the red papers- learning to read the river

8/2/2019 The Red Papers- Learning to Read the River

1/26

The Red Papers:

Ogilvy &Mather

Learning toRead the River

Dimitri Maex


2/26

The Red Papers:

Ogilvy &Mather

Learning toRead the River

Dimitri Maex


3/26

The Red Papers:

Contents

Introduction

History o MathMarketingThe early days o direct response

Marketing goes massMr. Company, Id like you to meetMs. ConsumerThe golden age o Math Marketing

Making MathMarketing workor youErgonomic measurementThree-step insightThe new single enterprise viewOrganizing or Math Marketing

Conclusion

Key takeaways

Reerences

6

10

10

1213

14

16

18222630

38

40

42

February 2010, No. 1


4/26

4

The Red Papers:

Using Math Marketing to increaseyour return on investment

Learning to Read the River:


5/26

6

The Red Papers:

7

Learning to Read the River

My employer, i it wanted to, could track every hour I spend, not onlythrough the electronic timesheets I ll out but also via the badge I haveto use to open the door at every oor in the building. All o my computerdata goes through their system, and its a snap or them to check my onlinecalendar and track the activity on my laptop. Because I use a bonus cardwhen I shop, the supermarket knows it was pasta night at the Maexs. II had connected to Xbox Live (which I no longer do since I was verballyabused the rst time I used it by a twelve-year-old rom Wisconsin whowasnt impressed with my gaming skills), then Microsot would know that

I am one o the many thirty-plus players hooked on Rock Band. Theycould even classiy me as a social gamer by matching me up with my wiesperormance on the guitar. HBO and Time Warner Cable know everythingI watch on TV, and thereore, probably know me better than I know mysel.

These incredible volumes o data are waiting to be explored by companieswho seek greater demand or their products and services through a betterunderstanding o what their customers do, think and eel. The inormationthey gather, but hardly use, could give them valuable insights into howand why their marketing succeeds and ails. Knowing that would pay oin a dramatic improvement in return on investment. Mathematics, still aneglected skill in most marketing departments, is the only tool able toextract crucial learnings rom the masses o data every person generatesevery day. The age o Math Marketing is here.

Yesterday was just a normal day. When the alarm on my cell phone blastedme out o a dream at 6:15, I umbled or the snooze button to get a ewmore precious minutes o sleep. A ew winks later, I rolled out o bed andlurched rom the shower to the closet and then headed to the kitchen orbreakast. Thus rereshed, I stepped out o my apartment, picked up theNew York Times rom my doorstep and took the subway to work, all thewhile listening to the latest Brit pop sensation on my iPod. Once at theoce, I headed to the gym downstairs where I ran or hal an hour on atreadmill, staring slack-jawed at the ESPN screen in ront o me. My next

ten hours ashed by as I ran around the oce to various meetings, madea ew phone calls and ignored a ew others, checked my BlackBerry waymore than I should have and pecked away at my laptop writing emailsand PowerPoint presentations pausing or a spell o intense concentra-tion as I built a spreadsheet in Excel. Come evening, I took the subwayhome but got out one stop early so I could hit the supermarket to buypasta dinners linguini with clams or my wie and a hearty Bologneseor me. We ate, played a bit o Rock Band on the Xbox 360, watched anepisode o True Blood on HBO, and went to bed. Lights out: 11 p.m.

During this unremarkable day, I let behind a rich trace o data, most oit unconnected with time in ront o a computer. Not only can AT&T seewhen I woke up or at least tried to they can ollow me all around thecity that day. The New York Times, i its clever, will see that I still like toget my newspaper delivered in the morning despite my habitual check otheir website every day. The Metropolitan Transit Authority can count meas another passenger on the overcrowded L train rom Brooklyn toManhattan. The next time I sync my iPod, Apple will see that Im grooving

on the Arctic Monkeys album I bought rom iTunes more than some omy other recent purchases, and they ought to use that inormation tomake better recommendations next time. (Lady GaGa? Not or me.)I would love or the gym at work to have a swipe card that tracks and summa-rizes my workout. I would be able to see my recent orm and get suggestionson improving my regimen while the gym gathers a ton o inormation aboutmy health and workout habits. Yes, I preer to watch ESPN while I run, andno, I havent yet added a weights circuit even though I know I ought to.

Introduction


6/26

8

The Red Papers:

9


Incredible volumes o data are waiting to beexplored by companies who seek greater demand

or their products and services.


7/26

11

The early days o direct response

Aaron Montgomery Ward invented the mail order catalog in 1872.With that single sheet o eight by twelve paper, he also, albeit unwittingly,invented Math Marketing. Wards breakthrough, soon bloated to over500 pages, led Richard Sears and Alvah Roebuck to launch their owncatalog in 1886. For the rst time, large merchants had an objective meansto measure their marketing success and optimize their operations. We donot know exactly how Ward, Sears and Roebuck did it, but we know theydid a good job. Ater all, both catalogs are still around today!

