SELMA AVDIC: Hello, everyone.Thank you so much for joining us today.My name is Selma Avdic.And on behalf of the American Marketing Association,I'd like to welcome you to our members only webcast.It's very exciting.It is sponsored by Kantar Vermeer, entitled, Bridging
SELMA AVDIC [continued]: Marketing and Analytics--Navigating a Disparate Data World.And so before we get started, I justwanted to cover a few housekeeping items.You will be provided with a recording of our presentationafter we conclude.And it will be made available to you as soon as possible.We encourage everyone to continue this conversation
SELMA AVDIC [continued]: on Twitter.And you can do that by adding or searchingfor #AMAMOW for our members only webcast.When tweeting about this webcast,if you have any technical or content-related questionstoday, please feel free to ask them at any time
SELMA AVDIC [continued]: by using the chat box on the left-hand side of your screen.And we'd be happy to assist in any way we can.So now, I'm very pleased to introduce our speakersfor today.And we're very honored to have such experts to sharetheir expertise with us.
SELMA AVDIC [continued]: And so first, we have Mario Simonwho is the chief executive officerof Millward Brown Vermeer.And he has a very extended resume.But I will just offer a little bit of information.He has led the global expansion of this practice, whichnow operates in the US, the UK, the Netherlands, Australia,
SELMA AVDIC [continued]: China, South Africa, United Arab Emirates, Mexico, Brazil,Singapore, and Japan.And we also have Bill Pink with uswho is the managing partner, Millward Brown Analytics.And in addition to this role, he alsoserved as chairman of Millward Brown's Global Marketing
SELMA AVDIC [continued]: Science Council.And so with that, I'm very happy to get us officially started.And I'll turn it over to Mario.
MARIO SIMON: Thank you very much, Selma,and good afternoon to everybody.So we are presenting here the studythat was just published on the Harvard Business Review, which,overall, is called Insights2020 and has a wealth of information
MARIO SIMON [continued]: in terms of understanding best practices that the insights aregoing to actually play in the future of our businesses.And it has been a very, very broad study.I'm going to take you through a few slides from that.And then we're going to go into deep dive.And Bill will help me.And I will help him sort of really give you a sense of,
MARIO SIMON [continued]: like, from the very organizationaltheoretical underpin all the way to the very actual and factualday to day on how to actually bring insights to the boardroomand make a huge impact based on the wealth of datathat we have today.
MARIO SIMON [continued]: So when we look at the world today rapidly changing,the scale of data has exploded in the last decadeor so to a level that has actually caughtalmost everyone by surprise, not only in termsof the quantity of the data, but also, the velocity of the data
MARIO SIMON [continued]: and the ability of it to really make meaningful changesto our lives.And most of the time, very, very, positive.And sometimes also actually quite destructive.And actually a huge set of practices and thinking
MARIO SIMON [continued]: around professional services has to do with, of course,the guardrailing of data security.So it does create a lot of uncertainty in the marketplace.It does create a lot of new kinds of threatsbut also huge opportunities as data is democratized very, verymuch accessible everywhere.
MARIO SIMON [continued]: And I think I'm not saying anythingthat is revolutionary here.But it sets the context for why we actually didthe study that I'm going to present to you withto begin with.And when we look at the value drivers in the 1900s to 1960,
MARIO SIMON [continued]: obviously, a huge amount of productivity gainswere based on manufacturing efficiency liberalizationof all sorts of supply chain and programs that's were the erathat the Six Sigma was created--the whole idea of really improving top quality
MARIO SIMON [continued]: in terms of production.And then for the next few decades, distributionand globalization became a huge driver of financial valueon the planet, which then was succeededby the next level of technology with the explosion
MARIO SIMON [continued]: of the internet and all the ancillary servicesthat IT helped the world.And today, we see that, actually,all of these things that have becomeenablers and have become table stakes.And now, the big next frontier for value creationis actually around the connective customer
MARIO SIMON [continued]: or connected consumer.And I actually speak about the customer and the consumeras interchangeable, because we areaddressing both B2B and B2C, like, absolute opportunities.So customer centricity is somethingthat was talked about in the past as a nice to have.
MARIO SIMON [continued]: Now, it is a have to have.But really understanding it in depthis actually very, very important.And so Insights2020 focused on whatare the drivers of customer centricityand how to achieve it in terms of our day-to-day functioningof our businesses.So we did a big effort on the heels of the Harvard Business
MARIO SIMON [continued]: Review article on the cover of July 2014,which was about marketing 2020, and where the talkis going around marketing.Now, it is about Insights2020.And so we had a great coalition of friends,
MARIO SIMON [continued]: including the American Marketing Association, who reallysupported us in terms of both publishing as partners,as well as founding partners, including our parent company,Kantar and our parent company Millward Brown.And it is the largest and most comprehensive studyin terms of insights analytics that has ever been conducted.
