Attribution Beyond Catalogs

Yesterday, I outlined the methodology I use to perform catalog attribution (click here please).  The article got good numbers, and generated questions.  This is the theme of the most commonly asked question.

Question:  You only care about catalogs.  We live in an omnichannel world.  How do I account for paid search, you moron?

Paid Search is the most complicated case, simply because we have to also ascertain the click-through rate within catalog / email driven searches.

First, we need to have both catalog and email holdout test results available.  You have email holdout test results available, right?  Right?  Because if you're going to do attribution work, you're going to apply real science, not just hokum-based guesses used by other practitioners.  Promise me that you have catalog and email holdout results.  If not, don't go further, you're just guessing, and that's really dangerous.

Here are the results from a sample catalog holdout test.

And here are the results from a sample email holdout test.

We only need a few additional pieces of information to identify the profitability of paid search.

First, in the catalog test, we learn that 50% of paid search demand is catalog-driven.  In other words, if we took catalog marketing away, 50% of paid search demand disappears.  Therefore, half of paid search demand is immediately allocated to catalog marketing.

Second, in the email test, we learn that 10% of paid search demand is email-driven.  In other words, if we took email marketing away, 10% of paid search demand disappears.  Therefore, 10% of paid search demand is immediately allocated to email marketing.

Let's say that our paid search program possesses the following metrics.

• Total Budget = $1,000,000.
• Total Clicks = 2,000,000.
• Conversion Rate = 2%.
• Average Order Value = $100.
• Total Demand = $4,000,000.
• Flow-Through Rate to Profit = 40%.
• Total Profit = $4,000,000 * 0.40 - $1,000,000 = $600,000.

Here's where things get a little bit messy.  You need to know the conversion rate of clicks attributed to catalog marketing, and to email marketing.  Few people possess this knowledge.  Go talk to your analytical gurus, vendors, or Google Analytics ninjas, and have them derive this number for you.

Let's pretend we know this number.

• Catalogs and Email Paid Search Conversion Rate = 2.5%.

Ok, we're making progress now.  Let's calculate the conversion rate for non-catalog and non-email clicks.  First, we know that 60% of paid search demand is caused by catalog and email marketing.  So we subtract that out of the equation.

• $4,000,000 * (1 - 0.60) = $1,600,000.

In other words, $2,400,000 paid search demand is catalog and email driven.  We know that the average order value is $100, we know that the conversion rate is 2.5%.  Therefore, we can calculate catalog/email driven clicks:

• $2,400,000 demand / $100 AOV = 24,000 orders.
• 24,000 orders / 0.025 = 960,000 clicks.

If 960,000 clicks are catalog/email driven, then 1,040,000 have to be paid search driven.

• 2,000,000 - 960,000 = 1,040,000.

Let's run the profit and loss statement for catalog/email driven clicks.

• Total Clicks = 960,000.
• Total Budget = 960,000 * $0.50 = $480,000.
• Conversion Rate = 2.5%.
• Average Order Value = $100.
• Total Demand = $2,400,000.
• Flow-Through Rate to Profit = 40%.
• Total Profit = $2,400,000 * 0.40 - $480,000 = $480,000.

By simple subtraction, we can calculate the impact of paid search, outside of catalog/email.

• Total Clicks = 1,040,000.
• Total Budget = 1,040,000 * $0.50 = $520,000.
• Conversion Rate = (1,600,000 / $100) / 1,040,000 = 1.54%.
• Average Order Value = $100.
• Total Demand = $1,600,000.
• Flow-Through Rate to Profit = 40%.
• Total Profit = $1,600,000 * 0.40 - $520,000 = $120,000.

There you have it.  You just attributed paid search driven orders to catalog and email marketing, and you know what remains.  What remains is still profitable, though it converts at a much lower rate.

Now, you have yourself a dilly of a pickle here.  It's not terribly easy to identify a customer as a catalog/email driven visitor to Google, then make separate decisions based on that information.

Because of that, some of the attribution talk is nonsense.

Let's pretend that the non-catalog and non-email clicks were unprofitable.  You have to have a working relationship with Google that allows you to tell Google, at the time somebody visits Google, that the visitor is catalog or email driven - and if not catalog/email driven, don't pay for the click.

So if you can do that, then the attribution exercise is actionable.

If not, then the attribution exercise is done for knowledge, but is not actionable.