The catalog merchants and other direct response advertisers usheredin an era o utopian optimism in the new science o marketing.To contemporary eyes the problem o successul advertising had beensolved. The time has come when advertising has in some hands reachedthe status o a science, wrote a smugly sel-satised Claude Hopkins in1923. It is based on xed principles and is reasonably exact. The causesand eects have been analyzed until they are well understood. The correctmethods o procedure have been proved and established. We know whatis most eective, and we act on basic laws. For a ew years he and his ellowcommentators were on target. It took the invention o mass media toprove them wrong.

Direct response

Mass media

Customer centricity

Digital

1950 1990 2000 2008

10

THE RED PAPERS

History o Math Marketing


8/26

12

The Red Papers:

13


It took another ew decades, but, by the 1990s, marketers laid their handson another batch o cool toys to play with. Computers. The plungingcosts o computing power and inormation storage made it possible tocapture, store and analyze vast amounts o data. Enough data, in act,to know a little something about every one o their customers, especiallytheir most valuable ones. Businesses invested heavily in data warehousingtechnology to store all customer inormation in one database, andmarketing entered the age o customer relationship management (CRM).Direct marketing transormed rom a blunt, one-way sales channel into

an elegant means or companies to stay in touch with their customers, nomatter how they bought. Marketers were reed rom total reliance on theimprecision o mass media measurement. Using these tools, FrederickReichheld demonstrated that a ve percent improvement in customerretention yields a twenty-ve to one hundred percent increase in prot.Garth Hallberg ollowed up by showing that a small proportion o theaverage companys customer base represents a disproportionate share orevenue. No wonder marketers poured their lives and their dollars intothese new tools.

Now that companies knew how much their best and most loyal customersmeant to the bottom line, businesses strived to gure out who these goldmines were and how to keep them in the old. They subjected singlecustomer views to rigorous analysis o transactions, value, response ratesand even demographics. Established statistical techniques such as logisticregressions and discriminant analysis enabled marketers to predict theircustomers next move, and other techniques classied customers accordingto recency, requency and monetary value o purchases (RFM). Lietime

value models predicted what a customer would be worth to the companyover that individuals lietime, and anti-attrition models assessed the likeli-hood a buyer would drit away.

Soon, the exponential growth in computing power brought new data-mining techniques, machine learning and articial intelligence into themarketing toolkit. The stage was set or the vast quantities o data derivedrom digital media. Math Marketing had nally come into its own.

Mr. Company, Id like you to meet Ms. ConsumerMarketing goes mass

While Americas rst radio station, KDKA, went on air back in 1920,it took until the 1950s or marketers to approach broadcast media withanything like the mathematical precision they once enjoyed. Anyone witha pulse could see the appeal in reaching millions o people at one time,but guring out what worked and what didnt remained opaque. Massmarketing could not be measured with the old tools, and so leading edgemarketers turned to operations research and management science modelsto gauge eectiveness. Data was scarce, and the application o even thelatest scientic methods had its limitations. Absent unimpeachable acts,

marketers turned to econometric modeling to understand better theimpact o broadcast marketing on brand awareness, consideration and,ultimately, sales and prot. It worked surprisingly well. Econometricmodeling still helps modern marketers judge the eectiveness o massmedia. I anything, econometric modeling is underused today, and, asWarren Buet might tell you, an undervalued asset can be a powerulmoney maker.

Econometric modeling is underused today, and,as Warren Buet might tell you, an undervaluedasset can be a powerul money maker.

Historyo Math

Marketing


9/26

14

The Red Papers:

15


My average day is a trail o digital spoor or aclever marketer, eager to bag my next purchase.

The golden age o Math Marketing

My average day has gone rom being an evanescent journey o interestto me and my loved ones alone to a trail o digital spoor or a clevermarketer, eager to bag my next purchase, to ollow. The easy stuis already being done. Closed-loop systems, such as e-commerceenvironments, are as close as were going to get to marketing nirvana.Digital traces shows us exactly which media individuals have seen. Web-site tracking notes where individuals came rom (or, in the case o search,what terms they typed in to arrive at a site). We can observe an individualsentire shopping experience, all the way to their actual conversion to a sale.

With more and more media becoming digital, soon most, i not all, mediaexposures will be traced to an individual sale. The world has become ourocus group. Now imagine what could happen i you could learn to readthe river o data all o us leave behind every day, even when were not sit-ting at a computer.

Math Marketings possibilities seem endless, and they are. They are alsoeasily dened: Math Marketing can increase your return on investmentthrough a thorough appreciation o the eectiveness o all your marketingeorts and a undamental, transormational understanding o yourcustomer. Bringing those benets into your organization requires somesignicant organizational and philosophical adjustments. Fortunately,others have already made the rookie errors. I have learned rom theirmistakes and my own and am condent that you can make MathMarketing work or you.

