October 11, 2010

Digital Profiles Part N: Statistical Modeling

Alert: This post is for statistical computer geeks ... people like me!

Digital Profiles are an absolutely perfect complement to the statistical modeling process.

In this simple example, I am using Logistic Regression to predict the likelihood of a customer purchasing in the next month, given how much the customer spent in the past year (divided by 1,000) and the Digital Profile that the customer belonged to.

Fifteen of the sixteen digital profiles are listed ... the sixteenth digital profile (Rotary Phone) is assumed to have a value of zero.

Three Digital Profiles have coefficients greater than one, suggesting that they significantly contribute to future response (Gold Mine!, Multi-Channel Mavens, Retail Fanatics).

Digital Profiles with a negative coefficient are those that deflate future response.

I use Digital Profiles in my Multichannel Forensics projects, as well as my Online Marketing Simulations.
  1. Digital Profiles are combined with historical recency and historical spend by year to predict the probability of a customer purchasing.
  2. Digital Profiles can be used to predict how much a customer will spend if the customer chooses to purchase again.
  3. Digital Profiles are THE MOST IMPORTANT DRIVER in calculating the "ORGANIC PERCENTAGE", the percentage of future demand that will happen catalogs are no longer mailed to customers. Digital Profiles do a very nice job, especially when your organic percentage is close to fifty percent.
It's been my experience that, after recency/demand, Digital Profiles are the most important predictor of future response, future spend, and future organic demand.

October 10, 2010

Dear Catalog CEOs: Marketing Profit and Loss Statement

Dear Catalog CEOs:

I am assuming that you already ask your Marketing VP to produce for you a Profit and Loss Statement, measuring the effectiveness of all marketing activities.

You do ask for this, right?

Here's what the Marketing Profit and Loss Statement looks like:

Marketing Profit and Loss Statement



Net Sales Expense Profit Profit %
Catalog Marketing $40,000,000 $12,000,000 $2,800,000 7.0%
E-Mail Marketing $15,000,000 $500,000 $5,050,000 33.7%
Search Marketing $10,000,000 $3,000,000 $700,000 7.0%
Affiliate Marketing $2,000,000 $160,000 $580,000 29.0%
Social Media $150,000 $10,000 $45,500 30.3%
Mobile Marketing $150,000 $100,000 ($44,500) -29.7%
Other Online Marketing $1,000,000 $450,000 ($80,000) -8.0%
Organic Demand $31,700,000 $0 $11,729,000 37.0%
Totals $100,000,000 $16,220,000 $20,780,000 20.8%
Fixed Costs
$10,000,000

Earnings Before Taxes

$10,780,000 10.8%

This is a healthy, 10% pre-tax profit business. But take a look at where profit comes from. Of the $20,780,000 variable operating profit generated by this business (before subtracting fixed costs), different marketing channels deliver different levels of profit, rank-ordered from best to worst.
  • Organic Demand = $11,729,000.
  • E-Mail = $5,050,000.
  • Catalog Marketing = $2,800,000.
  • Search Marketing = $700,000.
  • Affiliate Marketing = $580,000.
  • Social Media = $45,000.
  • Mobile = ($44,500).
  • All Other Online Marketing = ($80,000).
Good golly, Miss Molly!

Those of us who measure the "organic percentage" (you know your organic percentage, right? RIGHT?!) can easily calculate the amount of sales generated without the aid of marketing. In this case, only 32% of total company demand is organic, but it accounts for more than half of company profitability!

And take a look at e-mail marketing. So few marketers ever parse the profit and loss statement by marketing channel. If you do this, be prepared to realize that e-mail marketing delivers a disproportionate amount of profit.

Now go think about the number of human beings you have working on catalogs, and the number of human beings working on e-mail marketing. Tally up the counts, and be prepared to be shocked!

This is a typical catalog profit and loss statement, by marketing channel. This is not unusual. Notice that the catalog only contributes $2.8 million of the $20.8 million in variable operating profit generated by this business.

Run the numbers, my friends. Be prepared to be shocked.

And if you don't have the right information to do this, hire me to calculate your organic percentage!

October 06, 2010

Digital Profiles Part M: Search

Is Search something that is embraced by your best customers, or is Search something that helps you attract new customers?

A surprisingly small number of marketers actually know the answer to this question.

Why not use Digital Profiles to help explain the phenomenon known as "Search"?

In this example, we freeze our customer file at the end of the month, we assign every customer a Digital Profile based on activity during the past twelve months, and then we count how many Search buyers in the next month are sourced by each Digital Profile.

