Monthly Archives: February 2008

Marketing is the New Finance

That’s OK, you can spit your tongue out now.

Not my words, but those of Hal Varian, Google’s chief economist.  I think what he is saying is Marketing is about to undergo a performance (ROMI) revolution.  Here’s the context for the quote, from the WSJ:

“I think marketing is the new finance.  In the 1960s and 1970s [we] got interesting data, and a lot of analytic fire power focused on that data; Bob Merton and Fischer Black, the whole team of people that developed modern finance.  So we saw huge gains in understanding performance in the finance industry.  I think marketing is in the same place: now we’re getting a lot of really good data, we have tools, we have methods, we have smart people working on it.  So my view is the quants are going to move from Wall Street to Madison Avenue.”

Varian recently took questions over on the Freakonomics blog that are an interesting read.  On analytics, check this out:

“If you are looking for a career where your services will be in high demand, you should find something where you provide a scarce, complementary service to something that is getting ubiquitous and cheap.  So what’s getting ubiquitous and cheap?  Data.  And what is complementary to data?  Analysis.”

Amen, Brother

Check out the rest of the questions and Varian’s answers on the Freakonomics blog.

 

Share:  twittergoogle_plusredditlinkedintumblrmail


Follow:  twitterlinkedinrss

Natural Born Clickers

Did you catch the research piece published a couple weeks ago about the behavior / demos of folks that click on display advertising?  You can read the whole thing here, but for the sake of getting to the point:

“6% of the online population account for 50% of all display ad clicks”

“heavy clickers skew towards Internet users between the ages of 25-44 and households with an income under $40,000”

“more likely to visit auctions, gambling, and career services sites”

The incredible conclusion the purveyors of this Research for Press Release come to is this: you should still buy display advertising, just don’t measure the value of it using clicks, because “the branding effect of the ads is what’s really important.”

Wow.  So, the first thing I want to know is this: where were these ads running?  Because where they run has everything to do with the demos; the sites create the demos, not the ads.  Did they run on NetZero, for example?  Any other free to almost free access systems?  Because if the audience won’t pay for the service, you get the audience quality you deserve.

Another thing that would be really interesting to know is the demos of folks who click on PPC ads.  Want to bet more upscale, higher income, older?  Again, it’s not “clicking” that is the issue here, it’s how the ads are delivered; the network creates the audience.

If the above data is without network bias, my guess is this: the 6% of people that generate 50% of display ad clicks are people that are bored because they have no jobs.  And those people click on ads because they have nothing better to do.  They’re “surfing” the way the web worked pre- decent search engines, when you found out about other sites by clicking on ads.  An ad meant there was a site somewhere other than where you were!

Now, most folks that have any kind of intent at all use a search engine.

That brings us here: if the majority of clicks on display ads are generated by people who are jobless and bored, then is there any real Branding effect worth talking about from display ads?

In other words, what if the entire audience for display ads is younger and low income, just because of the stupid, untargeted nature of the format / distribution.  Some of this audience clicks a lot, because they’re bored and have nothing else to do.

The desirable audiences are off using a search engine, because they have purpose and intent.

So if you run a gambling or a job site, it makes a lot of sense to do “Branding” with display ads.  Likewise, debt-related.  LowerMyBills.com comes to mind, the banners with those idiotic dancing women to get your attention.  Here kid, click this.

But if you sell Expensive Vodka?  Luxury Cars?  High End Toys?  Electronics?  Do the targets for those products even see banner ads unless the site is super-targeted

What if display ads are only seen / paid attention to by “aimless surfers with no purpose”.  Gee, that’s an audience I really want to create a Brand impression for – regardless if they click the damn thing.  In other words, what if the Natural Born Clicker (NBC) demo is not just of a segment, but of display as a whole?

Otherwise, why didn’t they give the demos of the non-NBC clicks or exposures?  Instead, they compare to “total U.S. online population”.  That’s a Research for Press Release slight-of-hand trick.  What we need to know are the demos of the “light clickers” and the demos of the “exposed” – people who saw / remember display ads.  Not the entire online population.

Unless, of course – gasp – these demos are the same as the NBC folks.

Most people are too busy to pay any attention to irrelevant display ads in the first place.  They use search and completely ignore the side bars where display ads run on a site. 

Active, lean forward, get it done folks focus on the content.

Kinda makes you wonder…

Your thoughts?

 

Share:  twittergoogle_plusredditlinkedintumblrmail


Follow:  twitterlinkedinrss

Hitting the Wall

Looks like Red Envelope is Hitting the Wall.  What they need now is a Relationship Marketing Strategy based on understanding dis-engagement.  It’s the only way to dig out of this. 

They need higher Marketing Productivity.

Check out this statement:

“We hoped that our renewed creative statement would drive an up-tick in performance in fall and holiday, but unfortunately this was not the case” – CEO John Pound

Got some old schoolers there.  That’s a “pray and spray”, offline Customer Marketing approach, not an interactive customer strategy.  They have to understand why the behavior is different

You know, the difference between Brand and Branding.

They have the data.  But they didn’t do the analysis, because the effects of dis-enagagement were masked by new customer acquisition.  They thought everything was just fine and changing creative could “give them an uptick”.

Not.  I’m just guessing, but it seems to me this business model is a case where the Catalog – as they are implementing it – is only driving incremental business in very small segments, and is a loss overall.