Everything has become measurable, even mass media. Everythinggenerates data, and the volumes are enormous. Googles digital data-

base, perhaps the worlds largest, captures almost ten billion searches permonth. With these huge quantities o data, we have unprecedentedinsight into how our customers engage with brands and how thatengagement leads to revenue. Thats the good side. O n the otherhand, what had been a nourishing trickle o data has turned into a roaringtorrent. The trick, as any whitewater guide could tell you, is learning toread the river.

Historyo Math

Marketing


10/26

16

The Red Papers:

17


AccountabilityErgonomic measurement

InsightPrioritization, personalization

and precision

PeopleMath Marketing organization

DataNew single enterprise view

You dont go plunging down some terriying class ve rapids without proper preparation.Yet that is exactly what many companies do when aced with the sudden inux o beguilingdata owing into their servers. Its no wonder they get swamped or capsize. Moving yourorganization into top Math Marketing orm requires careul preparation. This our-step

structure will help make the transition smoother.

Making Math Marketing work or you


11/26

18

The Red Papers:

19


Use a mix o nancial and non-nancial metricsThe last couple o years have seen a number o eorts by marketingproessionals and academics to summarize marketings perormance inone metric. Some o these metrics, such as Reichhelds Net PromoterScore, Peppers & Rogers Return on Customer and Doyles Increase inShareholder Value, can be useul perormance indicators, but they willnever oer a ull explanation o the spectrum o your marketing. At thesame time, nance departments are orcing marketing departments toadopt nancial metrics. While this is a good thing, ocusing on nancialmeasures to the exclusion o all others is not. Counting the money is

not enough; marketers (and all non-marketing decision makers) needto understand where the revenue comes rom. IBM does this well. Inaddition to their nancial calculations, the IBM marketers measure howtheir eorts inuence brand awareness and image. Their innovativeFavorable Selling Environment Index assesses how receptive theenvironment is to their sales orce.

Use statistical modeling to identiy drivers o business perormanceSometimes the basics arent enough. Getting an understanding o theimpact o various elements o the marketing and media mix on outcomessuch as brand awareness, consideration and, ultimately, sales and protcan be a complicated task, especially in todays multichannel world.

Advanced and proven techniques such as econometric modeling canreveal the impact o dierent marketing investment levels and mixes. Theinsights you gain will lead to recommendations or allocating, dispersingand timing your TV, radio, print, out-o-home, search, display and othermedia investments. Econometric modeling has been around or over hala century, and its p ower in helping marketers understand what works

and what doesnt is well established. Despite that, its use is relatively limited.The UKs IPA awards set the international gold standard or advertisingmaterial, but only teen percent o the case studies submitted usedeconometric modeling to assess the eects o the campaigns.

Ergonomic measurementMath Marketing will help you understand how your marketing activities impact prots andhow you can optimize that marketing by doing more o what works and less o what doesnt.

And the way to do that is with measurement. Now heres the unny thing about the state omarketing measurement in 2010: almost everything is measurable. Yet plenty o surveyso marketers suggest that measurement is their main challenge. They get lost in the delugeo metrics. I see their thousand-yard stares every day. The cure is what I call ergonomicmeasurement. Ergonomics concerns itsel with understanding and improving the interac-tions among humans and everything else, and we have a crying need or it in marketingmeasurement. We need systems that do or measurement what the iPod did or mp3 players.

Despite the high degree o technical complexity under the hood, the user experience is simpleand elegant. Ergonomic measurement is a process o constant renement, but building onthis ramework will get you started o right.

Begin with the basicsI you want to understand the impact o your marketing eorts, you muststart with measurement rst, moving on to analysis o the results aterthat. Only then can you use the insights you generate to optimize yourmarketing eorts. With marketing budgets under increased pressure,many companies skip measurement and analysis and shoot or instantoptimization. This never works. You need the oundation o metrics andinsights beore you can optimize. To do otherwise will cost much more inthe long term.

Align metrics to objectivesMeasure what you need to measure, not what you can measure. Onlyocus on the data that aligns with your goals. All-but-limitless storageenables you to store everything you learn, but siting through data

unrelated to your business goals distracts your marketers. Consult withthe other business units in your organization to ensure that the metricsyou measure dovetail with their objectives. Mapping metrics to objectivesoten improves the quality o the objectives themselves.