This process, of course, is repeated for all advertising channels.  Demand/Expense are allocated to catalogs and email marketing, with the remainder allocated to each individual marketing channel.

If the remaining marketing channels are shared, most would just allocate fractionally, based on pre-determined rules.  This, of course, is largely hokum, but there is a market where people are willing to pay for hokum-based research, so be it.

ACTION REQUIRED FOR CATALOG PROFESSIONALS: NAME YOUR FAVORITE MATCHBACK VENDOR

This is a picture of the Red Cross.  They are in the process of responding to a disaster.

If you had to attribute events that caused the Red Cross to be here, you'd have a challenge on your hands, right?  It would be easy to demonstrate that the disaster gets most of the credit.

It would be hard to show how individual contributions and corporate contributions led to this response.  It would be hard to show how the actions of a Red Cross volunteer or employee led to the purchase of the vehicles, which contributed to the response.

Similarly, in catalog marketing, we have to attribute response.  In the rest of the marketing world, this is called "attribution".  In catalog marketing, catalogers frequently take credit for actions that should be attributed to online marketing ... as a result, the methodology is called "matchback", not "attribution".

There are many matchback vendors.  Often, I'm asked to comment on the strengths and weaknesses of matchback vendors.

Instead of my opinions, let's focus on your thoughts.

In the comments section, please leave an anonymous comment.  Here's what I want you to share.

1. Name of the Matchback vendor you use.
2. Grade you would give their services ... "A", "B", "C", "D", or "F".
3. Tell us what your matchback vendor does best.
4. Any additional considerations.

If you wish to put your name behind the comments, go ahead.  If not, please leave an anonymous comment, and give your thoughts to the catalog community.  Via the magic of search engine optimization, this can become the post that folks look to for your thoughts about matchback vendors.

Attribute This!

Maybe you watched the Kennedy Center Honors event earlier this week.  If you did, you witnessed this performance of Led Zeppelin's "Stairway To Heaven" ... performed by Ann and Nancy Wilson.

You had a band, an orchestra, a choir, and Ann Wilson belting out vocals.  By most accounts, this was a spectacular performance, if you like music from this genre.

If you work in my industry, you'd be asked to parse this performance, attributing the reasons why it was successful.  Based on the elements that were successful, you'd be asked to invest more in the areas that yielded success, less in other areas.

1. What percentage of success do you attribute to Ann and Nancy Wilson?
2. What percentage of success do you attribute to the living members of Led Zeppelin being in attendance, being honored, enjoying a version of a song they created, what, 35 years ago?
3. What percentage of success do you attribute to the choir?
4. What percentage of success do you attribute to the orchestra?
5. What percentage of success do you attribute to the band?
6. What percentage of success do you attribute to the fact that the drummer passed away years ago, leaving his son to play drums in this performance?
7. What percentage of success do you attribute to Jimmy Page, at the 3:58 mark of the video, checking to see that the guitar solo is played properly?  Or to the soul playing the guitar at the 3:58 mark of the video?
8. What percentage of success do you attribute to the lighting?
9. What, specifically, caused Robert Plant to have tears in his eyes?

It's funny.  When it comes to content / product / merchandise, we'd never think of tearing the thing into bite (or byte) sized bits, would we?  The whole is greater than the sum of the parts.  We fully accept that you don't separate the choir from the orchestra and assign value to each.

In marketing, we want to reverse engineer everything, parsing the parts so we can sum things up on a scorecard, regardless how accurate or inaccurate we are.

Just a little something to think about as we head into the New Year.

Why Attribution Efforts Fail Miserably: A Basketball Example

I want to walk you through a series of images from a basketball game.  The images will explain why attribution efforts fail miserably.

Look at the first image (below).  The player in the middle of the court just secured the rebound of a missed shot (the number on his jersey is "4").  He is passing the ball to the point guard (the number on his jersey is "0").

Now, we move a few seconds ahead in the action.  I positioned two arrows on the two players we are following.  The player with the ball (number "0") dribbled to the top of the key.  Next, he will pass the ball to the right wing, where his teammate is wide open.  But more important, look at the left arrow.  This is player number "4".  He has hustled down court, passing all of his teammates.  Because of his hustle, he had drawn two defenders near him.  And because he drew two defenders near him, his teammate (number "12") is wide open on the right wing.

Ok, we'll look at one more image, at the end of the play.