In other words, a prime candidate for a Controlled Test to resolve this matter of catalog incrementality quickly by avoiding the “matchback” problem so common when one takes an old school approach to an “always on” interactive medium.

I’m not saying the catalog is a bad idea.  I’m saying it matters more than most folks understand who you send it to and when.

So, have you ever shopped with Red Envelope?  How was the experience?  Did you purchase again from them?  What’s your Net Promoter Score with respect to them? (giggle)

More details on what Hitting the Wall is all about are here.

Share:  twittergoogle_plusredditlinkedintumblrmail


Follow:  twitterlinkedinrss

MultiChannel / Omnichannel Mayhem – CAO Solution

Well, the deluge of comments on the Managing MultiChannel Mayhem post was underwhelming for such an important idea, and I bet I know why.

You thought the system of checks and balances on communicating properly with customers was overly complex and unwieldy.  You could not imagine a self-correcting marketplace of permission within your company that would boost Marketing Productivity and the bottom line.

You’re right, it probably won’t happen at the company you are working at now, even though it’s the right way to go about it.  So let’s look at an alternative that might achieve the same result, shall we?

Even when this kind of marketplace system is set up, there are a couple of problems worth mentioning.

The first is a “black market” in permission to avoid the scrutiny of ROMI.  Channels / divisions can “cheat” in a number of different ways – using backdoor customer contact trading agreements, withholding / hiding customers, keeping and re-using copies of lists, etc.

It takes a lot of enforcement and a lot of effort, unless the company was born into this marketplace mode of thinking, and has a centralized customer database.  Then people tend to play fair, which happens at a lot of pure DM companies.  It’s just the mindset; Database Marketing people play it by the numbers – win, lose or draw.  Nobody fears analytics; failure is a learning experience to them, they say “Bring it On!”

The other cheating problem you run into is data torture by the analysts in each channel or division.  In a marketplace model, everybody wants to win, so there is incentive to bend the rules at every turn to produce stats that push your channel or division ahead of the others.  Modify the definition of a customer slightly, for example.  Stretch or compress time periods.  Intentionally use data that has not been cleansed.  Take advantage of slippage in the data update process.  All of that good stuff.

So what’s the answer, how do we solve the MultiChannel Mayhem problem?  How do we correctly implement the Interactive Customer Marketing triad of Relationship MarketingMeasuring dis-Engagement, and Control Groups, but avoid the cheating problems that go on with the Marketplace approach?

Here’s the answer: Every MultiChannel Company involved with at least one Interactive (always on) medium needs a Chief Analytics Officer (CAO), independent and reporting directly to the CEO, with all the senior analysts in the company reporting directly to this CAO.

Really.

One of the things you find out when trying to apply the permission marketplace concept is you need lots of standards and a way to enforce them.  So you have to build this CAO capability / authority anyway, and rulesets to manage the whole thing.  Then everybody still tries to cheat the marketplace but with a CAO sometimes they get caught – usually only after the deed was done.  That’s Suboptimal.

So, to address that, the next phase of the marketplace fix consists of dragging all the analysts that matter out of all the silos, and having them report directly to the CAO, creating a giant Research & Analysis group (R & A).  Then you easily have standards and controls for everything across all the marketing silos.  Standard definition of a customer.  Standard definitions of customer dis-Engagement and defection.  Standard definition of ROMI.  Standard data cleansing and list pull procedures.  Standard data update sequences.  Standard implementation of Control Groups.  All of it.

And then, magically, the importance of the complex marketplace idea actually goes away.  Because with an independent CAO and R & A group reporting directly to the CEO, you get all the benefits of the marketplace without the marketplace itself.

The CAO only has one directive from the CEO – maximize profit.  Allow the strong Customer Marketing ideas to blossom and the weak Customer Marketing ideas to fail.  It’s the same result as the Permission Marketplace, only with a lot less complexity.

Here are some direct benefits of this approach:

1.  If you have at least one always on, interactive channel, you cannot measure the success of marketing without using Control Groups.  Let’s say you want to implement them.  How are you going to keep other channels / divisions from stomping on your controls and destroying your tests if anybody can access the database at any time and pull contact information for Customer Marketing Programs?

It’s impossible.  You need centralized control of the database and one entity pulling the contact data so that tests are flagged, controls are flagged, and as a result extracts for contact are pure and free from the outside influence of any other direct-to-customer marketing efforts.

2.  There is something quite magical that happens when you put many analysts, each specializing in a different piece of the business, together in the same space reporting to the same person.  They talk to each other.  They go to lunch together.  They discover things the business would never know about otherwise.  Linkages.  Correlations.  Root Causes.  Because much of the analytical gold is buried not in a silo, but across and through a silo into others.

There’s really no way to put a value on the kind of exploration and discovery that can occur when you put many smart people, each concentrating on a different piece of the business, in a group reporting to one person, who in turn reports directly to the CEO.  Standards are clear and alliances are to the customer truth, not to the marketing silo.

I can hear the wailing already on the centralization of analysts.

“Jim, analysts are always more productive and have a better understanding of the channel if they report to and live with the other channel people”.  Personally, I don’t think that’s true; I know of very successful centralized groups that have great knowledge of the units they cover.  And from a management perspective, the reality of analytical oversight is you don’t want the analysts to become best friends with the folks whose work they are analyzing.