Making

MathMarketing

work

for you


12/26

20

The Red Papers:

21


Build a marketing dashboard

Dont keep your marketing story to yoursel. Parcel out the marketingintelligence you gather to marketing and non-marketing olks alike.Once this practice required lengthy, highly structured reports, but nowelectronic marketing dashboards provide visible and tangible sharing oyour measurement eorts. Dashboards oten become the collaborativeorum through which ergonomic measurement comes to lie. The design o adashboard leads to a conversation about what metrics are important, which,in turn, orces you to align your metrics to your objectives. You will needto populate the dashboard, and that will spur you to post up nancial and

non-nancial metrics as well as the more sophisticated statistical modelsyour organization can use to optimize its marketing. Dashboards bringthe data to the desktops o the decision makers. Having a dashboard initsel will not make an organization more ecient and accountable, butthey provide the means to make that possible and are an essential elementin creating a culture o Math Marketing.

Making

MathMarketing

work

for you


13/26

22

The Red Papers:

23


These questions will determine whether you need a value-segmentation,an attitudinal segmentation, a lietime value model, an anti-attrition model,a browsing typology or any other Math Marketing tool at our disposaltoday. Answer these questions or others that you may pose through athree-step process: Prioritization, Personalization and Precision. In otherwords, nd out how you can prioritize your most valuable customers andprospects, personalize the way you communicate with them, and nd theright time and place to target them.

Do you know who your best customers are? How can you nd prospects with the same prole?

What drives your protability? What makes your customers and prospects engage with your brand? How should you communicate with customers and prospects?

Three-Step insightData is only as valuable as the insights you distill rom it. Those discoveries are whatMath Marketing brings to your organization. The valuable consumer insights MathMarketing uncovers can be turned into powerul ideas that uel uture growth. The rangeo techniques available is nearly as broad as the spectrum o data you can measure. Tobring order rom that swirl, start with the basics and prioritize your data mining eortsbased on the questions they will answer.

PrioritizationDetermining whom to target

Prioritizing your most protable and highest-potential customers andprospects ocuses resources where they really matter. Prioritization toolsdescribe how customers or prospects behave and how that translatesinto current or uture prots. They allow the marketer to rank targetsaccording to a positive business outcome. Prioritization tools such asdecile analysis, share o wallet segmentations, lietime value models andrecency-requency-monetary value schemes combine multiple behavioralattributes into one scale or ranking targets rom high to low importance.

Most traditional database segmentations all into this category.

Data is only as valuable as the insightsyou distill rom it.

Making

MathMarketing

work

for you


14/26

24

The Red Papers:

25


In a digital world, consumers oten reveal in real time what theirintentions are, and Math Marketing has taken advantage o that withan ever-growing suite o breakthroughs in precision targeting. Precisiontargeting tools enable the marketer to identiy individual customers andprospects at the right time and in the right place. A drill manuacturer,or example, wants to know whether a person has searched or drills onGoogle, whether he clicked on a banner or drills, or whether he saw adrill-related video online. These observed events are direct indications

o a consumers interests and desires at that moment. A whole industryhas sprung up oering Behavioral Targeting (BT) solutions based onthe instantaneous availability o this data. In its early days, BT generatedtwenty to thirty percent lits in advertising perormance. As a result, ithas become standard industry practice. Innovations in BT are shorteningthe lag between when the inormation is gathered and when it is actedupon. When someone searches or a drill on Google, that is very valuableinormation i I can target that person immediately. On the other hand,knowing that someone searched or a drill three months ago doesntdo me much good. The predictive power o sel-disclosed data starts todecline minutes ater the observed event. Future innovations will ocus oncapturing multiple events in real time, increasing the value o even olderdata, and shortening the cycles between observation and action. Real-timebuying, an early entrant into this eld, is the vanguard o what will nodoubt become a transormative phenomenon. In the digital world, timelyinsight equals prot.

PersonalizationDetermining what to target them with

In order to tailor your messaging, you need a thorough understanding owhy targets behave the way they do. Personalization tools uncover groupso targets who share similar needs, liestyles, demographics and manyother revealing attributes. Brand planning and media planning researchsegments tend to all into this group. Since most personalization toolsstart rom what consumers need, they can also be powerul drivers o newproduct development.

PrecisionDetermining where and when to target them

Consumers oten reveal in real time whattheir intentions are.

Making

MathMarketing

work

for you


15/26

26

The Red Papers:

27


Moving beyond the traditional single customer view

Many companies have been building single view corporate data ware-houses since the early CRM days. Creating these has been an expensiveand labor-intensive task. Most likely, you now have a valuable asset. Forexample, Tesco in the U.K. estimates that sixteen percent o their marginis attributable to the knowledge they get rom their customer database.That makes their database a $3.2 billion asset.

Tescos customers, and everyone elses as well, are generating massivequantities o data outside o the corporate systems. This data is owned

by platorms like Facebook, Twitter, Google, Microsot and Yahoo!, andcompanies need learn how to integrate it into the databases they alreadyhave in place. The data gathered in these new integrated databaseswill extend beyond the individual to the entire enterprise o selling.