Because player #4 drew a defender close to him, his teammate on the right wing (#12) was wide open, and received a pass from player #0.  A defender rushed to the teammate (#12), blowing right past a shot fake.  Now, the teammate (#12) has a wide open three point jump shot attempt, which he buried!  The crowd went crazy, because this player hit a wide open shot that counted for three points, he's credited with one more point than he would be credited for on easier shots taken inside the three point arc.

But look at our player, player #4 ... he is just above the black arrow on the image.  If the shot is missed, player #4 has hustled into position, and has blocked out the defender.  Most long shots rebound to the opposite side the ball is shot from.  In other words, player #4 is in good position to get the rebound if the shot is missed ... and if he doesn't get the rebound, he has a defender blocked out, so that his teammate (to his left, #35) will get the rebound.

In the box score (basketball's version of a KPI attribution dashboard), here is what is documented from this offensive series:

• Player #0 = 1 assist.
• Player #12 = 3 points.
• Player #4 = No credit for anything.

This is exactly the same situation we observe in all of our attribution/matchback work.  We'd give player #12 credit for closing the deal (last touch attribution).  We'd give player #0 credit for an assist (multi-touch attribution).  We'd give player #4 no credit for anything other than a defensive rebound.

Player #4, who is truly the reason for the success of this offensive series, gets no credit.  In large part, he gets no credit because we don't have a measurement system to assign value to his efforts (hustle, positioning, basketball IQ, fundamentals).

Coaches, however, do assign value to his efforts.  They watch film, and can visually see his value to the team.  That's why he is playing in this game.

We, as a measurement community, will fail until we find a way to provide attribution to the activities that truly make a purchase happen (customer service, merchandise, creative).  Today, we give too much credit to channels, discounts, and promotions (i.e. player #12 and, to a lesser extent, player #4).  

We don't give credit to customer service, merchandise, and creative (i.e. player #4).

Your Opinion Wanted: Attribution

Please use the comments section to offer your opinion as to how to attribute this order properly --- I am not judging whether any answer is right or wrong --- I am simply seeking opinions:

1. Customer receives a catalog on October 1.
2. Customer receives email campaign on October 2.
3. Customer receives email campaign on October 4.
4. Customer visits site on October 5 via paid search, branded term.
5. Customer visits your mobile website on an iPad on October 6 via affiliate website, purchases item featured in October 1 catalog, uses free shipping promo code from affiliate website.

Describe the thought process you go through to make your attribution guess.  This is really what I am looking for --- your thought process.  So please provide that to our readers in the comments section, thanks.

Attribution Week: Your Website and Paid Search

Take a good, hard look at the "All Other Online Demand" row in this table.


Which column has the best performance for the online channel?


The website is pulled in two opposite directions.

1. Catalogs cause website demand to happen.
2. Emails significantly cannibalize website demand.

Pure website demand is the most elastic of any channel.  Some forms of marketing drive a customer to a website, while other forms of marketing drive the customer away from the website.

It is easy to understand the impact of Catalogs and Email Marketing on website demand.  All we have to do is execute mail/holdout tests.

In Paid Search, it's important to geo-target.  You see, we want to measure the impact that Paid Search has on all other channels.  To this point, we've described the impact that Catalogs and Email Marketing have on Paid Search.

You can set up a brief test ... geo-target specific areas, then execute normal Paid Search activities vs. minimal Paid Search spend, and overlay this with Email mail/holdout test results.

This will clearly tell you how important Paid Search is to all other channels.

Ok, given what you've learned this week (and it is a lot), answer these questions.

1. Do you currently have this level of business intelligence at your company?
2. If the answer is "no", are you willing to execute the tests necessary to answer these questions?
3. If the answer to both questions is "no", please describe in the comments section why you don't think you need to answer these questions?  Go!!

Attribution Week: Channel Impact

This is our table from earlier in the week, measured via a 90 day long mail/holdout test of catalog marketing and email marketing.


There are many nuggets of information in the table.  Let's summarize how each channel is impacted by catalog marketing and email marketing.


Paid Search:  We generate $2.00 per customer via Paid Search in the quarter.

• Catalog causes $0.72, or 36% of Paid Search volume to happen.
• Email causes $0.18, or 9% of Paid Search volume to happen.
• 55% of Paid Search demand is independent of Catalog or Email marketing.
• We have to attribute 36% of Paid Search volume/cost back to catalog marketing, and we have to attribute 9% of Paid Search volume/cost back to email marketing.