MARIO SIMON [continued]: With the type of methodology, we first did visual interviews.So over 200 visual interviews with academics,with CEOs, CMOs, heads of insides of large organizations,heads of agencies, different kindsof partners of large corporations and also startupsand small corporations to really understand
MARIO SIMON [continued]: and the scope of this study in what we need to actuallyunderstand more deeply.It was global, over 60 markets.And also, we talked to over 10,000 peoplewho were executives at those organizationsfor a quantitative survey to really understand
MARIO SIMON [continued]: the drivers of customer centricity.And we linked it to behavior analyticsfrom LinkedIn who was one of our partners,worked on crowdsourcing platform,and many, many research teams.And lots of hard work went into generating that study.
MARIO SIMON [continued]: And the first and most important findingis that if we look at revenue growth as also independentlyvalidated by a team at NYU, if we look at revenue growth,and then we look at customer centricity,there's a huge correlation between the companies thatare actually customer centric and actually
MARIO SIMON [continued]: being able to drive both absolute revenue growth,as well as in category relative to their peerset.And so the study really looks aroundthe underperformers and overperformersin terms of what distinguishes them and makes them special.
MARIO SIMON [continued]: In the Harvard Business Review article,you would be able to read all the top 10drivers of customer centricity.Here we're really zooming in on the onesthat then talk about bridging marketing analyticsand how that actually in practice happens.So a few excerpts from that study.One huge driver was that the overperformers
MARIO SIMON [continued]: create experiences based on data-driven insights.The way to read this chart is to saythat on this specific top driver of customer centricityand customer centric growth, what we seeis that 73% of overperformers said that they agreewith this statement in their organizations,
MARIO SIMON [continued]: versus 31% of underperformance.So a huge bias in preference towards creatingexperiences based on insights thatcome from the data for overperformers.A second very powerful one is around linking different datasources to distill insights.
MARIO SIMON [continued]: And so here we see that almost double the numberof overperformers say that they actually do that.And the example of LinkedIn is a really interesting one,because, not only were they a partner,but also, we saw that in practice in the abilityto not only have an amazing data set of people and members
MARIO SIMON [continued]: that they can mine the data and understand the databut also merge that data with demographic data with statisticfrom the different countries and understand patternsof labor migration of skills migrationand being able to advise large corporations, as wellas governments and NGOs on actually how to approach
MARIO SIMON [continued]: sourcing labor and what bets to make in the macrotrends.Then a third one is around customer centricityis fully embraced by all functions.And here, there's a huge people dimension-- people component.And the HB article has a much more depth
MARIO SIMON [continued]: around whole brain skill set that the modern marketand the modern insights executiveneeds to have to actually be able to translatethose insights and that data into action.But it is really about having customer centricity thatis embraced by all functions and actually
MARIO SIMON [continued]: is able to bring action in the boardroomthrough a stewardship of data by everyone in their organizationrather than being a siloed function.So those are the sort of big driversas it relates to specifically data integration
MARIO SIMON [continued]: on the Insights2020 article.And in terms of recommendations that have come out of it--and it's interesting that these recommendationsmight sound quite obvious to many of the people in the call.But the reality is that when we do the study,we see that still, many, many companies actually do not
MARIO SIMON [continued]: focus on these drivers of innovation and transformationfor their own company.So going from focusing on products and servicesto also focusing on the total experiencethat we offer as a brand or as a portfolio brands focusing
MARIO SIMON [continued]: on the detail of data, which, of course, has become, on the onehand, much more accurate, but on the other hand,much harder to do because of the volume and the velocityof data, so actually focusing on the ability of the data.And what are the big bets that our brands and companiesare doing.
MARIO SIMON [continued]: Focusing from managing risk to actually alsoincluding experimentation as a way of doing business.And this is an important one, because if wethink about most corporations, most corporationscreated and/or sustained through the ideathat risk is actually only a bad thing
MARIO SIMON [continued]: and has to be managed down.Therefore, there should not be any experimentation in the waysthat we do business.And what we see in the modern economyand through this fragmentation and explosion of data,actually, experimentation is a must-have capability
MARIO SIMON [continued]: in order to have a thriving brand and company.And finally from delivering to customers,which has been very much the mindset of many large brandsto co-creative and innovative with our customers.And when I say innovating and creating,
MARIO SIMON [continued]: I don't only mean in terms of adjacencies and the typicaland '40s way of thinking about reinventing marketingbut actually also innovating our business models and the waythat we serve our customer in very new and surprising ways.