Any measure that can generate consumer insights or improve marketingaccountability will be part o this new single enterprise view o the uture.

The new single enterprise viewData is the raw material or Math Marketing, and the volumes gathered every day arestaggering. Imagine or a moment that you had a database that held every word ever spokenby a human being, every utterance rom the dawn o language on to babble you hear around

you right now. Thats a damn big database, right? You would need twenty databases o thatsize to hold the data that will exist by the end o next year. And the amount o data we spoolout skyrockets every year! Not all o this data can be used by companies to improve theirmarketing eorts. Ater all, much o it is junk, but quite a bit o it can be helpul. Companiesthat adapt to the new scales o available data will thrive. Those that do not, well, no one

said that evolution was kind. Adapting organizations will adopt a single enterprise view

o their business, not to replace the traditional single customer view but to augment it. Thiswill allow them to access huge volumes o data, producing intricate analytics and targetedmessaging that has both mass economies o scale and single-customer precision.

Companies that adapt to the new scales oavailable data will thrive. Those that do not,well, no one said that evolution was kind.

Making

MathMarketing

work

for you


16/26

28

The Red Papers:

29


The new single enterprise view

Companies and a new entrant, third-party data clearinghouses, are devel-oping systems to integrate these data sets with existing data warehouses.To ully realize the potential inherent in the single enterprise view, thesenew solutions must include these new data sources: Search-intent data: Search data is rich in information. And its free.

Search intent modeling tools mine that data and eed back insight onthe search terms consumers use to nd your brands and products. Inaddition to the obvious benet o improved keyword targeting, this

analysis reveals the thought processes o your target audience. Knowingthe words they use and conjure to talk about your brand can improvethe eectiveness o your overall communication.

Listening posts: Social web platforms such as blogs, microblogs, forums,

social networks, and opinion and review sites are another easilyaccessible source o consumer data. Mining this data or insights maynot be quite as straightorward as it is with search-intent inormation,but it is worth the extra eort to learn what your consumers are saying.Listening post technology collects data rom the social web and perormssemantic analysis o the conversations happening there. From this,marketers can learn how many people are talking about a brand,whether those conversations express positive or negative opinions,and which other brands or characteristics are brought up during theexchange. Ogilvy used careul analysis o listening post data to help hotelchain Harrahs change the oers and language in their advertising eorts.In online reviews, customers raved about the views rom the hotel anding that would not have been picked up by any other means.

Harrahs now eatures those views in their ads, and that simple,transormative insight resulted in a iteen- to thirty-percentimprovement in return on their online advertising.

Primary research: If you want to know what consumers think, just askthem. That is how primary research works and why it has been the maindata source or generating market intelligence or a long time. Todaysprimary research is an entirely new discipline rom what it was even a

decade ago. Social communities and online survey tools have cut thecosts o primary research dramatically. Companies use these tools oridea generation, polling and even in-depth interviews. Instead o waitingweeks or months or research results to lter back in, decision makers canlearn almost instantaneously what their customers think. Careul cultivationo these research panels can open a eedback channel with engaged cus-tomers, providing marketers with a constant inux o resh perspective.

Results repository: A results repository to store all marketing metricsin one place is probably the most important asset a marketer can havein the quest or eciency and accountability. Though this sounds very

obvious, it surprises me how ew companies have one. The challengeis not technological that part is remarkably easy. The diicultyis in the discipline. With the constant pressure to do more with less,ew marke ters take the time to capture perormance metrics over time.This is especially true when it comes to cost metrics. Data on marketingcosts tends to be stored as batches in nancial systems making it useless orperormance analysis. By keeping a robust results repository includingboth perormance and cost data marketers can add a crucial layer oeciency analysis to their single enterprise view.

A new single enterprise view o your data can have as proound an impacton your company as single customer views did two decades ago. This crucialstrategic asset will soon become more than just a competitive advantage;it will be the mark o a healthy organization.

Making

MathMarketing

work

for you


17/26

30

The Red Papers:

31


Organizing or Math MarketingAll the Math Marketing in the world wont do you a bit o good unless your company isorganized to employ it. People, not numbers, not algorithms, make the dierence in successulMath Marketing. Throwing bodies at the problem will not work. A company needs the rightorganizational ramework and leadership mindset, which ew have in place at this time, toenjoy the benets o Math Marketing. To get this right without the expense or dislocationcaused by organizational turmoil, set up a base o skills, construct a Center o Excellence(COE), orge the right external partnerships and oster a pervasive Math Marketing culture.

Skills

Good Math Marketing talent is scarce. Math Marketers must be ableto use advanced mathematical techniques and explain their ndings in amarketing context to a nontechnical audience. I youve ever hung aroundwith a bunch o mathematicians, you know this is an extremely rarecombination o talents.