Natural Search:  We generate $2.00 per customer via Natural Search.

• Catalog causes $0.58, or 29% of Natural Search volume to happen.
• Email causes $0.18, or 9% of Natural Search volume to happen.
• 62% of Natural Search demand is independent of Catalog or Email marketing.

Affiliates:  We generate $0.50 per customer via Affiliates.

• Catalog causes $0.34, or 68% of Affiliate demand to happen.
• Email causes $0.10, or 20% of Affiliate demand to happen.
• 12% of Affiliate demand is independent of Catalog or Email marketing.
• We have to attribute 88% of Affiliate marketing costs back to catalogs or email campaigns.
• Since Affiliate marketing is directly tied to Catalogs/Email, we need to strongly consider if the demand will continue to happen if Affiliates are dropped?

Display:  We generate $0.25 per customer via Display:

• Catalog causes $0.15 per customer of Display demand to happen (60%).
• Email causes $0.01 per customer of Display demand to happen (4%).
• 36% of Display demand is not driven by other marketing activities (36%).
• We have to attribute 60% of Display costs back to the catalogs that caused the demand to happen.

Social Media:  We generate $0.15 per customer via Social Media.

• Catalog causes $0.01 per customer of Social Media demand to happen (7%).
• Email causes $0.04 per customer of Social Media demand to happen (27%).
• 66% of Social Media demand happens independent of Catalog/Email.
• Given the strong link between Email and Social Media, there should be some level of integration between the two channels.  Social Media, to some extent, depends upon a strong email marketing program.

Tablets:  We generate $0.65 per customer on Tablet devices.

• Catalogs cannibalize $0.06 of demand from Tablet devices.
• Email cannibalizes $0.04 of demand from Tablet devices.
• In other words, when Catalogs/Emails are sent, customers shift their focus away from Tablet devices, and instead spend money in other channels.  When Catalog/Email activities are not happening, customers shift their attention back to Tablets.  This strongly suggests that, in the future, when customers spend more and more time with Tablet devices, this business will be able to scale back a bit on Catalogs and Email marketing campaigns, because the demand will be recaptured by Tablet devices.

Mobile:  We generate $0.15 per customer on Mobile devices.

• Catalogs cannibalize $0.03 of demand from Mobile devices.
• Email has no impact on Mobile devices.
• In other words, when Catalogs are sent, customers shift their focus away from Mobile devices, and instead spend money in channels congruent with Catalogs.  At this time, Mobile is a very tiny portion of the total story.  In the future, we need to watch this relationship, to see if it continues.  If the relationship continues, the data strongly suggest that we can cut back a bit on Catalogs, because demand will reallocate to Mobile devices.

At minimum, the attribution process should include three things.

1. Mail/Holdout Testing, where applicable.  For catalog marketing and email marketing, this couldn't be easier.  For Display Ads, this couldn't be easier.  For Paid Search, you can vary your budget and compare the impact.
2. Modification of Ad Cost by Channel.  In our example, paid search costs should be moved from the paid search budget to catalog marketing, and to email marketing.  Too often, we don't focus on this aspect of attribution.  Mail/Holdout tests illustrate why we need to do this.
3. Business Intelligence.  The goal of any attribution project should not be to just allocate demand/expense, but should be to teach every employee how channels fit together.  We just don't do enough of this, do we?  How often do we step back, and try to teach employees how our business works?

Contact me (click here) for additional assistance with attribution work.

Attribution Week: Email Productivity

If you ask 100 email marketers the following question, how many do you think will answer "yes".

• Question for Email Marketers:  Do you measure the success of email marketing by executing holdout groups, not mailing customers email campaigns for up to three months at a time?

Ask 100 email marketers this question, and fewer than five will answer "yes".

What a shame.

You see, all of the ways that email marketing interacts with the rest of your business are illustrated by email holdout tests.

Take a look at our example:

Here's the deal.  Email marketers typically measure performance via opens/clicks/conversions, adding average order size to the strategy to yield demand per email delivered.  In our case, over the course of a three month period of time. we get $4.94 demand per customer ... across 26 email campaigns (2x per week by 13 weeks) ... each campaign is generating $0.19 for a total of $4.94.

We run a profit and loss statement.