Without the analysts physically near each other, I think you miss out on much of the analytical synergy.  With the communication tools we have today, it’s just not clear to me why a great analyst cannot have intimate knowledge of a channel just because they report elsewhere or even live elsewhere.

As I said, the highest ROI analytical gold is often found understanding what happens between and across silos – since when the analysts are each in a silo, this cross-border work is rarely pursued and tends to be undiscovered.  You want to encourage a “world view” of the company at this senior analyst level, not a silo-specific view.

I mean MultiChannel, by definition, means across channels, right?  Who is responsible for that analysis?

And I’m not just talking about “customer” analysts.  If I had my way, I’d have sales, customer, merchandise, supply chain, service, fulfillment – all of the senior analysts – in the same tank to encourage this world view.  You can always have operational analysts in the silo itself – your future senior analysts.

Think about it.  Who is doing cross-silo analysis at your company?  An analyst in a silo?  Right.  Who is going to figure out a check-buyer Marketing Program is not working because they are shredding checks over in Customer Service?

But if physical location is truly important, fine, the reporting relationship – the loyalty / trust chain to the CEO – is the most important part of this idea, not the physical location of the analyst.  The reason this is so important is these folks need to be protected from the inevitable pressure and flack that will be coming at them from silo heads, if the CEO is really looking for the truth.

3.  With the centralized analytical setup, rather than the trading complexities and vagaries of the Permission Marketplace, you simply have smart people who play by the numbers controlling Marketing access to the customer.  So you’re the Catalog Division and you want to mail to new customers of the TV Division?  Fine.  We’ll let you try that a few times, and if it looks like an incredibly bad idea financially, then we simply won’t let you do it any more.

In other words, the message from the CEO to the CAO on down through the R & A Group is this: Folks can test anything they want.  But we will not allow channels / divisions / Marketing silos to interfere with anybody else’s test programs, and we will not allow channels / divisions / Marketing silos to intentionally and repeatedly destroy customer value.  That goes for every business unit and Marketing silo, regardless of who they are or where they are from.

When you get to this point, you know Customer Marketing is actually being run by the numbers.

What does this analytical culture look like in practice?

You can still have your own analysts in your silo, that’s fine.  Each silo with analytical folks gets a standardized, clean data feed to work from – minus contact information if that is what is required to stop cheating – from R & A.  Same data set for everyone.  Nobody from outside of this R & A group pulls from the customer database.

And as far as any contact data to be used for Marketing, programs that will actually touch the customer, you get your lists from the R & A group, and you get your customer marketing campaign results from the R & A group.  You can track things on your own, sure.  But the final bottom line analysis on customer impact / ROMI comes from R & A.

Any kind of Marketing Success analysis from outside R & A is not used for significant decisions.  If you are trying to argue budget, priorities, people, program success, Strategy, whatever at the senior level, analysis from outside the R & A group doesn’t even make it through the door.  And this has a really interesting effect on senior people.

They actually start making very large and significant decisions based on the data.  Why?  Because they trust the results.

Perhaps for the first time in the history of the company.

Think about it.  If every silo creates their own results, then why would a senior exec really trust them?  These senior folks are not stupid.  So as long as the decisions being made are not that important, as long as the company doesn’t really “live and die by the numbers”, it’s OK to have self-produced, un-audited, un-controlled, “who cares if this is reality” numbers.

Because the numbers don’t really matter.

If you are serious about running a company by the numbers, and that company has any interactive, always on components, you need a R & A Group that is completely independent of any channel or Marketing silo.

Otherwise, you can’t trust any of the Customer Marketing analysis.

Interactivity is, indeed, different.

Share:  twittergoogle_plusredditlinkedintumblrmail


Follow:  twitterlinkedinrss

MultiChannel / Omnichannel Mayhem – Example

It’s been over 10 years since I left Home Shopping Network, so I’m pretty sure I can tell this Relationship Marketplace story without enraging the Gods.

The mail order channel had a jewelry-only catalog that did quite well, in terms of mail order ROMI.  This book generated the majority of the profits in the mail order division.  What we wanted to know was this: if you looked at the book from the company perspective – across all channels at the customer level – what was the contribution to profits this book made?

When we started using Control Groups on the book and looking more deeply into the Relationship Marketplace, here is what we found:

1.  The book actually lost money versus control.  It virtually had zero impact on sales at the customer level, meaning it was almost completely cannibalistic to TV.  That means the entire cost of the book, plus the cost of discounts, was a net loss on every book mailed.  The more they mailed, the worse the losses were.

I remind you this book generated the majority of the profits at the mail order divisional level.  This revelation was not pretty.

2.  Like the good Marketers they were, the mail order folks had tested all kind of segments across the entire customer database, and found brand-spanking new TV jewelry buyers to be the most productive target.  In other words, they were mailing to the folks with the highest level of Engagement.  This is why the book was very productive at the divisional but not at the company level.  The “always on” nature of the TV channel demand was pulling in buyers all by itself, and the book was essentially just getting these sales that would have happened anyway to switch channels.

3.  Further, and perhaps even scarier, the mail order division was involved with a list exchange that had a “hotline” (new buyer) component.  That meant as soon as TV acquired these highly Engaged jewelry buyers, their contact info was rented by the mail order division to competing jewelry catalogs if they became mail order buyers.  That’s OK for the mail order division, which gets to keep the list rental income.

I should also say list exchanges of this type are standard practice in the catalog business.  So the catalog folks were not doing anything “wrong”, from the catalog division perspective.

