TVNZ AdEffect Series

TV drives mass reach and boosts the effectiveness of other media

Seven case study series summary What we did This case consumer campaigns study series 6,647 interviews represents the 7 largest and most Local and categories comprehensive 6 global brands research study into campaign All major media channels effectiveness conducted in TV Online Video Online Display Social Media Radio New Zealand OOH Cinema eDM PR Print

TVNZ | AdEffect Series Summary Page 2 Delivering impact through effective marketing is a combination of both great ideas and an effective delivery mechanism

X = Creative Media mix Results

i.e. 100 X 0 = 0 0 X 100 = 0 100 X 100 = 10,000

TVNZ | AdEffect Series Summary Page 3 The TVNZ AdEffect Series understands these two essential components to provide holistic insights on how to deliver impact through effective advertising

How effective is the campaign overall and what actions could be taken to improve its effectiveness in the future?

How effective is the media mix? How effective is the creative execution?

Reach and Media mix Impact Standing out Change brand Change frequency perceptions behaviour How does each channel What campaign Does the campaign perform? How well do outcomes result from stand out from other Does the media plan How does the ad What is the effect on different channels work the media allocation? campaigns? generate enough impact the way the short- and long- together? exposure? consumers think about term KPIs of the brand? the brand?

TVNZ | AdEffect Series Summary Page 4 Depositing memories: how much does it cost to deliver your message? There are two components to this question

Cost of X Average = Cost of conversation / number of delivering impression impressions a message required

TVNZ | AdEffect Series Summary Page 6 CPM differs significantly by channel, but TV’s economies of scale mean TV can deliver mass reach economically CPM by channel (Local currency, WARC 2016)1

217. 2 55.4 66.5 21.2 8.4 10.4 14.0

OOH Broadcast Radio Magazines Newspapers Online Cinema TV

1. Source: WARC media spend database 2016

TVNZ | AdEffect Series Summary Page 7 Within online, CPM also varies significantly as does the advertising product being sold, i.e. display vs. video and skippable vs. not

CPM by channel (Local currency, WARC 2016)1 CPM by channel (Local currency, TVNZ AdEffect Series 2017)2

Different data sources lead to different results at total level 10.4 66.5 14.4 31.9 73.6

Broadcast Online Online Display YouTube OnDemand TV

1. Source: WARC media spend database 2016 2. Source: TVNZ AdEffect case study series average of 4 case studies with data supplied to a sufficiently granular level

TVNZ | AdEffect Series Summary Page 8 New Zealand is different to other western markets. A relatively low level of TV media fragmentation helps TV deliver strong economies of scale and low CPM compared to the market average CPM by channel by market (USD, WARC 2016)1

OOH Broadcast TV Radio Magazines Newpapers Online Cinema Market average 7.0 New Zealand 38.0

27.3 Australia 41.9

8.6 UK* 12.5

43.6 USA 34.6

NOTE See appendix page 39 for details on WARC calculation methodology 1. Source: WARC media spend database 2016, *UK 2015 data as 2016 is unavailable

TVNZ | AdEffect Series Summary Page 9 On average it takes 3.3 exposures to get a viewer to remember your ad, with the exact number required varying by channel and category

Average number of exposures to deliver message 4.2 3.9 2.3 2.8 (Range 1.0 – 7.9) (Range 1.9 – 5.9) (Range 1.0 – 4.0) (Range 1.4 – 3.5)

Broadcast TV Radio Print Online

NOTES: Source: TVNZ AdEffect case study series; measured as number of impressions required to deliver 20% of awareness; not available for Cinema or OOH; Newspapers and Magazines combined into print.

TVNZ | AdEffect Series Summary Page 10 Overall average On average it takes 3.3 exposures to get a viewer to remember your ad, with the exact number required varying by channel and category 3.3 Average number of exposures to deliver message 4.2 3.9 2.3 2.8 (Range 1.0 – 7.9) (Range 1.9 – 5.9) (Range 1.0 – 4.0) (Range 1.4 – 3.5)

Broadcast TV Radio Print Online

NOTES: Source: TVNZ AdEffect case study series; measured as number of impressions required to deliver 20% of awareness; not available for Cinema or OOH; Newspapers and Magazines combined into print.