The last 15 years have seen an explosion o analytical tools. In order tomaster them, Math Marketers need a staggering variety o skills.

I your organization wants to build out their technical Math Marketingcapabilities, the skills matrix on the next page is a good road map. Butother talents are also required to make Math Marketing work. Chieamong them is the art o making the complex seem simple. Math Marketersspecialize in analyzing the vast amounts o data in our digital world. Bytemperament and culture, they oten preer the company o like-mindedspecialists with whom they can talk shop without having to explain them-selves every other sentence. When youre in the weeds o complex problems,simplicity seems irrelevant and reductive. Yet simple, elegant explanationsare essential i Math Marketers are to have any impact on the organizationoutside o their own priesthood. Training Math Marketers to embraceboth the complexity o data and the simplicity o actionable insights isnt

just goo d politic s it s the ultimate goal o the work it sel. Thereare plenty o techniques and tools that can help Math Marketingbecome more actionable. These range rom personas and customerportraits to rulebooks and guidelines. Simplicity is key or all o them.

Making

MathMarketing

work

for you

People, not numbers, not algorithms, make thedierence in successul Math Marketing.


18/26

32

The Red Papers:

33


Businessplanning

Budget setting and allocation

Business case development

Scenario planning

Online media analytics

Online behavioral targeting

Social media analytics

Online mediaanalytics

Campaign perormance measurement

Marketing mix modeling

Media mix modeling

Econometricmodeling

Perormance reporting

Dashboard practice

360-Degree measurementand reporting

Segmentation

Predictive modeling

Targeting

Website reporting

Site testing and optimization

Web analytics

Tracking studies

Online research

Quantitative research

Making

MathMarketing

work

for you

Center o excellence

Many companies sensibly nurture advanced Math Marketing skills withintheir organization by establishing a Center o Excellence (COE) or MathMarketing. This group usually resides within a marketing intelligencegroup that, in turn, oten sits within a broader strategy group. Mosto todays marketing intelligence groups ocus on market researchand econometric modeling. For them, most, i not all, o the more modernMath Marketing skills are novel. These departments and their leaders grewand matured during the years o mass marketing, when research andeconometrics were ascendant.

Todays Math Marketing COE must consist o at least one Web analyst,a website optimizer, a social metrics expert, a database marketer, a searchanalyst, a quantitative market analyst, a qualitative market researcher,a media analyst, a digital media analytics expert, an audience researcher,an econometrician, a data miner and a PR measurement specialist.Since ew companies can sustain an internal COE with such a degreeo s pecial ization, many o them look outside or partners who bringsome, i not most, o these capabilities.

These are the abilities that Ogilvys Analytics department, a group o more than200 Math Marketing specialists worldwide, shares with clients.


19/26

34

The Red Papers:

35


The current Math Marketing landscape, as you can see above, is lacking a clear top dog.

BroadAperture

Specialty Skills

A true Math Marketing partner would have a combination o broadaperture vision and sharp speciality skills. This partner would appear in theupper right corner. No one is there at the moment. The main contenders management consultants, integrated communications agencies, mediaagencies, Google and Microsot will soon be ghting or dominance inthat top spot. That competition will be great news or companies eager toset up their COE.

ManagementConsultants

CommunicationsAgencies

MediaAgencies

PerormanceMarketing GoogleMicrosotYahoo!

CRMVendors

ResearchAgencies

Digital AnalyticsVendors

Making

MathMarketing

work

for you

External partnerships

The role o external partners will only grow larger as Math Marketingevolves. Companies eager to acquire Math Marketing by way o outsourc-ing need to establish the right organizational ramework to support theserelationships. On the other side, the ideal Math Marketing partner o theuture needs to be strong across two dimensions.

Broad Aperture: Math Marketers must comprehend the marketing land-scape in its entirety, understanding all media and all elements o themarketing mix. They should understand the impact o marketing on

overall business results and have the senior-level client relationships thatenable them to take action.

Specialty Skills: Math Marketers must have the specialty skills necessary tomaster the modern Math Marketing toolkit. They include econometricmodeling, data mining, statistics, web analytics, online ad servingcompetencies, quantitative market research techniques, dashboard andvisualization technologies, and social media analytics. They need theintellectual exibility to grasp and implement the latest developments inthe Math Marketing industry.


20/26

36

The Red Papers:

37


and expertise to themselves, securing their place as indispensable wizards.Not only does this undermine the very mission o the COE, it also sets uptension between the COE and other parts o the marketing organization.To prevent this rom happening, make clear that the main goal o theCOE is to drive adoption o the insights it produces. Measure this acrossthe organization and reward decision making based on Math Marketing.Members o the COE in particular should understand that their success, aswell as that o their colleagues, depends on communicating their expertise.Write it into the personal perormance goals o COE members and

incentivize them accord ingly.