Demand	$4.94
Net Sales	$4.20
Gross Margin	$2.31
Less Email Cost	$0.08
Less Pick/Pack/Ship	$0.42
Variable Profit	$1.81

That's some sweet action!

Except, of course, that this analysis is providing us with the wrong answer.

When we do not send email campaigns to customers, as measured via mail/holdout tests, we generate just $3.33 demand per customer, not the $4.94 as measured via opens/clicks/conversions.  We know this, because in our example, one set of customers did not receive a single email campaign for three months, and continued to generate incremental demand.

Here's what the profit and loss statement looks like, with our new and more accurate style of measurement.

Demand	$3.33
Net Sales	$2.83
Gross Margin	$1.56
Less Email Cost	$0.08
Less Pick/Pack/Ship	$0.28
Variable Profit	$1.20

Now, there's nothing wrong with $1.20 profit per customer ... but it is a lot less than $1.81 per customer, right?

Let's take a look at what happens to other channels, when email marketing is discontinued for ninety days.

Paid and Natural Search results decrease, by 11% and 16% respectively.  In other words, email marketing causes searches to happen.  In theory, you have to take 11% of your paid search marketing cost, and allocate it back to your email marketing budget.  Yes, you have to do this, this is what attribution is all about.

Demand	$3.33
Net Sales	$2.83
Gross Margin	$1.56
Less Search+Email	$0.15
Less Pick/Pack/Ship	$0.28
Variable Profit	$1.13

Now, that doesn't make a difference in this case ... but when your email marketing program lacks a lot of productivity, well, it is enough to push the whole program under water.

Look at all other online marketing demand.  When you don't mail email campaigns to a customer, that customer changes behavior.  The customer re-directs demand away from email marketing, back to the website ... spending $6.57 per customer online instead of $4.35 per customer online.  In other words, customers are using email as a navigational tool to get to the website.

Look at affiliate marketing demand.  When emails are stopped, customers spend $0.24 each at affiliates ... but spend $0.34 each at affiliates when emails are delivered.  Now, the affiliate is probably getting a cut of each order, right?  Well, you have to attribute that cut on the $0.10 incremental difference, attributing that to email marketing, not to affiliates, because email marketing caused the affiliate order to happen.

Demand	$3.33
Net Sales	$2.83
Gross Margin	$1.56
Less Marketing	$0.16
Less Pick/Pack/Ship	$0.28
Variable Profit	$1.12

Display / Retargeting are not significantly impacted, are they?

Look at social media ... 28% of social media demand is caused by email marketing.  In other words, if you take email marketing away, your brand advocates don't have as much to yap about, and consequently, they spend a little bit less.  In this example, email marketing causes social media demand to happen, so you want to know that, don't you?

Email has minimal impact, when it comes to tablets/mobile.

We learned that email marketing isn't as effective as we believe it to be, when measured via traditional channels.  That being said, it's still highly profitable.

Tomorrow, we look at each channel, decomposing the role of each channel with/without advertising.

Attribution Week: Catalog Profitability

One secret to marketing success is knowing how customers behave if you stop marketing to them.


When you stop marketing to customers, you are left with brand recognition and merchandising love.


By the way, both of those things are really, really important.


So we execute a four-panel mail/holdout test.  One set of customers is treated "as usual" ... that's the catalogs+email column.  One set of customers is not mailed email campaigns ... for three months.


Yes, three months.


One set of customers is not mailed any catalogs ... for three months.

Yes, three months.


One set of customers is not mailed anything ... no catalogs, no email campaigns ... for three months.


Yes, three months.


You learn an awful lot by doing this.


We know that catalog marketing, when measured via the traditional mail/telephone channels, doesn't look very profitable.

Demand	$4.00
Net Sales	$3.40
Gross Margin	$1.87
Less Book Cost	$2.25
Less Pick/Pack/Ship	$0.34
Variable Profit	($0.72)

This is the outcome, over the course of a quarter ... it looks like we are losing $0.72 per customer ... and if you have 200,000 customers, well, that's more than $140,000 per quarter, more than $560,000 per year.


Well, that's doesn't work.  So we created the "matchback".  Any online order that happened within "x" days of sending a catalog is credited back to the catalog.  This "changes everything", as they like to say on Twitter!  Look at the table at the start of this post.  If we sum up all demand across all channels, we get $19.70, not a paltry $4.00.  Now look at the profit and loss statement.