But I think we could all agree that situation sucks for the TV channel, which deployed assets to acquire the customer in the first place.  Not only is TV getting hosed on the (company perspective) non-incremental sales stolen by the catalog division through this book, TV is not getting a piece of the list rental income from selling the contact info of their best customers to the competition!

Further, as one might guess would happen with highly Engaged customers, when we held back a control group from the list rental process, those folks who were not rented out delivered higher profits to the company overall than those rented out.

Just makes sense, right?  Relationship Marketplace, indeed.

What to do?  We simply re-configured the targeting of this catalog to dis-Engaging TV jewelry buyers.  As they stopped buying from TV, this very well-executed book was able to extract another purchase or two from the customer and actually maintain some of them longer term.  Sales for the book dropped dramatically, but these sales were truly incremental versus control, and the book was still profitable.  This also affected list rental, but since the buyers had already dis-Engaged from TV, this issue was not nearly as critical.

Got any multichannel misallocation stories of your own you can share?

Background on this post here; solution I’ve seen succeed, here.

 

Share:  twittergoogle_plusredditlinkedintumblrmail


Follow:  twitterlinkedinrss

Managing MultiChannel / Omnichannel Mayhem

So, you’re saying to yourself, all this talk about Control Groups and Measuring dis-Engagement and Relationship Marketing, is there any other reason to care about this stuff besides Campaigns?

Yes – addressing MultiChannel Mayhem – the fact that few companies have a solid plan for really optimizing the multichannel system.

Kevin shows us why MultiChannel Mayhem matters here.  Turns out multichannel customers are not always the best customers.  Of course, you knew that, right?  I mean, you have relationships with multichannel vendors you conduct through a single channel, right?  And you would consider yourself a pretty darn good customer, right?

The reason Control Groups, Measuring dis-Engagement, and Relationship Marketing are so important is they comprise an entire MultiChannel Interactive Marketing System.  It is pre-built and ready to go, customized to take advantage of and optimize Interactive Customer Relationships, where sometimes the most profitable Marketing is to do none at all.

If you are an analyst / technical background, you would probably be more comfortable studying the system from the bottom up, from the very measurement-oriented idea of Control Groups up through Measuring dis-Engagement to the Marketing Strategy idea of Relationship Marketing – what, how, then why.

If you are a Marketer, you would probably be more comfortable studying the system from the top down, from the Marketing Strategy idea of Relationship Marketing down through Measuring dis-Engagement to the measurement-oriented idea of Control groups – why, how, then what.

Why should you care about this system?  Because addressing the problem of MultiChannel Mayhem is built right into the model, along with all the cultural and organizational tools you need to implement.

Let’s go though a common challenge with Multichannel Marketing and see how the system solves it.  First, we need some assumptions:

1.  The goal is to optimize the MultiChannel system for Return on Marketing Investment – not sales, not share, etc.
2.  Given proper measurements, people will make rational decisions.
3.  The general concept of long-term customer value is important to the company and people act accordingly

Now, let’s address a common problem – no controls on who uses the customer database or what it is used for.  Any division or channel can market to any customer at will, doing whatever it is they want to do, regardless of previous messages.

This is a mess, and you know it.  Customers are being bombarded with conflicting brand and offer messages.

Here’s how the above system solves that problem:

1.  The division or channel that acquires the customer owns the customer exclusively for some period of time.  After all, they started the Relationship, they invested in acquiring the customer, they should get the rewards of the Potential Value they have created.  No other divisions are allowed to communicate with the customer without the permission of the acquiring entity.

2.  The acquiring entity retains this exclusivity to the customer only as long as they can maintain an active Relationship with the customer.  Once dis-Engagement has grown significantly, the customer is open season for any division or channel.  This is why it is so incredibly important to define and measure the dis-Engagement process.  The customer is evolving, and this rule ensures the company is flexibly responding and evolving with the customer, following the Relationship, and maximizing value.

3.  After the initial acquisition and subsequent dis-Engagement, whichever channel or division has the most current relationship owns the customer.  If a customer defects from one channel to another, the “ownership” passes as well.

4.  The channel or division with the most current relationship can at any time sell communication permission (basically, rent the list of customers Engaged with it) to other channels or divisions if it wants to.  This is not only a nice way to help offset acquisition costs, it also creates an internal marketplace of relationship permission, where only the communications which generate incremental customer value survive the Controlled Testing process.

If you have to rent a list of new customers from the acquiring entity and pay for that list, you are going to be more thoughtful and careful regarding what kind of communications you send.  Realizing that prior communications are going to affect the success of your communications to this rented list, you are going to tend to stay on brand message and within established communications guidelines.

In other words, this construct enforces financially responsible testing.

The rental fees are negotiated between buyer and seller.  The seller can charge different fees for access to different segments, for example, known best buyers cost more to rent than predicted dis-Engagers.  Any segmentation you want.  This is a CPM-based rental fee arrangement to keep list access prices fair across channels.  I suppose you could go to CPA, but that’s probably going a bit too far, given the lack of ability for some channels to track such things.

5.  If the customer is a true multi-channel customer – active in multiple channels or divisions simultaneously – then all channels the customer is active in have permission to communicate with the customer.  Any activity with any division or channel – including service requests – indicates Engagement is current and resets the dis-Engagement clock to “Currently Engaged”.  This addresses the above scenario Kevin has painted, which is more like the reality of multichannel relationships – they ebb and flow between channels over time.

