online Store.
- From Alan Rimm-Kaufman --- sku-level videos.
Co-Ops, E-Mail Targeting, and Modified RFM
Long time reader "Ray" asked if the co-ops offer a comparable product to Modified RFM for e-mail targeting. That's a good question!
Many of my readers work at co-ops. Co-ops already use product-specific information to decide which catalogs customers receive. Co-op readers, do you have a comparable product to Modified RFM for e-mail targeting?
It would seem likely that the depth of customer information across companies would give co-ops significant competitive advantages over e-mail vendors who do not possess the depth of data held at co-ops.
Co-op readers, please provide your answer in the comments section of this post.
Many of my readers work at co-ops. Co-ops already use product-specific information to decide which catalogs customers receive. Co-op readers, do you have a comparable product to Modified RFM for e-mail targeting?
It would seem likely that the depth of customer information across companies would give co-ops significant competitive advantages over e-mail vendors who do not possess the depth of data held at co-ops.
Co-op readers, please provide your answer in the comments section of this post.
November 18, 2007
Emphasizing New Product In Modified RFM E-Mail Targeting
Here's another quirk. You have new product that you want to feature in your e-mail targeting strategy.
If the has never been sold, you probably cannot come up with a weighting score for this product. In this case, you make up manual rules for deciding who gets that version of the e-mail campaign.
But if the product has been available for a few months, you may get a decent weighting score. In these cases, you may need to artificially increase the weights, in order to allow enough customers to receive this version of the e-mail campaign (multiple versions on the same day, customer only receives one version on that day).
If the has never been sold, you probably cannot come up with a weighting score for this product. In this case, you make up manual rules for deciding who gets that version of the e-mail campaign.
But if the product has been available for a few months, you may get a decent weighting score. In these cases, you may need to artificially increase the weights, in order to allow enough customers to receive this version of the e-mail campaign (multiple versions on the same day, customer only receives one version on that day).
More On Modified RFM For E-Mail Targeting
Many wonder what to do when you have two potential versions of an e-mail campaign. Which version should a customer receive?
Companies loaded with analytical talent have interesting algorithms to make these decisions. Yesterday, we talked about a shortcut that gets us 80% of the benefit for about 5% of the work.
But what do you do when one version of an e-mail campaign is so much more productive than another? In other words, say you have a Mens and Womens version of an e-mail campaign, and the customer could receive either version, but the Mens version is much less productive (sales per e-mail) than the Womens version?
A shortcut is to evaluate the historical difference in productivity, and apply that to the "weighting" score from yesterday's post. In other words, if the Mens version performs at 65% the level of a Womens version, multiply your Mens weighting scheme by 65%.
Again, this is statistical blasphemy. But you don't work at a company where you have thirteen statisticians sitting around waiting for new and exciting challenges. You're lucky to have one good analyst, and the demands upon this person's time are many. So take the shortcut, and get 80% of the benefit for 5% of the work. And when you have the money and/or human resources to do e-mail targeting the right way, by all means, pursue the ideal answer.
Companies loaded with analytical talent have interesting algorithms to make these decisions. Yesterday, we talked about a shortcut that gets us 80% of the benefit for about 5% of the work.
But what do you do when one version of an e-mail campaign is so much more productive than another? In other words, say you have a Mens and Womens version of an e-mail campaign, and the customer could receive either version, but the Mens version is much less productive (sales per e-mail) than the Womens version?
A shortcut is to evaluate the historical difference in productivity, and apply that to the "weighting" score from yesterday's post. In other words, if the Mens version performs at 65% the level of a Womens version, multiply your Mens weighting scheme by 65%.
Again, this is statistical blasphemy. But you don't work at a company where you have thirteen statisticians sitting around waiting for new and exciting challenges. You're lucky to have one good analyst, and the demands upon this person's time are many. So take the shortcut, and get 80% of the benefit for 5% of the work. And when you have the money and/or human resources to do e-mail targeting the right way, by all means, pursue the ideal answer.
November 17, 2007
Modified RFM For E-Mail Targeting
RFM is great for targeting one catalog to one customer. However, RFM is tough to manage in a multichannel environment.
This becomes clear in e-mail targeting. Say you have a Mens version of an e-mail campaign, and a Womens version of an e-mail campaign --- a customer could receive either version on the same date. Use this customer as an example:
This is where we apply "Modified RFM".
Have your statistician build a regression model one time --- and use the "weights" or "coefficients" for your modified RFM scheme. I realize this is statistical blasphemy, however, we aren't managing clinical trials for cancer drugs, we're deciding which version of an e-mail campaign a customer receives.
Step 1: Pick a "dependent" variable for "Mens". I like to look at the past twelve months.
Step 2: Create a series of "independent" variables:
Step 4: Repeat Steps 1-3 for Womens merchandise.
Now, we can evaluate which version of an e-mail campaign a customer should receive. Let's look at our example:
For the Mens version of the e-mail campaign, the customer receives a "weight" of 160.
For the Womens version of the e-mail campaign, the customer receives a "weight" of 75.
So, you should send the customer the Mens version of the e-mail.
For your next campaign, you don't have to build models again --- remember, we're not trying to cure cancer, we're just figuring out which version of an e-mail campaign will improve response a bit. Just apply the same weights built in your prior modeling process, and decide who gets which version.
The key here is to not build separate RFM schemes. Instead, you build variables in your database that summarize purchases by 0-3 month, 4-6 month, 7-12 month, 13-24 month, and 25+ month time periods. Then you "weight" those purchases based on importance. This gives you a good targeting strategy.