Demand	$19.70
Net Sales	$16.75
Gross Margin	$9.21
Less Book Cost	$2.25
Less Pick/Pack/Ship	$1.67
Variable Profit	$5.29

Wow, catalog marketing works!  And that's where we quit.  This all happened back in 2000/2001/2002, our industry realize that we were "multi-channel" and we just stopped in time. We didn't innovate.  We just assumed that all of these online orders were caused by the catalog.


Not true.


Look at the results across catalog mail and catalog holdout groups.

Well, the test tells us something different.


Look at the incremental lift ... the difference between mailed results and holdout results.  We only generated $7.74 of incremental demand per customer.


In other words, if we don't mail any catalogs, the customer continues to generate 57.9% of the demand the customer was going to generate anyway.


I know, you don't want to believe this, you want to believe that customers crave catalogs.  Well, these days, customers are influenced by customers --- demand will still happen, regardless.


What does the profit and loss statement look like?

Demand	$7.74
Net Sales	$6.58
Gross Margin	$3.62
Less Book Cost	$2.25
Less Pick/Pack/Ship	$0.66
Variable Profit	$0.71

Ok, there's nothing wrong with this.  But the outcome is not what matchbacks suggest.  Instead, the results are 58% of what matchbacks suggest.


In this case, your job is to discount your matchback results, multiplying whatever your vendor tells you by 0.58.  Then, you run your profit and loss statement on what remains.


Later this week, we'll dig into the specific channel-based results, showing what happens to other channels when you stop mailing catalogs.

Dear Catalog CEOs: Attribution Week

Dear Catalog CEOs:


This week, we're going to tackle a concept known to the catalog industry as the "matchback", or to the rest of the world as "attribution".


As you already know, our industry incorrectly associates many online orders with the mailing of a catalog.  Based on mail/holdout tests, we know that some customers will continue to spend money, half or more of their dollars, when no catalogs are mailed.


This is not something that the vendor community wants for you to know.  The entire vendor ecosystem benefits when we over-attribute online orders to catalogs.

This week, we'll look at examples of mail/holdout tests.  We'll more accurately understand how catalog marketing impacts a business.  You'll be much more profitable if you adhere to the knowledge learned in the series of posts that will be published this week.


Good luck!

Crutchfield Attribution: Re-Visited

Last week's "Crutchfield: An Attribution Nightmare" was one of the three most popular posts of 2011, and for good reason.  The text of the post outlines a shopping experience where attribution becomes challenging, even overwhelming.

I received a lot of feedback, via comments, via email messages, and via Twitter.  You don't get to view a full picture of reader feedback.  I do.  So it's time to share some of the feedback, and my reaction, don't you think?

Question:  Kevin, you clearly doubted your purchase, or you would not have waited a month to place your order.  Nobody waits a month to buy an item without advertising playing a major role in the purchase.

Kevin:  Oh boy.  In the article, and multiple times in the comments, I noted that the only reason the purchase was delayed was to make sure that delivery would coincide with a window in which my builder could install the television.  It's that simple.  If you choose to not believe me, well, then that's your choice.  I'm being 100% honest.

Question:  Kevin, marketers who think customers will find you without advertising will lose market share.  Focus on last click attribution or fractional attribution, ok?

Kevin:  Oh boy.  Have you ever conducted a test?  Hold out catalogs for three months, and hold out e-mail campaigns for three months.  Do sales decrease by 99%?  No.  Anybody who has analyzed a mail/holdout test knows that customers will find you without advertising --- and the stronger your brand is, the more likely the customer is to find you without advertising.  We're attribution professionals, and yet, we fail to execute Attribution 101 --- simple mail/holdout tests.  If your attribution professional fails to recommend mail/holdout tests as the cornerstone of a proper attribution algorithm, go find yourself a new attribution professional.

Question:  Kevin, marketing clearly played a role here, you said so yourself when you decided to search for a lower price.

Kevin:  I purposely placed this comment in here, because quite honestly, this is the only place where one can logically suggest that price comparisons played a role.  If I had found the item for $350 less with comparable shipping, sure, I may have made a different decision.  Why almost nobody talked about this is beyond me.  This is the essential piece of the attribution solution!  I literally told you that nothing influenced my decision, I literally told you that I did execute a search, and yet, the feedback focused on how to parse shopping cart abandonment emails into the attribution routine, or how to parse the gift with purchase into the attribution routine.