From the company point of view, this system is a self-regulated, self-optimizing customer value generating machine.  It ensures that from whatever channel or division perspective you view the system, the strong survive and the weak fail, but the Potential Value of the customer is always maximized.

6.  If, from a Strategic perspective, the Company wants to intervene in this system, that is permitted, as long as there is compensation to the Relationship-generating asset gatherers.

For example, when launching a new channel or product line with no customers to start with and no cash flow.  The Company, on behalf of the new channel or line, pays rent to all the other channels for the permission assets they have generated.  The channels still get a return on their assets, and for the Company, it’s simply accounting for a real cost – the cost of cannibalizing existing sales in another channel / line or acquiring the customer.

From a practical perspective, whether money actually changes hands or not in the permission marketplace is a company culture thing.  The more sophisticated database marketing companies will often force real monetary exchange because they understand customer value so well.

Other companies might simply run debits and credits in a spreadsheet, creating more of a “list exchange” type of scenario but one that still provides controls on (for example) inundating best customers with communications and ignoring defecting customers.  These credits can be used to pay for other cross-channel marketing activities as well, for example, the mail order division wants to put catalogs in the web division’s outbound shipments and pays with credits received from the web division’s use of catalog e-mail addresses.

A self-regulating, self-optimizing system for Managing MultiChannel Mayhem, built on Relationship Marketing, Measuring dis-Engagement, and Control Groups to keep everybody in the system honest.

It’s really no different than, say, optimizing a PPC campaign – except it’s only about customer communication, and instead of optimizing across keyword sets and landing pages, you are optimizing across the entire company, every campaign in every channel against each other.

So, do you have Multichannel Mayhem at your company?

If so, does it matter to you or not?  Matter to anybody?

If it does matter to you, do you think the system above would address the problems you have?

 

Share:  twittergoogle_plusredditlinkedintumblrmail


Follow:  twitterlinkedinrss

*** A Paradigm Shift

Cisco’s passage to the new world of customer-driven marketing

I tried to explain the importance of measuring dis-Engagement from the Analytical side and from the Marketing side.  Didn’t seem to get much traction with all of you.

So now I think I will let somebody else explain it – Cisco.

From Target Marketing Magazine, we have this article about how Cisco is measuring and using Engagement to drive higher close rates on leads.  There are some really interesting phrases spoken by the marketing folks in this article you should pay attention to:

In the old world, a campaign was temporal in nature.  In the new metaphor, the campaign is always on 

need to shift from disruptive model to an engagement model

moving from a monologue, where we just send e-mails to people and they respond when they can, to a wave, where we are actually creating a dialogue with that customer

customers will self-propel through the buying cycle

self-serve themselves into the buying experience as they need to

be more responsive to customers when they’re leaning forward

listening to your customers is an ‘old world’ phrase, and it means something different in the ‘new world’

empowering a customer to move through the sales process on his own

new type of evangelist to do marketing this way versus doing marketing in the disruptive manner

All sounds pretty familiar, if you read my last post.  This is the way you optimize Interactivity, the Strategy of Relationship Marketing.  “Interactive” means continuous two-way exchange, right? 

Not outbound sledgehammer to the head every week?

It’s a story of using the LifeCycle, of measuring Engagement and dis-Engagement, in action.  When the customer is Engaged, you leave them alone.  When they start to dis-Engage, you open a dialogue .  It is really as simple as that.  It’s a dialogue based on behavior.

This marketing team took Cisco’s lead quality from 13% of leads expected to close to 75% expected to close.  “Expected” here means so darn likely these leads get added to the sales forecast.  That’s my kind of “expected” – 90% actually close.

Gotta be some smiling folks in that Sales department.  And those kinds of numbers are common with Relationship Marketing, in both B2B and B2C.  It’s just a better model.

Here’s another link to the article: A Paradigm Shift

So a question for you: Have we forgotten the essential premise of the web – Interactivity?  What’s the word “Interactive” mean to you

Share:  twittergoogle_plusredditlinkedintumblrmail


Follow:  twitterlinkedinrss

Can Marketers (and Analysts) Have it Both Ways?

Well, the Framework for Engagement did not generate much discussion.  Perhaps it was too technical, or not technical enough.  That duality is always a problem when you are trying to unravel an issue at the junction of Marketing and Technology; you have to speak down the middle and sometimes, you don’t “engage” either audience.

So let’s try attacking this issue in another way…from the “generally accepted Marketing speak” perspective. 

First, a couple of short definitions:

Brand – the actual personality of your company or products, the experience and feelings people have when they think  about and interact with your company or products.

Branding – the act of trying to communicate this brand or product personality to people though media.  As we know, this Branding effort can project an image that is different from the actual Brand.

Here is what is happening:

1.  Marketing communications used to be one-way, and the “media” was in control of these communications.  This was the era of Branding, where companies tried to establish a personality and set of benefits associated with products.  For the most part, the only challenge to this effort, unless it was a spectacular failure or blatantly false, was conversations over the backyard fence with neighbors.

2.  Marketing communications are now at least two-way, if not many-to-many.  This means in terms of the shear volume of communications, it’s more often Brand (the true personality or experience) than Branding being communicated, e.g. “the customer is in control of your Brand”.  While I don’t think this is true, I think it is possible to allow the customer to take control of your Brand, but you have to give up control for that to happen.