Math Marketing is a rapidly evolving discipline, and the innovationspushing its growth come rom practitioners. COE members should begiven the reedom and the time to discover new ways o improving thecompanys return on investment while they sling their own proessionorward. Urge your COE to automate the cutting-edge work they do now,making it common practice. The more they do this, the more timetheyll have to reinvent themselves, especially i you maintain anongoing investment in training and research and development tosupport them.

Communicate your business case to thedecision makers and then translate the

action items to the people on the ront line.

Culture

As daunting as the talent crunch may seem, the biggest challenge companiesace is internal. Eective Math Marketing requires a culture shit romthe most senior leadership on down. Marketers need to be seen as ullparticipants in an organizations nancial and creative decision makingrather than nding themselves kept to the side as a murky but necessarycost center. Brilliant insights wont help a company that neglectsto implement them, and Math Marketers, no matter their rank, are thebest choice to make their case to the C-Suite ocers; they speak the samelanguage. Already, marketing departments are under pressure to

hire high-potential marketers who understand return on investment andanalytics. With the average tenure o CMOs now at 28 months, it wontbe long until a new generation o marketers, well-versed in the workingso Math Marketing, rise to the top.

Empowering the Math Marketers can help companies strike the rightbalance between analysis and action. Math Marketing should ocus oneducating the end users, particularly those outside o the senior leadership.Most companies spend ar too little time on this. They put all their eortinto generating insights, ormatting them into simple, easy-to-digestormats or the bosses to absorb and then let their work languish on theshel o the head o marketing intelligence. Communicate your businesscase to the decision makers and then translate the action items to thepeople on the ront line. These can be customer service representatives,new product development engineers, creative and design teams, or anyoneelse who could end up beneting rom your insights. Involving them earlyin the process can help hone your eorts and will increase the adoptionand use o those insights you generate together.

The biggest cultural challenge o all oten comes rom within your MathMarketing department itsel. The main objective o a COE is to makedecision makers outside o it smarter. The COE provides them with theinsights, knowledge and tools to make better decisions independently.I a COE is successul, it will make itsel obsolete. This, o course, haspolitical implications as very ew COE teams enjoy working toward theirown obsolescence. Most react to this threat deensively, keeping knowledge

Making

MathMarketing

work

for you


21/26

38

The Red Papers:

39


Huge chunks o the advertising that people sit through has little or noconnection to their lives. It is disruptive and oten outright annoying, butis xing that worth sacricing a measure o your privacy? Math Marketingcan make advertising content more relevant to the consumer, increasingthe eectiveness o individual ads. What i I told you that knowing some-thing about you improves eectiveness so much that I can cut the amount oadvertising you have to put up with by hal? Now Ive got your attention.

Ads that are personalized provide a better return or the advertiser, andadvertisers will pay more or a personalized thirty second ad than or a

general one. Youll see ewer ads, or which the content providers receivedmore revenue, and those ads will be relevant to your lie.

Consumers are the reason all o us go to work each day. They pay our salaries,and when you get right d own to it, theyre really i n charge. We needto educate them, preparing the way or widespread acceptance o MathMarketing, just as we will within individual organizations. The advertisingworld we create together will benet us all.

I never thought o my proession as controversial. Marketing thatsa middle-o-road gig, right? With the growing sophistication o digitalmedia, things have changed. Now when I talk to my amily and riendsabout the details o what I do or a living, I put on welding goggles rst:Youre collecting all that data about me?!? Thats creepy! Consumersare rebelling against widespread data collection or commercial purposes,and they have valid reasons. Data gathering is largely unregulated andhidden, but things are changing. In the absence o other standards, theadvertising industry has begun to sel-regulate. We have pledged that

we will not collect personal identiable inormation (PII) without theconsent o the consumer, and we will be transparent about when and howwe use behavioral inormation or advertising purposes. We will give theconsumer the opportunity to opt out o any data collection, but we willalso educate consumers about our reasons or collecting this inormationin the rst place.

I hope that our code o conduct, when paired with consumer education,will help people understand that what may seem like an intrusion at rstis, in act, to their benet. I know that sounds Orwellian, but considerhow ew people outside o the marketing proession understand thatadvertising is needed to create content, especially in an age when consumersare expecting more and better ree material. Consumers have demonstratedagain and again that they will not pay ull reight or the media they enjoy.I the situation were reversed, every website would be behind a pay wall,and every television channel would be premium.

Advertising-supported media isnt going away, but Math Marketing can

make it much less o a nuisance in consumers lives. Lets say you are atwenty-ve-year-old gamer, and the TV networks need to expose you tothirty minutes o advertising a day to pay or the production o your avoriteshow. Would you rather have thirty minutes o advertising about the latestgames or thirty minutes o general advertising? Most gamers would preerthe rst scenario, but in return or that youre going to have to let the MathMarketers know something about your tastes, habits and preerences. Youmight be on the ence now.