Question:  Kevin, if activities correlate with an outcome, it's probably a good idea to participate in the activity that causes a correlation, right?  So if a shopping cart abandonment email is highly correlated with a purchase, we should execute shopping cart abandonment emails, right?

Kevin:  Let's view this from a different angle.  Let's say that you send two email campaigns per week.  By doing this, all purchases are correlated with the activity of sending email campaigns to a customer.  Therefore, all purchases should be allocated to email marketing campaigns, right?  You'd never agree to this if you were a search marketer, a television advertiser, a catalog marketer, or if you had retail stores and you saw that email campaigns took credit for all store purchases.

Question:  How did you know that Crutchfield offered free shipping?

Kevin:  I've been a customer since the 1990s, and I recall purchasing speakers in 2008.  So, obviously, there is some brand loyalty, prior e-commerce activity, and decades-old catalog activity that fueled my knowledge of the brand, coupled with email campaigns over the past three years.  That being said, you don't allocate today's sale to email campaigns mailed three years ago, do you?

I'm not saying I have any answers, other than to execute some mail/holdout tests, and to not discount the impact of search in the purchase process.  I am saying, however, that we, as attribution experts, have clear biases that result in us giving disproportionate credit to the marketing channels that we align with.  And that's the important thing to take away here ... execute some tests, minimize biases, and always remember that a significant portion of purchases (in spite of what some say) happen because of brand loyalty, and not because of marketing activities.

Crutchfiled: An Attribution Nightmare

Attribution experts express confidence in their ability to parse an order across the activities that caused the order.

Oh boy.

Let me explain an attribution nightmare, courtesy of Crutchfield.

I was about to purchase a new television, a 46 inch model.  I decided to purchase the television from Crutchfield, because I knew I would get free shipping.

I made this decision on or around May 1.

In the first week of May, I received a small page count catalog from Crutchfield.  I looked through the catalog.  My television was not, to my knowledge, featured in the catalog.

In the second week of May, I researched the television I wanted to purchase.  I used Google.  I visited Amazon.  I visited J&R Music World.  I visited Best Buy.  I found the television I wanted to purchase.  I compared prices across brands.  I then visited Crutchfield.  The television I wanted was available.  I placed the item in my shopping cart.

In the third week of May, Crutchfield sent me an e-mail marketing message ... and surprise surprise, the television I wanted was featured in the e-mail marketing message.  Who knew?!  It's almost like Crutchfield scanned my cart and then advertised the item to me.

Late in the third week of May, Crutchfield lowered the price of the television.  Good for me!

Early in the fourth week of May, while reviewing my shopping cart, Crutchfield made me an offer.  They offered to give me a soundbar/subwoofer system, valued at over $300, for free, if I purchased my television.  Wow.  I added the soundbar/subwoofer to my shopping cart.

Midway through the fourth week of May, Crutchfield sent me an abandoned cart e-mail message, reminding me of the items I wanted to purchase.

In the first week of June, I added a mounting bracket to my cart, and I added a separate item not related to the television purchase to my cart.  I placed my order, saving about $50 on the television, obtaining a soundbar/subwoofer valued at over $300 for free, and getting free shipping as per standard marketing practices at Crutchfield.

Ok, attribution experts.  Tell me what caused this purchase?

• What impact did the mailed catalog have, given I already decided to buy the television prior to receiving the catalog, but given that the catalog was mailed a few weeks prior to buying the television (hint --- this is the crux of my stance on the "organic percentage").

• What impact did the first e-mail marketing message have, given that it featured the television I already placed in my shopping cart?

• What impact did the price change on the television have on my purchase?

• What impact did the offer of a free soundbar/subwoofer have on my purchase?

• What impact did the abandoned cart e-mail message have on my purchase?

• What impact did free shipping as a standard offer have on my purchase?

• What impact did Google have on my research?

• What impact did the fact that I already decided to buy this item before the marketing began have on my purchase?

• How would you, the marketing/attribution expert, parse this information without knowledge of the fact that I made up my mind on the purchase before any of the marketing activities were initiated?

Discuss.  Remember, I decided to buy a television from Crutchfield before any of the marketing activities mentioned here were initiated.

E-Mail Marketing, Search, Matchback, Attribution

One of the mysteries of marketing in 2009 is the concept of attribution, a process where we matchback orders derived in one micro-channel to the advertising micro-channel that drove the order.