3.  With Brand more important than Branding, companies need to pay more attention to the actual product and service experiences they deliver.  This means product design and customer service / experience become much more important than they have been in the past, relative to Branding and media.

4.  Direct Marketers have always been hyper-sensitive to Brand – the products and experience – because that is all they have.  In a one-to-one Relationship with the customer not fueled by GRP’s, if they screw up the Brand part, they are toast.  They can’t afford to do business any other way but being deeply thoughtful about the end customer. 

This is why, for example, most catalogs and TV Shopping networks were very early winners in e-commerce and established a lot of the best practices, especially on the back end – they already got all this.  Same business, different medium.  Web analytics was happening at these companies way before it became accepted as mainstream.

This is what interactivity with the customer brings to the table.  It changes the customer Relationship.  Yes, the web is different, as is TV shopping.  It requires a lot more focus on actually delivering on Brand promises.  It requires more empathy with the customer, requires streamlined and flawless operations and execution.  Because now that every person can “be the media”, it is possible for a company to allow customers to take control of the Brand.  But they have to give this control up first, by not understanding the fundamental changes interactivity brings to the table through the shift from Branding to Brand as the centerpiece of Marketing execution.

Here is what needs to be done: 

1.  Unfortunately, most Marketing people have skills that are aligned around media and creative, not products and service.  This is unfortunate, and is leading to the Deconstruction of Marketing and rise of the Chief Customer Officer.  To correct this, Marketers need to expand their skill sets, and interact more deeply with Operations / Customer Service.

2.  For most Brands outside of packaged goods, Media dollars need to be shifted away from mass media and more into direct / database Marketing and yes, into Operations.  Personally, I have used Marketing budget to pay for customer service testing and upgrades, because I know how important they are when dealing with interactivity.  Apparently, I am not alone, as some of this is starting to happen already.  For example, $100 million out of TV into direct at Talbot’s.

3.  Marketers (and analysts, of course) need to change the way they look at customer measurement.  If the customer relationship is now different, it has to be measured differently, measured interactively.  I run media, I get response, that measurement is not accurate with existing customers.  I drop catalogs, I get response, that measurement is not accurate with existing customers. 

Because now you have Interactivity, and you have always-on scenarios that are self-perpetuating, people can visit or order from a web site any damn time they want – regardless of marketing stimulation.  The key variable to understand is not response, but what you get from the customer without doing any “outbound” marketing at all.  This is the true measure of your success with Interactivity – how “engaged” the customer is with your Brand.

If you are addressing Interactivity properly and doing a good job with the Brand – design, experience, service – consumers will want to interact and  / or purchase from you regardless of and sometimes even despite your marketing efforts

That is the way Interactivity works. 

In fact, it can be argued that the better you are at Interactivity and Brand, the less you will have to spend on Media and Marketing, because interactivity begs continuity – you don’t need to spend to drive response, because you have an ongoing relationship, the customer is drawn back to you time and time again through the power of your Brand.  That’s what you wanted, right?  A Relationship?

Engagement?

So I ask you, Marketers and Analysts.  Can you have it both ways?

You wanted Interactivity, now you have it.  Now that you have it, are you going to go about your business like nothing has changed?  Will you continue to measure your success in the many-to-many media world using the one-way media measurement model of “response”?  Or are you going to learn the ways of Interactivity, shed the old models, and begin to learn how Customer Interactivity works? 

After all, if all this Interactivity with customers is really new and very important, then shouldn’t “Strength of Relationship” be the most important KPI you could ever measure to determine success?

In other words, are you going to Measure Customer Engagement and dis-Engagement rather than Response as the true benchmark of your success in an Interactive Marketing world?

What are your comments on the above?  Does the scenario I painted on what is happening with Marketing / Brand / Branding make sense to you?  What about my suggestions on what Marketers need to do to address these changes?  What did I get right or wrong?  What is holding you back from making these changes, if anything?

Share:  twittergoogle_plusredditlinkedintumblrmail


Follow:  twitterlinkedinrss

A Framework for Engagement – e-Mail Example

Let’s take a specific example of what I was talking about in the last post in this series to show you what a Relationship Marketing Strategy looks like in action.  I have stated, for example, in this interview, that email is both taking credit for sales that would have happened anyway and generating more sales than you think it is. 

These statements are both accurate at the same time.  When you use Control Groups to measure the incremental behavior generated by e-mail campaigns, what you find is both effects occur at the same time but each one happens with a completely different behavioral segment – Engaged versus dis-Engaging.  Unfortunately, in most cases, the net effect is e-mail falsely takes credit for more sales than it doesn’t get credited with accurately, because the Potential Value (likelihood to buy) of the Engaged is far greater than that of the dis-Engaging.

Against a graph of Response by how long ago the last click in an e-mail occurred, these dual effects on the actual response rate of an e-mail drop versus control group look something like this (red line).

Click picture to enlarge:

On the left we have the highly Engaged, and moving to the right, the pattern of dis-Engagement.  On the left, we have a much higher percentage of “would have bought anyway”, which decreases as we move to the right.  The portion of the blue bars above the red line represent buying activity in the control group – they bought without receiving an e-mail.