MARIO SIMON [continued]: And so Insights2020 findings show that intelligence drivesoverperformance in terms of total experience in termsof co-creation in terms of actionability,as well as experimentation.And we've developed obviously solutionsthat a consultancy and as a research
MARIO SIMON [continued]: company to address and support our clientson that path of transformation based on data.So really connected, like, data intelligencecan be a huge platform for building scale and excellence.
MARIO SIMON [continued]: And I will pass it onto Bill, now,who would give us a sense of really how we execute thatspecifically, and how we're able to connect the dots to makea bigger impact in our brands.
WILLIAM PINK: Thanks, Mario.What Mario just walked everyone throughto really set the stage in your mindsfor how conducted data intelligenceis sort of playing out in market.You've seen sort of the trends, the macro-level trends.And there's very data-driven insightswe sort of pulled together to show how this
WILLIAM PINK [continued]: has a real business benefit.It isn't just a fancy term.I also did want to say that the phrase, connected dataintelligence, is something we've sort of put togetherquite purposefully, because there is tons of materialand webinars and all sorts of stuff being written and spokenabout, as the need to connect the dots, the needto tell an integrated story.
WILLIAM PINK [continued]: And what we've found is its phrase,"connected data intelligence", works quite well,because it takes everything you justsaw about how that does result in true business growth--what the implications are and from a strategy perspective.But it's one thing to say, throw a bunch of dataand research together.It's another thing to know how to do that.It's another thing to not have a point of view
WILLIAM PINK [continued]: around why you're doing that.And then the intelligence part reallycomes in, it can be not so easy.You have to think it through.You have to know what questions you're answering.You're now have to know a lot about the datasets you're leveraging.So you really do need that sort of whole mindsetof analytics research, understandingwhat you're trying to achieve to leverage connected
WILLIAM PINK [continued]: data intelligence and put it to action.And that's what we're going to walk throughin the rest of the presentation.We'll show you a bit of a frameworkthat we use to understand marketing effectiveness thatis basically reliant on having this kind of connected datato work.And eventually, we'll show you the framepieces of the framework in actionand then the whole framework in action.We're going to go through some high levels, some patterns
WILLIAM PINK [continued]: of relationships we found between looking at data sets,such as search and social sort of call it a more digital data,more modern passively observed data but see its relationshipwith branding campaign.We'll talk about some normative findingsfor sort of how we put different pieces of the puzzle togetherand just going to look at some automotive data for sake
WILLIAM PINK [continued]: of an example.And what we've then been able to learn about how brands workand how media impacts brand.And then we'll get into some more granular-level insightsat the consumer level.And how we're stitching data togetherto really complete the picture.So it's a lot to cover in a short amount of time.But hopefully you'll see how impactful this kind of work
WILLIAM PINK [continued]: can really be.And that will be to connect the data intelligence in action.So what is the framework we like to use?This visual you see here is sort of the waywe sort of high-level approach the research as a systemto answer very specific questions about how brands workand how brands achieve growth.
WILLIAM PINK [continued]: The phrase we like to put our headsaround is, we're really trying to drive behaviorstoday and cement demand for the brand for tomorrowor in the long-term.So what does that mean?That means we need to think about how both dataand analytics can be used together to kind of talkto the upper place of the chart here
WILLIAM PINK [continued]: to understand short-term movements in response--short-term movements in social in search, in shoppingbehavior, potentially sales.But we also need to understand how the marketing activitiesand the total brand experience are leading to different brandperceptions--which activities are working together to do that?Which activities are not working together?
WILLIAM PINK [continued]: And how does that all relate from a perceptions perspective?You just see a really important list here of perceptionsto kind of structure the brand for growthand lead to success over time.So this is sort of an analytic and research frameworkthat you're going to see in action through the deck.And I'm going to start with the upper half of the chartand talk about you leveraging searchand social as indicators of brand and campaign success.
WILLIAM PINK [continued]: So as a research firm, the questionabout which digital behaviors are most informativeon brand health and campaign performancehas been a very big question we've been wrestlingwith for the last few years.There's been a lot of debate about howto leverage certain social as a campaign or brand indicator.
WILLIAM PINK [continued]: There's a question about whether they are redundant with surveysif it means surveys aren't even needed.There's a question about whether they're representative and notuseful and which there's a whole raging debate about that.So what we wanted to do is we wanted to take a step backand say, OK, we're going to understandthe digital signals that matter from a branding campaignperspective.