Conclusion


22/26

40

The Red Papers:

41


1. Ergonomic measurement

Measure and analyze before you optimize Be sure to measure what you need to measure Use a mix of metrics Employ statistical modeling Build a marketing dashboard

2. Three-step insight Prioritization Personalization Precision

3. The new single enterprise view Move beyond single customer view Combine multiple streams of data in innovative ways Search intent Listening posts Primary research

4. Organize or Math Marketing Determine your talent needs Form a center of excellence Forge external partnerships Change your corporate cultures return on investment

Making Math Marketing work oryou requires our major steps:

Math Marketing will improve

return on investment advertising efciency the lives o consumers

Although Math Marketing seems

radically new, its roots go back tothe late nineteenth century

Key takeaways


23/26

43


The Red Papers

Ayres, Ian.Super Crunchers: Why Thinking-by-Numbers Is the New Way toBe Smart.Bantam, 2007.

Baker, Stephen.The Numerati.Houghton Mifin Harcourt, 2008.

Berry, Michael J. A.,

and Lino, Gordon.Mastering Data Mining: TheArt and Science o CustomerRelationship Management.Wiley, 1999.

Binet, Les, and Field, Peter.Marketing in the Era o

Accountability.World Advertising ResearchCenter Ltd., 2007.

Caples, John.Tested Advertising Methods.Prentice Hall, 1998(First edition, 1936).

Davenport, Thomas,

and Harris, Jeanne.Competing on Analytics: TheNew Science o Winning.Harvard Business School Press,2007.

Friedman, Thomas.The World Is Flat: A Brie Historyo the Twenty-First Century.Farrar, Straus and Giroux, 2006.

Hallberg, Garth.All Consumers Are Not CreatedEqual: The Dierential MarketingStrategy or Brand Loyalty andProts. Wiley, 1995.

Hopkins, Claude.Scientic Advertising.Waking Lion Press, 2008(First edition, 1923).

Kao, John.Innovation Nation: How AmericaIs Losing Its Innovation Edge,Why It Matters, and What WeCan Do to Get It Back.Free Press, 2007.

Moeller, Leslie, and Landry,Edward. The Four Pillars oProt-Driven Marketing: Howto Maximize Creativity, Account-ability, and ROI. McGraw-Hill,2008.

Reichheld, Frederick.The Loyalty Eect: The HiddenForce Behind Growth, Prots, andLasting Value.

Harvard Business School Press,1996.

Reerences


24/26

44

The Red Papers:

45


Dimitri Maex

Managing Director Consulting N. America

Head of Global Data Practice

Ogilvy & Mather Worldwide

Dimitri leads Ogilvys Consulting group in North America. This group consists

of approximately 100 consultants across Ogilvys CRM, Analytics and BusinessPlanning practices.

Dimitri began his career as a nancial controller at Kraft Foods, after which he

joined the Amsterdam Group (European Alcoholic Beverages Industry Group)

as a market analyst for the EU. He joined OgilvyOne Brussels in1998

as a data

analyst and has worked on various clients and on international product develop-

ment for the Ogilvy Group.

In 2001 Dimitri transferred to the London ofce to become Principal ofInternational Consulting, responsible for developing the consultancy offering for

international clients. He also became Ogilvys global analysis community manager,

a role he still fullls.

In 2004 Dimitri moved to San Francisco to work at Ciscos headquarters in San Jose. Concurrently, as associate director of marketing analytics at OgilvyOne,

he was responsible for developing the analytics department at Cisco Worldwide.

After his spell in Silicon Valley, he moved to the New York ofce to run the analytics

group which he helped double in size over the next three years.

In 2009 Dimitri was given the added responsibility of running Ogilvys NorthAmerican Consulting Group. He also became the head of Ogilvys Global Data

Practice, which made him responsible for developing Ogilvys data and analytics

offering globally.

Born and raised in Antwerp, Belgium, Dimitri studied econometrics at the

University of Antwerp. He then went abroad to get his MBA in marketing at the

Xavier Institute of Management in Bhubaneswar, India. His hobbies are tennis,

squash, soccer and music.

THE RED PAPERS

About the author


25/26

47

WriterDimitri Maex

EditorJeremy Katz

Creative directorGabe Usadel

Art directorAlex Fuller

IllustratorRod Hunting

TypeaceJ Baskerville

Design485 Brand Design, Ogilvy Chicago

February 2010Issue 1

The Red Papers are published by

Ogilvy & Mather

No article may be reproduced ortransmitted in any orm without thewritten permission o the publisher.

Copyright 2010Ogilvy & Mather

Edition1:5,350

Colophon

46

The Red Papers


26/26

the red papers- learning to read the river

Documents