Here's the results:
  • Gold Mine! = 104 customers.
  • Multi-Channel Mavens = 86 customers.
  • The Digerati = 32 customers.
  • A Long Drive = 70 customers.
  • Shop Online, Buy In Store = 46 customers.
  • Direct Newbies Looking For Eight = 40 customers.
  • Online One/Twos = 58 customers.
  • Ship It To Me = 32 customers.
  • Retail Fanatics = 7 customers.
  • Retail 4/5/6 = 15 customers.
  • Retail 1/2/4 = 7 customers.
  • In And Out = 17 customers.
  • Retail 1/2/3 = 3 customers.
  • Retail Newbies Looking For Eight = 6 customers.
  • Retail 1/2 = 7 customers.
  • Rotary Phone = 2 customers.
  • Non-12-Month Buyers = 283 customers.
Oh oh.

In this case, Search is overwhelmingly preferred by last year's customer base ... about 2/3 of the customers purchasing via Search in the past 30 days were buyers in the past twelve months.

Do you know what this percentage is for your business?

If you don't know it, and you are responsible for Search marketing, why don't you know it?

Search is viewed in a different light when it is a tool that is used by your best customers ... and in this example, Search is clearly preferred by the high-value customer who shops all channels. It means that your good customers, the ones that you call "loyal", those customers are comparison shopping. Even worse, it can often be demonstrated that when this situation exists, your direct marketing activities (catalogs, e-mail in particular) cause your "loyal" customer to comparison shop your competition.

That doesn't sound like it's that much fun, does it?

This does highlight the importance of Search. Your Digital Profiles clearly indicate who is using Search, and you can quickly surmise what might happen if you didn't have a great Paid Search program, or if your Natural/Organic Search program was in the tank.

We usually think of Search as a way to "intercept" a customer ... in this example, Search is a way to "save" a customer! Use Digital Profiles to see if this is a situation you should pay attention to!

October 05, 2010

Digital Profiles Part L: Social Shoppers

There are two directions that marketers are taking with Social Media.

Direction #1 = Counting followers, engaging users, creating conversations.

Direction #2 = Looking for actionable outcomes that drive sales increases and profit improvement.

Having been a VP at an eight billion dollar a year apparel/shoe department store, I have a feeling that most Executives are looking for you to head down "Direction #2".

If that's the direction you're heading down, then you want to analyze what I call the "Social Shopper". This is a customer that purchases because of your social presence, or because of a social media reference from another individual. These customers are not difficult to track, heck, you have the referring URL information in your web analytics platform to do this ... it just takes a little work, doesn't it?

Once you've done the work, you can review your "Social Shoppers" via the Digital Profiles framework. Let's count the number of "Social Shoppers" in the past year, by Digital Profile:
  • Gold Mine! = 1,079 customers.
  • Multi-Channel Mavens = 683 customers.
  • The Digerati = 397 customers.
  • A Long Drive = 516 customers.
  • Shop Online, Buy In Store = 241 customers.
  • Direct Newbies Looking For Eight = 219 customers.
  • Online One/Twos = 294 customers.
  • Ship It To Me = 66 customers.
  • Retail Fanatics = 0 customers.
  • Retail 4/5/6 = 0 customers.
  • Retail 1/2/4 = 0 customers.
  • In And Out = 0 customers.
  • Retail 1/2/3 = 0 customers.
  • Retail Newbies Looking For Eight = 0 customers.
  • Retail 1/2 = 0 customers.
  • Rotary Phone = 0 customers.
Oh oh, what do we see here? There's those darn "Gold Mine!" customers once again, followed by "Multi-Channel Mavens". In other words, the Social Shopper, predictably, is an online individual (duh) with an affinity for shopping in stores. Ask Kohl's if that's the case!

We can look at where next month's social shoppers will come from ... here, we freeze Digital Profile assignments at the end of a month, and then in the next 30 days, we measure the probability of a customer purchasing via Social Shopping, by Digital Profile. Take a look!
  • Gold Mine! = 0.81% response rate.
  • Multi-Channel Mavens = 0.79% response rate.
  • The Digerati = 0.44% response rate.
  • A Long Drive = 0.39% response rate.
  • Shop Online, Buy In Store = 0.32% response rate.
  • Direct Newbies Looking For Eight = 0.26% response rate.
  • Online One/Twos = 0.26% response rate.
  • Ship It To Me = 0.16% response rate.
  • Retail Fanatics = 0.05% response rate.
  • Retail 4/5/6 = 0.05% response rate.
  • Retail 1/2/4 = 0.02% response rate.
  • In And Out = 0.05% response rate.
  • Retail 1/2/3 = 0.00% response rate.
  • Retail Newbies Looking For Eight = 0.01% response rate.
  • Retail 1/2 = 0.02% response rate.
  • Rotary Phone = 0.00% response rate.
In other words, the most responsive customers to "Social Shopping" are the customers with an online presence, coupled with an affinity for store purchases.