WILLIAM PINK [continued]: We've got to do some real R&D. We have to test this.Being a global research firm, we havethe luxury of data, especially traditional research data.But we kind of start partnering withmany, many different sources of digital behaviors,whether that was web traffic, searchvolume, social sentiment.
WILLIAM PINK [continued]: I think, ultimately, the R&D lookedat over 100 different measures from a digital sideand wanted to see which one tied best to survey-based brandmeasures, which one tied best to sales offline,and what kind of analytics were neededto sort of pull it together.So this will give you a visual sampleof some of the categories and markets we looked at.
WILLIAM PINK [continued]: But that was really the question about thinking about itfrom a connected data perspective,trying to answer the questions about branding campaign,what data would be most useful, and whatdid we need to do that data to make it work.So let's start with a relatively simple finding with onethat has a lot of implications from a measurement perspective.
WILLIAM PINK [continued]: You're going to understand the relationships, say,between digital signals and sales.The nature of that relationship isgoing to vary a lot by category and by brand.So very simply if we start on the left-hand side,here, we just plotting on top of each other, search volumes,and weekly sales for a retail brand from the UK.
WILLIAM PINK [continued]: And as you could see, they're all almost righton top of each other.There's some bumps and bruises in the trend.But they pretty much go up and down together.Well, that brand happened to have a very stronge-commerce channel.So the fact that search and sale sortof fall on top of each other is really not surprising.Most of those searches' region, Idon't think most likely to relate directly to conversion.
WILLIAM PINK [continued]: You can eyeball the charts.You could see them on top of each how they kind of work.But in other brands and this on the other side of chart,we just happened to flag with some of the automotive testsfrom the US.If you want to understand the relationship between, say,search and sales for automotive, it's a much more complicatedprocess.You need mathematical transformations.
WILLIAM PINK [continued]: You need many other control variables.The purchase cycle is very different and very complicated.There's all sorts of intervening factorswhen you think of an automotive purchaseand what happens between potentially searchingfor a car, searching for a shopping momentjust to see what's going on, versus, actually buying a car.
WILLIAM PINK [continued]: So what we needed to do is build really the analytics,control for other factors, extract a long-term trend,and we could make it work.Here, it's a much more complicated process.And that was sort of the blinding insight and retrospectthat one of the reasons there's so many opinions about howthis works is it does work often a little bit differently
WILLIAM PINK [continued]: by situation.But we did see some common patterns.And one of the things we found wasif we were going to use digital set data for signalsof long-term trends, of long-term brand helpor signals of short-term reactions to campaigns,
WILLIAM PINK [continued]: the common patterns were the needto use analytics to understand the trend that weneeded to account for certain variables are extremelyconsistent.So across categories, you need to account for seasonality.Across categories, you'd always needto account for marketing activities.It just says campaigns spend years of summary.
WILLIAM PINK [continued]: Across categories, especially on social,you need to account for events--could be a PR moment, could be a celebrity moment.Whatever it may be, the data tends to be very spiky.And you have to account for those events.And then ultimately, you could get a long-term signalin terms of the base level-- whatwe call the base level here.
WILLIAM PINK [continued]: So if we did that, these analytics,and you see the phrase here, dynamic models--not to get into the weeds of the statistics.But one of the things we found wasby using this type of time series analysis,we could learn how brands, as they were performing on searchin terms of social mentioned.
WILLIAM PINK [continued]: We could understand how they evolved over time.And we could understand the impact of these advertising.We can understand the impact of eventson how those brands performed.And what did we found was that time series techniquewas consistently the most effective wayto extract the signal from the noise.So we built a system served to repeatedly apply that.