Statistical purists will blast me for misuse of appropriate statistical techniques. That's fine. We're just trying improve e-mail marketing performance, while minimizing use of internal resources, or minimize expense incurred when hiring consulting statisticians. This gets you 80% of the benefit for about 5% of the work.
This becomes clear in e-mail targeting. Say you have a Mens version of an e-mail campaign, and a Womens version of an e-mail campaign --- a customer could receive either version on the same date. Use this customer as an example:
- Customer spent $100 on Mens merchandise in the past three months.
- This customer also spent $200 on Womens merchandise 7-12 months ago, and spent $100 on Womens merchandise 13-24 months ago.
This is where we apply "Modified RFM".
Have your statistician build a regression model one time --- and use the "weights" or "coefficients" for your modified RFM scheme. I realize this is statistical blasphemy, however, we aren't managing clinical trials for cancer drugs, we're deciding which version of an e-mail campaign a customer receives.
Step 1: Pick a "dependent" variable for "Mens". I like to look at the past twelve months.
Step 2: Create a series of "independent" variables:
- Dollars spent on Mens in past three months (prior to the dependent time period).
- Dollars spent on Mens 4-6 months ago (prior to the dependent time period).
- Dollars spent on Mens 7-12 months ago.
- Dollars spent on Mens 13-24 months ago.
- Dollars spent on Mens 25+ months ago.
Step 4: Repeat Steps 1-3 for Womens merchandise.
Now, we can evaluate which version of an e-mail campaign a customer should receive. Let's look at our example:
| E-Mail Targeting Strategy: Mens Weights | |||
| Spend | Factor | Weight | |
| 00 to 03 Months | $100.00 | 1.600 | 160.0 |
| 04 to 06 Months | $0.00 | 0.600 | 0.0 |
| 07 to 12 Months | $0.00 | 0.300 | 0.0 |
| 13 to 24 Months | $0.00 | 0.150 | 0.0 |
| 25 to 99 Months | $0.00 | 0.050 | 0.0 |
| Total Weight | 160.0 | ||
| E-Mail Targeting Strategy: Womens Weights | |||
| Spend | Factor | Weight | |
| 00 to 03 Months | $0.00 | 1.600 | 0.0 |
| 04 to 06 Months | $0.00 | 0.600 | 0.0 |
| 07 to 12 Months | $200.00 | 0.300 | 60.0 |
| 13 to 24 Months | $100.00 | 0.150 | 15.0 |
| 25 to 99 Months | $0.00 | 0.050 | 0.0 |
| Total Weight | 75.0 | ||
For the Mens version of the e-mail campaign, the customer receives a "weight" of 160.
For the Womens version of the e-mail campaign, the customer receives a "weight" of 75.
So, you should send the customer the Mens version of the e-mail.
For your next campaign, you don't have to build models again --- remember, we're not trying to cure cancer, we're just figuring out which version of an e-mail campaign will improve response a bit. Just apply the same weights built in your prior modeling process, and decide who gets which version.
The key here is to not build separate RFM schemes. Instead, you build variables in your database that summarize purchases by 0-3 month, 4-6 month, 7-12 month, 13-24 month, and 25+ month time periods. Then you "weight" those purchases based on importance. This gives you a good targeting strategy.
Statistical purists will blast me for misuse of appropriate statistical techniques. That's fine. We're just trying improve e-mail marketing performance, while minimizing use of internal resources, or minimize expense incurred when hiring consulting statisticians. This gets you 80% of the benefit for about 5% of the work.
November 15, 2007
Yea, but ...
During my first year of consulting, the most frequently asked question is this:
... "What happened to customer behavior, sales, and profit, when you quit executing your traditional catalog marketing program at Nordstrom?"
At conferences, via e-mail, phone calls, lunch appointments, or client visits, I take the time to explain what happened.
And then I wait for the two most common words I hear in response to my experiences.
"Yea, but ..."
These two simple words are used to discount what was communicated, to suggest that one's opinion matters more than actual experiences held by another.
If you're an executive, try something different next week (it's a short week with Thanksgiving and all). Try going the whole week without saying the words "yea, but".
... "What happened to customer behavior, sales, and profit, when you quit executing your traditional catalog marketing program at Nordstrom?"
At conferences, via e-mail, phone calls, lunch appointments, or client visits, I take the time to explain what happened.
And then I wait for the two most common words I hear in response to my experiences.
"Yea, but ..."
These two simple words are used to discount what was communicated, to suggest that one's opinion matters more than actual experiences held by another.
- "Yea, but you grew your website without a traditional catalog program because you had 100 stores to support the website."
- "Yea, but that won't work for my brand because my customers love my catalogs, whereas your catalogs looked tacky to me".
- "Yea, but I've read that multichannel customers are the best customers, so you have to have a catalog".
- "Yea, but did you stop and think about how you spammed your customers by e-mailing them twice a week? Long-term, you can't get away with that."
- "Yea, but that won't work for my brand because Abacus/Experian told me that 83% of my online orders came from customers recently mailed catalogs".
- "Yea, but that was 2005-2006, and customers behave differently today."
- "Yea, but that won't work for my brand because my merchandising assortment is identical, regardless of channel."
- "Yea, but that won't work for my brand because my customer is 64 years old, and doesn't trust the internet."
- "Yea, but online marketing doesn't work unless you're an online pureplay offering free shipping."
If you're an executive, try something different next week (it's a short week with Thanksgiving and all). Try going the whole week without saying the words "yea, but".
Subscribe to:
Posts (Atom)
Upsets
On Saturday night, long after most of you went to bed, New Mexico scored what would become a game-winning touchdown with twenty-one seconds ...
