66
4 December 2013 Leadership Series How to Effectively Use Market Research to make your Branded Content more Effective

Leadership Series How to Effectively Use Market Research to … · 2016-05-03 · How to Effectively Use Market Research to make your Branded Content more Effective. The graveyard

  • Upload
    others

  • View
    2

  • Download
    0

Embed Size (px)

Citation preview

4 December 2013

Leadership SeriesHow to Effectively Use Market Research tomake your Branded Content more Effective

The graveyard of failed BEprograms is littered with greatideas, even great executions.

“”Ogilvy Entertainment: Making magic, using logic. BEAM

The more I practice, the luckier I get.“ ”Arnold Palmer

Estimating theROI

(& what BC is doing vs.other campaign

elements)

Measuring theresponse tothe branded

content

Evaluating theimpact on the

brand

Conscious response is ok, but limited.- Consumers tend to post-rationalise- They can’t articulate how they’re feeling

Measuring theresponse tothe branded

content

Measuring the unconscious response

Brain scansExpensiveDifficult tointerpret

Facial CodingSinglemeasurement ofengagementMeasures thenature of theresponseSimple toimplement andinterpret

BiometricsSingle measure ofengagementEasier to interpretNo measure of thenature of theresponseNot cheap

Be Emotive

2005 Luca Bruno, 2013 Michael Sohn - AP

Creative Media ROI+ =

Challenge

? ✓ !

Source: Gunn Report 2000-2008, IPA

Les Binet & Peter Field,Marketing in the Era of

Accountability, 2007

Emotional advertisingcampaigns are moreeffective and more

profitable than rationalcampaigns - even in

'rational' categories…

11xmore efficient

in market sharegrowth

The Gunn Report

Emotions?

1800s

DarwinThe Expressionof Emotions in

Man andAnimals

1960s

!

EkmanFacial Action

Coding System(FACS)

2010s1800s 1960s

CodingAutomated

FacialCoding

Emotions

Demo

Cannes Study

11kpeople

75%prediction

accuracy forshortlists

Entered 96

Shortlisted 72

Bronze 24

Gold 30

Silver 24

246videos tested

35k+viewings

Grow the trend

0.14

0.13

0.12

0.11

0.10

0.09

0.08

0.0 0.2 0.4 0.6 0.8 1.0

Bronze/SilverGold

Shortlisted

Nothing

Relative time

Happy aggregated trendlines

Steeper slope=

Better outcome

Value

What kind ofemotionsdoes your

video spark?

How differentdemographics

react to thecontent?

How does itcompare to closecompetitors and

industry?

15” to 5’content length

20+categories

EmotionalImpact

1,000+videos tested30+

countries

EngagementConnect via

emotions

AttractionHook in 8seconds

RetentionKeep theaudience

ImpactKahneman’s

Peak-End

Successful video components

Case study

Confusion isdominant

emotion for first45 secs, lowering

attraction

Happinessspikes anddominatesthroughout

Happiness drops,but video ends

with peak, leavingviewers withstrong impact

Minimalnegativeemotions

throughout

This video has very goodoverall performance.

Better than 95% of videos.

So Real it’s ScaryEmotional

Impact

Male79.5%

Female79.4%

EmotionalImpact

Age

Income

CountriesCompetitor

In-depth analysis

Conclusions

Confusion in the first45 seconds lowersattraction. Happinessspikes and dominatesthroughout.

Top5%

Compared withvideo A, LG leavesviewers with ahigher impact.

A B

• 18-34• Single, divorced,

separated.• Low or very low

income.

~

Get attention with a strong spike ofsurprise and happiness in the first eightseconds. Avoid confusion.

Prevent drops and endwith a spike for thestrongest impact.

Ending with an engagementspike benefits the syndicationstrategy of publishers withhigher play-to-end rates.

Syndicate to relevantaudiences with thehighest engagement.

Learning

€Activate the videoacross UK and USaudiences

http://beon.aolnetworks.com/en/advertisers/case-studies/technology

Emotional

CampaignData SetCampaignData Set

EmotionalEngagement

Audience Brand LiftResearch

CampaignData

CampaignData Set

Revolutionising measurement

The most advanced campaignreporting tool, utilising industrystandard brand lift measurement,audience profiling, engagementand emotive analysis to provideROI visibility.

Insights

Consumers love emotional content...

3xhigher averageview to end %

8xhigher click

through rate

20xbetter in

converting viewsinto social

actions

100xQuicker inattractingviewers

Source: Be On research

8ppincrease in BrandRecommendation

7ppincrease in

Purchase Intent

14ppincrease in Brand

Favourability

...and it also creates ROI

CreativeAttractionRetention

EngagementImpact

MediaSyndication

OptimisationSharing

ROIBrand Favourability

Purchase IntentBrand Awareness

+ =

Create Syndicate Measure

Impact Model

Estimating theROI

(& what BC is doing vs.other campaign

elements)

Measuring theresponse tothe branded

content

Evaluating theimpact on the

brand

11.5m views on YouTube

#1 tweeted about video inFrance (a tweet a second)

2000 facebook fans in 2 days

Non-survey based approach

Perrier’s “Le Club”

100,000 people interacted with Persil on the forum

25,000 people aged 18-35 bought tickets to the event

9% uplift in brand awareness among target audience*

8% increase in brand usage*

*Source: TNS Russia

Quasi-survey based approach

Survey based approach

Exposedto branded

contentN=200

NotExposed

N=200

Audience

NotExposed

Awareness Perceptions Purchase intent

Advocacy Affinity Associations(more metricsthe better)