WILLIAM PINK [continued]: And one of the learnings we find is on the back endand across many, many brands in many, many marketsand categories, if we run this time series analysis,and we extract in this case a campaignor a short-term impact, once we controlfor those factors we discussed, once we sort of look
WILLIAM PINK [continued]: at how the campaigns or the messages rated--in this case, low, medium, high is just a low, medium, highin terms of breakthrough, awareness, and memorabilityof the messaging.Once we control for spend, interesting impacton the amount of social conversation
WILLIAM PINK [continued]: is highly related to effectiveness of the creator.So this gave us a lot of confidence in terms of,oh, very interesting.If we apply a proper modeling technique to extractthe elasticity, does social respond to spendand different forms of creative as one would expect--
WILLIAM PINK [continued]: better breakthrough, higher response?Yes, we see that over and over again.So this is the kind of learning that sort of R&D led to.We're going to get into many other examples.But I would definitely tell you to look to our sitewhere there's all sorts of stuff written at millwardbrown.com.But just to keep this thought in mind that on a repeated basis,
WILLIAM PINK [continued]: we would see a very consistent and predictable relationshipbetween controlling for all these factors,controlling for media.We would see the amount of responseon social tied to creative effectiveness.However, that doesn't mean it's as simple because you'regenerating a big response that's always
WILLIAM PINK [continued]: going to result in offline equity, offline sales.You still have to understand the nature of the campaign.And you have to understand the nature of the social contact.This is a nice example of what we're talking about here.In this example, we had three brands--kind of a medium, large, small--
WILLIAM PINK [continued]: whatever I'd think of it.And what was interesting was brandA would advertise and do all sorts of marketing activity.And they would just look at the social trends.And they would see a bigger response than for other brands.So they could live side by side and say, hey, look, here's
WILLIAM PINK [continued]: our marketing plan over the last year.Here's the trend in social.When we delivered more advertising,we had bigger spikes than when other brands deliveredadvertising.And they were wanting to declare victory from that.But the question was, does that relationshipresult in offline impact for the brand?
WILLIAM PINK [continued]: So we started simple.We said, all right, let's look at the content.Let's look at the social content itself.For brand day, which saw the biggest effects,what was interesting is the actual conversationonline highly overindexed compared to the other brandsto be about the communications itself.A lot of conversation was about the advertising.
WILLIAM PINK [continued]: For the other brands in contrast,the conversation was much more about the brand experienceusing the brand, the consumption.And when you then rolled out to seewhether it was having an offline effect on equityand an offline effect on behaviors,even though you got a bigger response for brand day,it had a weaker effect, because people were talking
WILLIAM PINK [continued]: about the ads themselves.It didn't translate into an equity experience.For the other brands, it was less elastic to advertisingor marketing.But it had a bigger effect offline when it did spike.So we used this example to say, although we have allour global R&D, that shows we canuse social as a measure of campaign breakthrough
WILLIAM PINK [continued]: as a measure of creative elasticity.And we know the kind of measures that socialtends to impact offline.You still want to take a look at all the campaigns--all the brands uniquely to understand what's going onand understand how it's translatingto an offline effect.And then it gets even more complicated.And this chart is not meant to be an eye chart
WILLIAM PINK [continued]: and look a little bit like a subway map.But it's that way to make a point.One of the most consistent questions-- in fact, bettersaid, one of the most consistent desiresis to use digital measures as a sort of leading indicators.If my social goes up, if my search goes up,will I see a sales effect a week or two down the road?
WILLIAM PINK [continued]: And often, you will see patterns that look like that.But our point of view is to really understandthe relationships you need to go a bit deeperand not think of it as much of a linear relationshipbut think of a network of relationships.So in this particular example, whatwe found was if you look, again, at different beverages
WILLIAM PINK [continued]: situation, if you look at sort of the drinking moments,sort of on-premise, which would be in a bar in a club,you would often see social respond after the moment,because what was happening-- peoplewere talking about the brand after they consumed it.So really, social lagged the sales experience.But it lagged it by a matter of moments or hours.
WILLIAM PINK [continued]: Oops, let me change the slide and go back.What we found, though, is as that social increased,that did get picked up in surveysa few weeks later, which also got picked upin longer-term sales effects.So there was this kind of sales to social, social to equity,equity to sales.
WILLIAM PINK [continued]: And then guess what happened?When sales went up in the long-term,you had more social mention.So we used this example to say, youdon't want to get caught up too much from a connected dataperspective to saying, oh, look at digital.And then you'll know it'll happen offline.It's much more about understand the network of relationshipsin your business.Understand how your brand works.And then in that network, it'll probably
WILLIAM PINK [continued]: be a series of leading and lagging indicators each relatedto the other.Here's another example of the kind of thingswe want to put together.You'll hear phrase about, what's the sort of structureof the brand?What's the hierarchy?what ladders up to choice?What ladders up to decisions?