Here's another thing we can do. Let's dig into our "Gold Mine!" segment, and for the past three years, count the number of customers with a Social Shopping purchase, and the number of customers without a Social Shopping purchase.
  • 2008 = 8,773 with no Social Shopping, 576 with Social Shopping.
  • 2009 = 8,235 with no Social Shopping, 720 with Social Shopping.
  • 2010 = 7,544 with no Social Shopping, 1,079 with Social Shopping.
Notice that the "Gold Mine!" segment isn't growing ... in fact, it is shrinking. Just as important, it is being overrun with Social Shopping. It does appear that Social Shopping is becoming more important, but it isn't necessarily causing the overall customer base to become "more valuable", because if it caused the overall database to become more valuable, there would be more "Gold Mine!" customers, right?

Here's what the counts look like for "Multi-Channel Mavens" ... you'll see the same story here as well:
  • 2008 = 7,000 with no Social Shopping, 393 with Social Shopping.
  • 2009 = 6,399 with no Social Shopping, 506 with Social Shopping.
  • 2010 = 5,541 with no Social Shopping, 683 with Social Shopping.
If you are a marketer, responsible for growing your Social Shopping presence, you have a unique opportunity to see if your activities are truly improving the overall customer file. In this case, for this business, it appears that Social Shopping is inevitable, and is becoming a preferred channel for a minority of customers, but isn't causing more customers to be populated in the best Digital Profiles.

October 04, 2010

Digital Profiles Part K: E-Mail At A Crossroads

The e-mail community look at e-mail as the most important relationship-building tool that exists, one with the best ROI (when measured as a percentage of sales).

Can we look at e-mail within the context of Digital Profiles? Certainly!

Too often, we look at customers who possess an e-mail address. More important, of course, is the customer who purchases after clicking-through an e-mail campaign.

Let's count how many customers purchased via e-mail in the past year, by Digital Profile.
  • Gold Mine! = 3,194 customers.
  • Multi-Channel Mavens = 2,227 customers.
  • The Digerati = 1,447 customers.
  • A Long Drive = 2,637 customers.
  • Shop Online, Buy In Store = 1,114 customers.
  • Direct Newbies Looking For Eight = 1,040 customers.
  • Online One/Twos = 1,908 customers.
  • Ship It To Me = 671 customers.
  • Retail Fanatics = 1 customer.
  • Retail 4/5/6 = 0 customers.
  • Retail 1/2/4 = 0 customers.
  • In And Out = 0 customers.
  • Retail 1/2/3 = 0 customers.
  • Retail Newbies Looking For Eight = 0 customers.
  • Retail 1/2 = 0 customers.
  • Rotary Phone = 0 customers.
Look at that darn "Gold Mine!" segment ... once again, it is truly a Gold Mine ... about 20% of the customers who purchased via e-mail belong to this Digital Profile.

Now tell me, dear e-mail direct marketers, how does this change your perspective on e-mail marketing? Do you treat these customers the same way you treat a customer who has an e-mail address in the "Rotary Phone" Digital Profile, or do you treat them differently? Heck, you already know the answer!

Within the "Gold Mine!" and "Multi-Channel Mavens" Digital Profiles, your e-mail marketing message is likely to highlight products and benefits that apply across all channels.

Look at the "A Long Drive" Digital Profile ... that's where the second-most active e-mail buyers exist. If these customers don't live near a store, don't you think the marketing message should be calibrated to the online experience, satisfying the customer via convenience ... this customer doesn't have to drive a long distance to retail stores, you're bringing the store directly to this customer!

There are 1,114 e-mail buyers in the "Shop Online, Buy In Store" Digital Profile. The name tells you what the marketing message needs to be ... your job as an e-mail marketer is to stimulate this customer to get off her butt, get into a car, and visit a store!! It's the opposite message of the one provided to "A Long Drive", isn't it???

E-mail marketers frequently talk about segmenting customers, personalizing messages, using trigger-based campaigns, you name it. Why wouldn't you, if you are an e-mail marketer, apply Digital Profiles to your database, leveraging the results to your benefit, and to the benefit of the customer?

October 03, 2010

Dear Catalog CEOs: The Right Number of Contacts

Dear Catalog CEOs:

There are companies out there that you can pay a quarter million or half million dollars to, and they'll create a complex mathematical algorithm that allows you to determine the right number of contacts to send to a customer.

Or you can work with me, and we'll "keep it simple"!

Say you have a customer segment that receives twelve catalog mailings a year. The average cost of a catalog is $0.60. 37% of your demand "flows-through" to profit. The segment spends, on average, $30.00 a year. Your organic percentage (you actively track your organic percentage, right? RIGHT?!) is 40%.