Audience

Metrics

Exposedto branded

contentN=200

Survey based approach

NotExposed

Awareness Perceptions Purchase intent

Advocacy Affinity Associations

45%

22%

(more metricsthe better)

Audience

Metrics

The content has led to a 23%points increase in brand affinityAnalysis

Exposedto branded

contentN=200

Survey based approach

Aired on UKTV at a time when paid-for references for productsand services were NOT permitted in TV shows in the UK

• 900 adults aged 18-44 in the UK were interviewed

• Five groups of respondents1. Not exposed2. Exposed to promo only3. Exposed to promo and show4. Exposed to promo and website5. Exposed to promo, website and show

Impact #1: Unprompted awareness of Prius increased

24%

37%

Not Exposedto Show

Exposed toShow

% mentioning ‘Toyota Prius’ or ‘Prius’ (unprompted)

+13% points increaseas a result of the show

Exposed

NotExposed

Impact #2: More people would consider buying a Toyota

32%

43%

Not Exposedto Show

Exposed toShow

% considering Toyota for next car purchase

+11% points increaseas a result of the show

Exposed

NotExposed

Impact #3: The show strengthened perceptions of Toyota

+12% Leads the way in tackling environmental issues

+10% Safe

+9% Suitable for families

+8% Practical

+7% Honest

+7% Constantly innovating

+5% Warm / friendly

We can also analyse what each campaign element delivered…

13%

44% 39%62%

Promo Only Promo & Show Promo & Website Promo, Website &Show

% correctly identifying Toyota/Prius as the sponsor of Carpool once the nameof the show has been revealed

The showadded+31%points vs.

promos alone

Estimating theROI

(& what BC is doing vs.other campaign

elements)

Measuring theresponse tothe branded

content

Evaluating theimpact on the

brand

Pointlogic PPT Template v 1.0

Pointlogic Valuepoint - Respondent level modelling

ROI of branded content in a complex world

What does Return on Investment (ROI) mean?

ROI All the things a brand does

- Ads- Sponsorships- Branded content- Product experience

Investment types

What does Return on Investment (ROI) mean?

ROI

All the things that happen- Sales- Brand health

The challenge

What we would like to know?

Did the tacticalcampaigns alsomove brand health?

Was my TV spendtoo high or too low ?

Which KPIs weremoved the most?

54

Our approach in 4 steps

1

2

3

Measure their brand attitudes and behaviours

Understand what marketing consumers have been exposed to

Determine the links between marketing and behaviour

4 Feedback into future planning

Understanding Communication ROITraditional methods

Aggregate modelling - Look at each week and work with total spend, total TVexposures etc.

Struggles badly to isolate the effects of media in complex environments

We start with individuals…

WomanHas childrenAffluent

.

Light TV AdsHeavy facebook adsVisited brand pageMedium radio ads

.

TrialistTrusts brandAware of ad messageIntends to buy us

.

What havethey seen

?

What didthey do orbelieve ?

What elsemakes

them dothat?

As the numbers grow, we can isolate more effects…

WomanHas childrenAffluent

.

Light TV AdsHeavy facebook adsVisited brand pageMedium radio ads

.

TrialistTrusts brandAware of ad messageIntends to buy us

.

What havethey seen

?

What didthey do orbelieve ?

What elsemakes

them dothat?

ManHas childrenMiddle income

.

Heavy TV AdsNo facebook adsHeavy YoutubeLow press ads

.

AwareTrusts brandNot aware of admessageIntends to buy other

.

WomanSingleLow income

.

WomanHas childrenMiddle income

.

ManHas childrenAffluent

.

Medium TV AdsHeavy facebook adsVisited brand websiteHeavy radio ads

.

Light TV AdsLow facebook adsLow YouTubeMedium press ads

.

Heavy TV AdsHeavy facebook adsNo press adsNo radio ads

.

LoyalistTrusts brandAware of ad messageIntends to buy us

.

Not awareDoes not trust brandNot aware of admessageIntends to buy other

.

Re-purchaseTrusts brandAware of ad messageIntends to buy us

.

Once we get up to bigger numbers, we can start to unpick ROI

Once we get up to bigger numbers, we can start to unpick ROI

Saw TV ad 4 timesSaw TV ad 4 times

Once we get up to bigger numbers, we can start to unpick ROI

Visited brand websiteVisited brand website

Once we get up to bigger numbers, we can start to unpick ROI

Intend to purchaseIntend to purchase

Applying the power of modelling

• We then just need modelling totease out all the causes and effectsfrom the hundreds of individualswe have precise data for

• We can sort out the media effectsbut also understand what is downto their personal history andpreferences

• We use new and powerfultechniques to do this atrespondent level

Example - modeling results

Product AAwareness

Product APurchase

intent

Product BAwareness

Product BPurchase

intent

TelevisionCampaign 1

TelevisionBranded content

Print

OnlineCampaign 1

OnlineCampaign 2

Significant, highly cost efficient

Which media have a significant effect and are cost efficient to increase the KPI?

Significant, less cost efficient

The media agency and client supply details of all the media activity at adetailed level

In a survey (existing tracker or custom) we measure all behaviours, brandhealth and media usage

Branded content can be specifically probed for

From the media usage we can estimate how many exposures to all thecampaign activities

Then we do the modelling

Then we load the results into our planning tools

Practical implementation

65

Software to use insights for scenario planning…

Thank you.

Leadership SeriesHow to Effectively Use Market Research tomake your Branded Content more Effective