WILLIAM PINK [continued]: And there's lots of work done in this space--different statistical analyses.But for us, the real learning comesfrom the databasing of results.So connected data is not only pullingdifferent pieces of the puzzle together,as you saw on the other slides.It's also about continuous learning and debt databaseand so that we can build metalearnings.So when after we've done a series of analysis--
WILLIAM PINK [continued]: this happens to be in an automotive example--we start to see patterns emerge.And what is the sort of hierarchy?It's never particularly clean.It's actually tends to be a bit messy in termsof how things ladder together.But if you get the look for, OK, whatare the things that tend to ladder to the top?Trust, reliability, productive, or good vehicle,
WILLIAM PINK [continued]: with tends of ladder up in automotives.Those are sort of your higher order benefits.But what tends to be the things that kind of drive that?And we see these relationships between performance,quality manufacturer, fuel efficiency, technology,dealership.These are the sort of things thatare kind of under the hood, if you will,
WILLIAM PINK [continued]: of the equity story that only come outby using analytics that tie data together and only come outby doing that on a repeated basisand databasing the findings so that wecould see across many of these studies what patterns emerge.And here's another story.What we just looked at was more of the positioningside, the equity side.But it's the same kind of thinking
WILLIAM PINK [continued]: that goes on the media side.So in this case, we use our cross media researchwhere we creating estimates of impact of the various marketingchannels and seeing how they contributeto outcomes, such as message association,such as purchase intent.And it's a combination of survey and nonsurvey data required
WILLIAM PINK [continued]: to do this to get at the exposuresand to get at the outcomes.And what you find is quite intuitive.Once you look in this case has to be in the transportationcategory--when you look across the cat across the brandswe've measured, what you see sort of a first principlefinding of different media tending to work differentlyto achieve different outcomes.
WILLIAM PINK [continued]: So the way TV works works quite differentlythan the way magazines work, which,in turn, works differently than the way cinema works.And even some of these terms might be data.And we'd like to talk more about video and videoacross platforms than TV.But just for a classification perspective,you can see how the more traditional mediaworking a particular way--
WILLIAM PINK [continued]: what their average reach levels tend to be--some of the other media that are working quite differently.Me personally, I find the magazine one is alwaysquite fascinating, because the reach for magazineshas changed.It's gone down.And magazines are consumed very differentlythan they have in the past.But for those smaller number of folksthat are exposed to the magazines,
WILLIAM PINK [continued]: you still see an overindex from an impact perspective--the impact on messaging, impact on intent.You could see a big effect from point of sale activities, whichis when you're in that mode and you'rein that moment, that has a big effect.And TV is still benefiting a lot from its reach,although it's obviously there whenyou see the share of spend.So it's quite expensive.
WILLIAM PINK [continued]: So while these are really useful findingsto help drive and set context for decisions, for purposestoday, we can only get to these kinds of findingsby bringing together survey, nonsurvey, and analyticsin a smart and intelligent manner.Now, what does it mean to put it all together?If you remember, we started this section of the presentation
WILLIAM PINK [continued]: with me saying, we have a marketing effectivenessframework that leverages different researchand analytic techniques.So this is a snapshot of an actual examplewhere we included three stages, if you will.We included an analysis of short-term sales.We included an analysis of long-term sales over, say,
WILLIAM PINK [continued]: a two to three-year period.And we included deep dives into media and perceptionsand kind of pulled it all together.Without getting into the details of the research mechanics,it was really important to have all three pieces of the puzzle.Otherwise, you could miss certain parts of the story.If you could see just by the size of the bar charts here,
WILLIAM PINK [continued]: more of the TV effect was coming through in the long-term.More of the promotions effect was coming throughin the short term.And you need the full puzzle to see the full impact picture.Different media come through differently again,as you're looking short term, as you're looking long term.And if you're looking to make allocation decisions
WILLIAM PINK [continued]: and only looking at it in one perspective or the other,you have the potential of making the wrong decision, whichis what we're trying to avoid.So that's where we're headed.This is the kind of output you can ultimately generate.But again, it requires integrating the analysisinto creating the research together.
WILLIAM PINK [continued]: So one other research comment--everything you just saw was metalearnings,connected data, a combination of survey, nonsurvey analysis.And in the last section there, westarted to move more towards consumer level analysis.That's really where things get much more powerful,
WILLIAM PINK [continued]: not that the aggregate analysis aren't powerful.They're extremely powerful.But in our more modern media landscape,we want the ability to slice and dice.We want as much consumer level or individual leveldata as possible.Yet, we know most data does not exist in one single source.If it did, life would be a little easier from a research
WILLIAM PINK [continued]: and analytics perspective.So what we need to continue to push onis using analytic techniques--sometimes the referred reference,data fusion techniques and the like--to take individual data streams that areseparate from each other, regardless of how they'recollected, whether a primary research or not,to bring them together to look at data sets in one
WILLIAM PINK [continued]: single source matter.In this example, we had data on TV behaviorsfrom things, like set-top box.We had online behaviors from digital panel.We had mobile behaviors from a mobile panel.We had survey and perceptions and equity
WILLIAM PINK [continued]: from survey research.And we had sales data from a receipt-based aggregationof sales beat decisions.So that's a big laundry list of various data setsthat are all important in their own right.But the goal was to leverage them and createone view of the customer and how to do that.