As we move to the right, this effect decreases, until the blue bars are now below the red line.  The space between the blue bar and the red line represents sales made because of your e-mail that were not tracked back to the e-mail drop.  Often, this is a result of simply not tracking the “campaign tail” for long enough, which is difficult to do without using a Control Group to find the long-term lift.

The implication: for commerce, you should be sending a different message to these different behavioral segments depending on where they are in the LifeCycle if you want to maximize profit.  In the segments with highest likelihood to buy, you should take it easy on the discounts; one way to optimize commerce profit across the entire engagement spectrum is to use a Discount Ladder

For all other business models, it’s highly likely that you could benefit from the same approach, if you have clear value KPI’s and understand this dis-Engagement process.

Now, I am well aware the above sounds insane to offline retail folks.  Most if not all of you lack the data to measure these effects, but that doesn’t mean they don’t exist.  For the folks that do have the data, the day will come.  Hey the web is interactive, the web is different, right?  Well, yes it is, so why measure the effect of promotions like they do offline if you have a superior method of optimizing for profit?

When a visitor / customer is highly Engaged, they often generate visits or sales without needing any Marketing at all.  That’s what the Relationship Marketing Strategy is all about – the Relationship drives the business.  That’s why, for example, people have found that including a lot of relevant and customer-focused content in a commerce newsletter gets higher response rates than just sending people coupons.  It’s why creating a new customer kit drives higher repeat purchase rates – it’s the Relationship building.  And that’s why you will find (if you use Control Groups) that for the highly Engaged, your e-mail program is taking credit for sales it did not generate, and that if you are providing discounts to the highly Engaged, you are probably wasting money on them.

A portion of the Engaged segment would have bought anyway, and the fact that you dropped a coupon in their lap with e-mail is simply coincidence.  Or, if you send a coupon every week on the same day, the customer simply waited for the coupon they knew they were going to get so they could make a discounted purchase they would have made anyway at full margin.  That’s how e-mail takes credit for sales it does not generate, and anybody who is managing to ROI / profit should care deeply about this.

Now, if you’re a “share” thinker, this subsidy cost related to “would have bought anyway” doesn’t matter to you, because any sale the other guy didn’t make is a good sale.  But last time I looked, you can’t put share in a bank account, and the logical extension of this share mindset is you can get 100% share by selling product below cost, so I have never understood it.  If your directive is increasing gross sales, that’s pretty much the same thing – you get there by unproductive ad spending, which in the end is the same thing to the bottom line as selling product below cost.

By the way, I’m not saying the Engaged should receive no communications, but they should get a different kind of communications tailored to their behavioral state.

On the flip side, no matter what your directive, you should care about not getting credit for sales your e-mail generated.  E-mail to another segment, those in the process of  dis-Engaging, almost certainly generates sales you are not tracking and not crediting to e-mail.  And that’s because the dis-Engaging are changing their behavior with the company for some reason.  They are seeking alternate channels, for example.  In other words, they are responding to your e-mail but they are not responding through your e-mail, they are not using whatever devices or links you give them in the e-mail but are still making a purchase because of the e-mail.  Again, you don’t see this unless you use Control Groups.

A third segment, the dis-Engaged, doesn’t respond to your e-mails at all.  And they’re not going to, because your company is now irrelevant to them.  The company has not been tracking the dis-Engagement process so it didn’t take any specialized action to slow or stop the dis-Engagement.  In fact, the company is probably just damaging their Brand by sending these folks any e-mail at all.

This is Relationship Marketing Strategy; it completely redefines how you communicate with customers based on where they are in the Engagement / dis-Engagement cycle.  And it works amazingly well.  The bottom line is customers remain customers longer – this Strategy tends to extend the LifeCycle – with the result customers end up with higher LifeTime Value.

Are you surprised?  You shouldn’t be.  People talk about this incessantly on the web all the time, don’t they?  Relevance?  Customer centricity?  Customer experience?  Blah blah blah? 

Then how come so few people are using a Relationship Marketing Strategy?  How come so few are using Control Groups to measure the true net influence of e-mail?  Why are people blasting out the same irrelevant message to all their customers once a week?  A lot of talk and very little action, methinks.  Perhaps you just needed a framework to put everything into perspective, a roadmap to getting it done? 

Now, I realize many folks in the community don’t have the tools they need to measure dis-Engagement; typically only the high-end tools have metrics like Recency and Latency and even though Google Analytics tracks Recency, it isn’t easy to do much actionable segmentation for that metric in the tool. 

There’s a very simple reason for this, if you think about it – it’s a lot harder to measure something that doesn’t happen than measure something that does happen.  After all, servers are all about requests, they’re not really thinking about “did not request”, if you know what I mean.  That job takes a database that’s remembering the last date a request was made forward in time, and calculating “did not request”.  That capability is a lot more expensive, at least for now.

But, I hope the “tool problem” doesn’t mean the community will ignore the concept of dis-Engagement while screaming to the skies about how important Engagement is to measure.  Those of you with access to a transactional database don’t have to wait for web analytics tools, you can profile customers for Engagement and the dis-Engagement process right in the transactional database with a simple query tool.  And I hope you now have a Strategic framework to think about why dis-Engagement is so important, at least from a Marketing perspective, so when you get your hands on that high-end tool, you will know exactly what to do with it.

If you really take some time to think about the ramifications of the Relationship Marketing Strategy that Engagement is a Tactical part of, you just might come to believe that dis-Engagement is even more important to measure than Engagement.