For whatever reason, the e-mail blogosphere and vendor community fails to capitalize on this opportunity.

My Mutichannel Forensics projects repeatedly indicate that e-mail marketing and search marketing play a unique micro-channel role. E-Mail marketing is a "love" channel, if you will. The 10% to 50% of your twelve-month buyer file that subscribes to e-mail marketing "loves you" more than the average customer. These customers have better "RFM" characteristics, not because of e-mail marketing necessarily, but because the customer is a good customer who wants to learn more.

And then we have search, which works in the opposite direction. The customer who "loves you" doesn't implicitly trust you. As a result, she wants to make sure that she's getting the best deal possible, the best combination of merchandise and value.

When you have customers who want to see your e-mail campaigns and then want to use search, you have a classic micro-channel combination that must be tabulated in your database, and analyzed going forward.

At minimum, we need to run matchback algorithms for e-mail marketing. Catalogers have been running matchbacks for the past fifteen years, taking credit for orders that were not necessarily driven by catalogs. E-Mail marketers, however, have been exceptionally slow to embrace attribution and matchback programs. I don't understand why.

It's a fairly simple process. Say you deliver an e-mail marketing campaign on a Tuesday. Take all customers who ordered on Tuesday, Wednesday, Thursday, and Friday, and match them back to your e-mail campaign. And by the way, make sure you have a holdout group, a group who did not receive the e-mail campaign, and do the same process --- subtracting the difference between mailed and holdout group for true incremental value.

Now, any orders that are generated by search marketing are matched back and attributed to the e-mail marketing campaign. And here's where we need to make an adjustment ... we need to make a guess at all of the unconverted searches that were caused by e-mail marketing, and allocate the cost of those unconverted searches back to the e-mail marketing campaign. If the typical search conversion rate is, say, 3%, you have to multiply converted searches by 33, and then multiply that total by the cost-per-click, in order to get at the right advertising cost.

Two things usually happen, two things that are highly relevant to e-mail marketers.

1. E-Mail marketing causes search activity, and that search activity results in orders that are normally credited to search and should be credited to e-mail. This can result in e-mail marketing being more productive that usually measured to be.
2. E-Mail marketing causes the "search audience" to do a bunch of unproductive searches. As a result, the "search segment" is actually unprofitable --- causing the e-mail marketer to withhold e-mail marketing campaigns to customers who search all of the time.

The latter point is worth noting ... the e-mail marketer should be creating segments in the database of customers who utilize search on a frequent basis ... electing to develop a different contact strategy for the "E-Mail / Search" micro-channel combination.

Mega-Metrics: Attribution Combinations

Oh those online marketers ... looking to attribute orders to the right marketing channel.

Allocation / Attribution is a largely futile effort. We simply cannot get in the mind of the customer, can we?

We can test. Testing works really well in the direct marketing world, but testing for attribution purposes works less well in the online marketing world.

So here's what you can try ... I've had success with this.

• Create what I would call "combinations". When a customer visits your site on March 3 via paid search, and purchases on March 5 via an e-mail campaign, you create a "combination" that captures the first touch and the last touch. In this case, the combination is "PAID SEARCH / E-MAIL".

• If the customer only had one visit (say from an affiliate marketer), then the combination is "AFFILIATE ONLY".

• The pundits will jump in and suggest that we're missing everything that happens in-between first touch and last touch. Let them jump in and argue.
  

• Pay CLOSE ATTENTION to what happens on the next purchase. What are the combinations that are most likely to happen NEXT? Does "PAID SEARCH / E-MAIL" lead to "E-MAIL / PAID SEARCH"? Or does it lead to "PAID SEARCH / E-MAIL"? Or does it lead to "ORGANIC ONLY"? The mega-metric to calculate is the percentage of folks who migrate to different combinations in the future.
  

We care about what happens next, because what happens next tends to dictate what the primary driver is. Over time, you'll learn that certain advertising channels tend to "drive" orders, while other advertising channels tend to "complement" orders.

Once you identify the channels that drive orders, you begin to understand how you might attribute orders to ad channels better --- and you'll make better investment decisions.

The combinations help you see how customers migrate over time ... helping you figure out "what to do next" as opposed to helping you figure out "what just happened". Ultimately, you'll plug the most popular combinations into a Multichannel Forensics simulation, and you'll see what is about to happen to the future trajectory of your business.