WILLIAM PINK [continued]: So to do that, we had to frame it in our minds.We had to use some analytics to impute,given what I know about the people over herein the TV data set.And what I know about them over here in the surveydata set and their answers to certain questions.What would be my best estimate of their TV behaviorsbringing it together based on people that look alike?
WILLIAM PINK [continued]: What did I know about their sales data over here?What do I know about their choicesover here and their demographics?Can I make an estimate of how they link together?So basically, we create the single source.It's an imperfect data set, but it'squite strong in terms of the quality of the imputations.And then we can learn things we couldn't otherwise see.
WILLIAM PINK [continued]: So what did we ultimately learn?Here is just an example.You get to see some of those maps,like we were showing you earlier, but in a verygranular level.And you get to see all the various interrelationships.So things, like, exposure to TV advertisinggenerates search volume.And in this case, it also generated in-store purchasing.
WILLIAM PINK [continued]: Great.The online exposure's led to more online shoppingand online purchasing, as well as in-store.So you say, oh, what happened to TV and in-store?Well, it's interesting.What happened to TV and online?Well, what we found was the strongest relationship for TVwent TV to search, search to shopping to online.
WILLIAM PINK [continued]: To get to understand how the journey is happening,we also got to see a lot of relationshipsfrom purchase decision leading to social commentaryimpacting brand imagery in the survey.So you really get to start to get a view into the consumerjourney, which is what is much moreactionable from a marketing, perspectiveand why we refer to this again as connected data intelligence.
WILLIAM PINK [continued]: And then one kind of wrapper slide--you can really start to pull this all together.And you can see how the data flows from insightsthrough activation to measurement and a systemonce you're at the consumer level of insights.And in this case, we included some names,because you could see it needs to be researchers and agencies
WILLIAM PINK [continued]: and different research groups working togetherto make this work.It's not just analytics.It's not just smart design.It's also that people element--taking it across stages, thinkingabout the different stages of understanding messaging,understanding segments and audiences,scaling them in different systems through modeling, testand learning against delivery from different agencies,
WILLIAM PINK [continued]: and ultimately measuring again.That's a pretty hefty system of research and media targeting.But it's only made possible by the current data possibilities,the current data applications, and working together,ultimately, again, towards connected data intelligence.So that was a really fast run through
WILLIAM PINK [continued]: some pretty high-end stuff.And I think I'm actually handing it back to Marioto bring us to the conclusion.
MARIO SIMON: Absolutely.And again, our aim here was to really gofrom a very sort of organizational capabilities-ledmindset for data analytics all the way to actual executionand an actual models that are employed today
MARIO SIMON [continued]: that really sort of changed the gameas integrated models that's take into account a lot of data.So happy to open it up for any questions,which if we can answer them on the fly, we will.And if there are technical questions
MARIO SIMON [continued]: that we need to think about it a bit more,then we're happy to also have a dialogue post this call aswell.
SELMA AVDIC: Yes.Thank you so much to Mario and to Bill.Yeah, we have some great questions.So I'll just go ahead and get started.And then the first question for you both.We have an audience member that wants to know,what was used to gain the qualitative natureof social content?
WILLIAM PINK: In the example we showed or in general?
MARIO SIMON: I think of example that was shown here.
WILLIAM PINK: Oh.Good question.[LAUGHS] I forget which system it was.I believe it was an analysis of usingCrimson Hexagon and other sort of social listeningtools in various social research queries to extract the content.But it's important to remember, that analysiswas a combination of things.
WILLIAM PINK [continued]: It was a content analysis of the terms--a very cleaned up content analysis.As you would imagine, that happensto be when you're talking beverages,that is a category where quite often, there's a lot of wordsyou need to clean up, whether that be something,like, you just used the example, Mountain Dew.
WILLIAM PINK [continued]: The word, DBW, you could be talking about morning dew.You could be talking about skiing in the mountains.You could talk about all sorts of things.So besides the cleaning of spam, the cleaning of languageoccurs first, as well as the spam,reduced to a meaningful set of commentsthat are about the brand.And then we're using various formsof qualitative and machine-driven
WILLIAM PINK [continued]: coding to pull the content analysis together.So that's piece.But then it was the dynamic modelingto understand how the media impacted social.And how social translated into equity and sales.There's really two separate analyses pulled together.
SELMA AVDIC: Wonderful.Thank you so much, Bill, for that clarification.We have another participant that wanted to knowthe difference between--they want to know what is online display versus online video?