With this information, I use a proprietary curve that estimates the "optimal" number of catalogs to send to a customer. Here we go:


Catalog Organic Total Total Total
Catalogs Demand Demand Demand Cost Profit
0 $0.00 $12.00 $12.00 $0.00 $4.56
1 $5.20 $12.00 $17.20 $0.60 $5.93
2 $7.35 $12.00 $19.35 $1.20 $6.15
3 $9.00 $12.00 $21.00 $1.80 $6.18
4 $10.39 $12.00 $22.39 $2.40 $6.11
5 $11.62 $12.00 $23.62 $3.00 $5.98
6 $12.73 $12.00 $24.73 $3.60 $5.80
7 $13.75 $12.00 $25.75 $4.20 $5.58
8 $14.70 $12.00 $26.70 $4.80 $5.34
9 $15.59 $12.00 $27.59 $5.40 $5.08
10 $16.43 $12.00 $28.43 $6.00 $4.80
11 $17.23 $12.00 $29.23 $6.60 $4.51
12 $18.00 $12.00 $30.00 $7.20 $4.20
13 $18.73 $12.00 $30.73 $7.80 $3.88
14 $19.44 $12.00 $31.44 $8.40 $3.55
15 $20.12 $12.00 $32.12 $9.00 $3.21

Wow, take a look at that! Now, at twelve mailings a year, the customer generates $4.20 profit, so naturally, you think you're getting profit in each of the twelve mailings.

My proprietary curves, created largely from data during my time at Lands' End, Eddie Bauer, and Nordstrom, tell us otherwise. Optimal profit is generated at just three catalog mailings.

Our measurement systems are simply not capable of truly measuring incremental profit. We break all of our activities down to individual marketing campaigns, we "attribute" or "matchback" orders to all marketing activities, and as a result, we completely ignore "organic demand", demand that is generated without marketing activities.

Yes, your customer will spend money, even if you don't market to the customer.

So we over-market to customers, given our poor marketing attribution/matchback systems, we spend way too much money advertising, money that actually delivers a loss.

Remember, the vendor ecosystem greatly benefits when you over-market to a customer.

And there are vendors that greatly benefit when you pay them $250,000 or $500,000 to solve this problem, when you could pay me eight percent of that total, and I'll solve the problem for.

Three questions for you, Dear Catalog CEO:
  1. Do you know your "organic percentage", the percentage of demand that will happen if you do not do any marketing?
  2. Do you know the optimal number of catalogs to mail to each segment of your housefile?
  3. Do you need to contact me to answer questions (1) and (2) for you? If so, click here!

Discount / Promotion Algebra You Need To Know

Many folks contacted me last week with information about Groupon and their presentation at Shop.org.

By the way, wouldn't it have been nice if Shop.org partnered with Groupon and offered you, the loyal marketing executive, a sweet 50% discount off of the Shop.org conference registration fee? ... and did I see that Brett Michaels has a keynote at the DMA conference ... wow ... maybe the DMA will partner with Groupon and give you a sweet 50% discount so that you can hear that one on the cheap).

Well, I wasn't in attendance, so I cannot vouch for this fact ... but I'm told that it was said that losses on Groupon transactions are "made up for" on the next purchase.

This "truthiness marketing math" concept can work. Let's work through an example:
  • You lose $3 profit on your Groupon transaction.
  • You make $10 profit on the next purchase a discount-oriented customer places.
We can use this really neat tool, called "algebra", to figure out the percentage of customers who need to purchase again so that you make money.
  • -$3*(1-x) + $10*(x) > 0.
  • -$3 + $3*x + $10*x > 0.
  • $13*x > $3.
  • x > (3/13).
  • x > 23.3%.
  • NOTE: The first posting had a rate of 42.9%, which was corrected by multiple readers.
So, the Groupon customer repurchase rate must be greater than 23.3% in order for truthiness marketing math to work.

Here's the problem, dear retail and e-commerce marketing experts:
  • It can be a challenge for a retailer or an e-commerce business to get 23.3% of first-time buyers or discount buyer to repurchase.
Do the math yourself, folks. You know the amount of twelve-month profit a new discount-oriented customer delivers to your business ... right ... you do know this number ... right? So assuming you know this number, calculate the repurchase rate you need ... you do track repurchase rate of first time discount-oriented buyers ... right ... you do know this number ... right? Just do the math, and determine for yourself if this type of promotion is appropriate for your business.

If it is appropriate for your business, good for you!

But if it is not appropriate for your business, then question why the marketing punditocracy are shoving this business model down your throats lately ... stand up for yourselves and make sure that the marketing punditocracy share profitable ideas and tips!

An Ecommerce Brand Is Constantly Reinventing Itself

A year or two ago I received an email ... the Leader asked "what happened to all of those COVID customers, it's like they just disa...