To summarize this series, the idea of Engagement, and a lot of notions surrounding it – customer centricity, relevance, and customer experience – are concepts within a Marketing Strategy known as Relationship Marketing that tosses out calendar-based communications in favor of communications based on the customer’s relationship with the company.  The ability to do this depends on an understanding of the Customer LifeCycle, the results of each interaction between the customer and the company over time.  The LifeCycle is tracked using various Engagement metrics, including dis-Engagement, which typically is the first sign of a problem with the customer Relationship.

If your company is having trouble understanding why Engagement is important to measure, perhaps it’s because senior management lacks the context of the Customer LifeCycle for taking action and the strategy of Relationship Marketing as a game plan.  Maybe you should send a link to the Wikipedia definition of Relationship Marketing to the CMO or CEO and ask, “Is this what you want?  Because if you do want this, I know how to measure the success of it”.

Questions or comments on e-mail engagement and response? 

Have you ever heard of the strategy called Relationship Marketing?

Share:  twittergoogle_plusredditlinkedintumblrmail


Follow:  twitterlinkedinrss

A Framework for Engagement – Implementation

Like I said in the last post, I’m sure there are quite a few different reasons why folks want to measure “Engagement”, and not all of them have to do with Marketing.  But if you are talking about Engagement as a metric to be used in Marketing, now you have the complete framework for why (as opposed to how) it is so important to measure Engagement – to define the LifeCycle of the customer, in order to communicate and act in the most customer-centric, relevant way possible.  Logically, this approach drives higher profits.

The LifeCycle is about both Engagement and dis-Engagement.  If you are in the Marketing camp, you can’t just talk about measuring Engagement.  After all, if Engagement is really important and valuable, then dis-Engagement has to be really important and valuable in the opposite way – it’s a bad thing.  Dis-Engagement means, literally, that your company is no longer relevant to the customer. 

In some businesses, online display advertising for example, it’s not clear that dis-engagement really matters, at least in the current model from the perspective of the advertiser.  Hey, an impression is an impression, right?  Who cares what happens after that.  At least they’re talking about some kind of engagement metric – Duration – which should relate to the quality or the likelihood of an impression.  Not much more they can do, in my opinion, for that business model.  But from the perspective of the site owner the ads run on, dis-engagement should be a big deal – especially if you paid something to get that visitor to come to your site in the first place.

So, we have Engagement, and we have dis-Engagement, which it seems nobody ever talks about.  I sincerely hope that changes in the future as we move forward.

Now, how do we track the LifeCycle, how do we actually implement?  It’s really very simple in concept:

1. Define / Measure Engagement – any way you want to, as appropriate for your business; whatever activity or combinations of activity you feel appropriate

2. Measure dis-Engagement – the absence of Engagement, as in the visitor / customer stopped doing whatever it is you define as Engagement for your business model

3. Take some kind of Marketing or Service action to slow or reverse the dis-Engagement with dis-Engaging folks

That’s not very hard, is it?  No.  If you’re looking for some kind of model to follow for planning and managing LifeCycle communications, take a look at Satama’s REAN.

However, even when you get the LifeCycle and learn to react to it, the system is not optimized yet.  The Relationship Marketing Strategy is not optimized until you start predicting dis-Engagement, and taking action to try and re-Engage the customer before they completely dis-Engage.  Because once your company becomes completely irrelevant, it’s very hard to change that for the visitor / customer – much harder than if you act before or when the dis-Engagement is occurring.  You can’t have an “annual re-Engagement campaign” and fix this – you have to fix it as it is happening, meaning you throw out all calendar-based communication and communicate based on where individuals or segments are in the LifeCycle.

Fortunately, dis-Engagement is usually a process – unless the company screwed up in a really big way.  And this dis-Engagement process is fairly uniform and actually quite easy to predict with simple tools.  The company most often becomes irrelevant to the visitor / customer over time.  In other words, the company gets second chances, the customer often gives the company leeway to become relevant again.  So as a company or analyst, the key is to:

1. Recognize dis-Engagement has begun with a customer or segment

2. Have a re-Engagement plan and implement the plan before the company becomes irrelevant to the customer

I’m pretty sure most people reading this know what comes next – how to measure dis-Engagement and act on it – given I have plastered this information all over my blog and web site.  If you don’t know how to predict dis-Engagement and the triggers you can use to take action, this is a good place to start.  Depending on your business model, you should probably also take a look at what Theo proposes in terms of Kind and Degree for survey work once you have dis-Engagement behavior as a trigger for the survey.

For most web sites, regardless of what you are using as a metric for Engagement, a good clue the dis-engagement process has begun is when a visitor stops visiting, posting, commenting, buying, or whatever is key to generating value on your site.  The challenge is you have to recognize this non-event has occurred right away, because the longer you wait to try and re-Enage the visitor / customer, or ask why they are dis-Engaging though a survey, the less likely it is you will be successful.  And in case you are wondering, those of you with e-mail tactics that consist of relentlessly pounding your list with the same messages and offers regardless of visitor / customer behavior are not addressing the re-Engagement issue – trust me.  Think about it.

I’ll get into how a Relationship Marketing strategy affects e-mail marketing and measurement tomorrow in the next post.

Does the above make sense to you?  Questions?  Criticism?  Problems?  Let me know, leave a comment.

Share:  twittergoogle_plusredditlinkedintumblrmail


Follow:  twitterlinkedinrss