WILLIAM PINK: [LAUGHS] That is a good one for me to confirm.Our database team did that coding.Oh, and I should say, these were the global average.These are not only US campaigns in hereI have a feeling it was different formsof static display versus video display.But I will confirm that just to make sure I'm
WILLIAM PINK [continued]: saying the right thing.But that was from our database team coding upthe various online advertising in the set of campaignswe have for automotive.
SELMA AVDIC: OK, great.Thank you, again.So next question here.We have Tony that wants to know about the scalabilityand economics of this approach dependingon the size of the business.And this is in relation to consulting as you've described.
WILLIAM PINK: Great question.
MARIO SIMON: And I think that goesto really a very right turning to sort of mentionthe question around scalability and with consulting.There are a few techniques in the marketplace that
MARIO SIMON [continued]: are less precise and less insightful but actuallymuch more automated.I think that our capabilities go from a quite contained costand the product eye view of this kind of measurement all the way
MARIO SIMON [continued]: when it is really a big investmentthat clients want to make.So I'll give an example.A beverages company that uses a lot of experience for marketingand wanted to sort of go into a new spaceof media purchases, which was that multi, multimillion-dollar
MARIO SIMON [continued]: investment.And therefore, they needed to have really robust analyticsbehind it.So there was a big consulting engagementin a couple of millions of dollarswhen it was all said and done.But it really supported optimization of investmentof $100 of millions of dollars.So that would be the case for big shift
MARIO SIMON [continued]: that marketers want to do.But then on a day-to-day on the fly,there's a lot of optimization thathappens and is actually quite automated and quite productive.So it's a big gamut.And if we understand the scope and the final consumerand the objectives of the data and of the events
MARIO SIMON [continued]: that we can definitely offer different approaches.
WILLIAM PINK: Dealing Mario--yeah, I do.And it's great, great lead in.What I wanted to add is more on the research side.One thing you'll notice from a Millward Brown perspectivewhen you go on our site, the tools and approaches wedescribed here-- they're meant to be highly scalable, highlystandardized.And you'll see we have a really a nice range of some
WILLIAM PINK [continued]: of these services, like this leveraging searchand social as measures of branding campaign.We have that at a relatively reasonable price, highlyscaled.And we've made it as standardized and automatedas possible building in the analytics.And then all of our higher-end servicesreally build off of that.It's the same first principles.
WILLIAM PINK [continued]: But you just get more and more and more.You connect more and more dots.So for us, it's actually been good business strategyin meeting the market for questions like yours,because we know not everyone will be a full-blown allthe research bells and whistles, all the consulting bellsand whistles.But we needed something that's highlyscalable and leverageable by many different typesof businesses, small through large.
WILLIAM PINK [continued]: So we've built these things so that youcan have a service that more of a do-it-yourself, moreof a simpler scaled downs version.And then you've got the full-blownwhat we call brand guidance systems that bring everythingtogether.So it's really meant to serve kindof all of end of the market.
SELMA AVDIC: OK, wonderful.Well, I want to say thank you to both of youagain, Mario and Bill.That is all the time that we have for todaybut thank you so much everyone for their commentsand these great questions.Just as a reminder to all of our AMA members--
SELMA AVDIC [continued]: you will be able to view an archive of this webcast by anytime by visiting AMA.org/webcast.And we will have all of this great insightavailable for you.I just want to thank also Kantar Vermeer again,and for presenting this with us, as well
SELMA AVDIC [continued]: as ReadyTalk who provided us with our platform for today.If you'd like to learn more about ReadyTalk,you can visit readytalk.com/ama.But ultimately, we really want to thank our audience,our wonderful members.Thank you so much for your participation and your time.And this does end our presentation for today.
SELMA AVDIC [continued]: Thank you again and enjoy the rest of your day.
Publisher: American Marketing Association
Publication Year: 2016
Keywords: analytical methods; B2C marketing; brand management; business growth; business models; business-to-business marketing; consumer behavior; customer centrality; data analysis; data collection; data integration; data management; data mining; database marketing; digital communication; experimentation; globalization (business); impact analysis; innovation and creativity; internet; interviews; large-scale research; marketing campaigns; marketing strategy; network analysis; network data structures; network indicators; research findings; risk (business); Sales; Social media; Survey research; time-series analysis ... Show More
Segment Num.: 1
Mario Simon, CEO of Millward Brown Vermeer, and Bill Pink, Managing Partner of Millward Brown Analytics, discuss their large-scale research on the relationships between marketing, digital communication, and sales growth.
Looks like you do not have access to this content.
Mario Simon, CEO of Millward Brown Vermeer, and Bill Pink, Managing Partner of Millward Brown Analytics, discuss their large-scale research on the relationships between marketing, digital communication, and sales growth.