TVNZ | AdEffect Series Summary Page 11 A direct comparison between TVNZ OnDemand and YouTube shows that TVNZ OnDemand is significantly more efficient than YouTube

Average number of exposures to deliver message 2.0 15.0 VS

TVNZ OnDemand YouTube

NOTES: Source: TVNZ AdEffect Series – Vod Case study; measured as number of impressions required to deliver 15% of awareness, extrapolated to 20% to be comparable with slide 39.

TVNZ | AdEffect Series Summary Page 12 Overall average Putting all this together, on average it costs 9.3 cents to deposit your advertising into a 92.7 customer’s mind... Cost of delivering message by channel (CPM multiplied by required frequency, local currency, WARC 2016)1

183.5 88.8 43.6 54.9

Broadcast TV Radio Print Online

NOTES: Source: WARC media spend database 2016, TVNZ AdEffect Series, Kantar TNS analysis.

TVNZ | AdEffect Series Summary Page 13 Overall average ...with Broadcast TV proving to be the most cost effective 92.7 Cost of delivering message by channel (CPM multiplied by required frequency, local currency, WARC 2016)1

183.5 TV proves to 88.8 be the most 43.6 54.9 cost effective

Broadcast TV Radio Print Online

NOTES: Source: WARC media spend database 2016, TVNZ AdEffect Series, Kantar TNS analysis.

TVNZ | AdEffect Series Summary Page 14 Building a mass reach campaign without TV: is it possible? TV was the top performing channel in all seven case studies; in fact no other channel evaluated ever achieved recognition higher than Broadcast TV’s lowest performance of 34% Advertising recognition by channel (% NZers 18-54)1

Min Average Maximum 55%

43% Broadcast TV minimum 34% 33% 27% 27% 19% 20% 20% 16% 16% 11% 9% 11% 10% 7% 4% 3% 1% 1% 1%

Broadcast TV Radio Online Display Online Video OOH Print Cinema

NOTES: Source: TVNZ AdEffect Series.

TVNZ | AdEffect Series Summary Page 16 If TV had not been included as a channel, recognition of the campaigns would have been 24% lower on average Average overall campaign recognition (% NZers 18-54)

59% -24% 35%

All channels Net impact on total All channels campaign recognition EXCLUDING Broadcast TV

TVNZ | AdEffect Series Summary Page 17 By modelling additional media spend, we can see the potential of other channels: all channels are limited in generating additional awareness and don’t come close to the level of reach that TV offers Potential recall – Maximum (% NZers 18-54)1

68%

Potential 42% 44%

33%

Average

43% 27% 22% 10%

Broadcast TV Radio Online Print

NOTES: 1. Source: TVNZ AdEffect case study series

TVNZ | AdEffect Series Summary Page 18 TV also helps make other broad reach channels more memorable Advertising recognition by channel by exposure to Broadcast TV1

High exposure to Broadcast TV 30% 18%

37% 28% Increase Increase Low exposure to Broadcast TV 22% 14%

Radio Online Display

NOTES: 1. Source: TVNZ AdEffect case study series.

TVNZ | AdEffect Series Summary Page 19 This is also true for radio, however TV’s superior reach means other channels are more likely to be leveraging off the power of TV Broadcast TV advertising recognition by exposure to other channels1

High exposure High exposure to Radio 49% to Online2 43%

37% 4% Increase Increase Low exposure Low exposure 2 to Radio 36% to Online 41%

Broadcast TV Broadcast TV

NOTES: 1. Source: TVNZ AdEffect case study series. 2. Exposure information only available to online as an overall channel.

TVNZ | AdEffect Series Summary Page 20 Effecting change: which channels influence consumers’ actions? While digital may have disrupted the traditional linear purchase funnel, as marketers we are still ultimately trying to move consumers closer to a point of purchase or action

Awareness Purchase Intent

Active evaluation Awareness

Consideration

Consideration set Loyalty loop Moment of purchase Purchase Intent Trigger Purchase Post-purchase

Consideration Purchase/ Loyalty Static & linear Looped Dynamic

NOTES: Kantar Asia Pacific, Getting Digital Right – Reflections for a better digital world.

TVNZ | AdEffect Series Summary Page 22 The following slides use indices to show the impact the campaigns we’ve studied have had on the advertiser’s KPIs

Impact of exposure on general brand KPIs (index)

High TV exposure increases High TV exposure increases KPI performance from 50% to 60% KPI performance by 20%

60% 50% To facilitate interpretation, the impact 120 of high channel exposure is converted 100 in to an index using the below equation

High Exposure Index = Control Control High TV Control High TV exposure exposure

TV TV

TVNZ | AdEffect Series Summary Page 23 TV performs well across all stages of the funnel, but is most powerful in establishing the campaign and driving core brand metrics

Generic classification of Average impact of campaign on KPIs (index, 18-54)(1)(2)(3) campaign KPI targets Average Broadcast TV

Specific actions e.g. Visit store, search online 137 120

Targeted brand perceptions e.g. Product is available, brand is easy to deal with 134 134

General brand measures e.g. Awareness, consideration, preference 111 120

NOTES: 1. Source: TVNZ AdEffect case study series 2. Control group (low exposure) = 100 3. FMCG targeted sales case study excluded as an outlier

TVNZ | AdEffect Series Summary Page 24 In both short and long purchase lifecycle categories campaigns as a whole drove similar results, however TV’s role in the media mix differed significantly

Generic classification of Average impact of campaign on KPIs (index, 18-54)(1)(2)(3) campaign KPI targets Average Broadcast TV

Specific actions 138 132 134 e.g. Visit store, search online 103

Targeted brand perceptions 143 135 126 132 e.g. Product is available, brand is easy to deal with

General brand measures 125 e.g. Awareness, consideration, preference 112 111 116

NOTES: 1. Source: TVNZ AdEffect case study series Short purchase Long purchase 2. Control group (low exposure) = 100 3. FMCG targeted sales case study excluded as an outlier lifecycle categories lifecycle categories

TVNZ | AdEffect Series Summary Page 25 In particular, in short purchase lifecycle categories, no other channel matched TV’s consistent ability to drive increased share/usage

Campaign impact on ‘usage’ KPIs (index, short purchase lifecycle case studies, 18-54)(1)(2)(3)

OOH and radio had no impact OOH and radio had no impact Online, print and OOH had no impact

121 121 117 114 105 114 105

Broadcast TV Online Cinema Broadcast TV Online Video Online Display Broadcast TV

QSR FMCG new product launch Retail (Share of purchase occasions) (Usage of product) (Share of spend)

NOTES: 1. Source: TVNZ AdEffect case study series 2. Control group (low exposure) = 100 3. FMCG targeted sales case study excluded as an outlier

TVNZ | AdEffect Series Summary Page 26 Brand vs. tactical advertising: is it possible for a campaign to do both? Peter Field and the IPA have discussed at length the importance of avoiding short-termism

“Because the effects of brand building only become apparent over the long term, short-termism is

dangerous. It can lead to excessive activation (which is inefficient), and under-investment in the“ brand (which can lead to long-term decline).

NOTES: Source: Media in Focus: Marketing effectiveness in the digital era

TVNZ | AdEffect Series Summary Page 28 As accepted as this view may be, executives and marketers are still tasked with short-term planning cycles. The challenge: is it possible to achieve both short- and long-term impact in one campaign?

87% 65%

of executives and directors feel most of executives and directors say pressured to demonstrate strong financial short-term pressure has increased performance within 2 years or less1 over the past 5 years

NOTES: 1: Source: Media in Focus: Marketing effectiveness in the digital era

TVNZ | AdEffect Series Summary Page 29 It is possible to achieve short-term results with a long- term campaign through effective planning and campaign integration with brand response campaigns

Building brands and driving sales are no longer mutually Brand response campaigns, exclusive activities. Critically, the two elements – if based around a brand integrated, and messaged idea that drives short-term Brand consistently – create a Response powerful virtuous circle behavioural activation, achieve where brand helps build response, and the response the best of both worlds1 itself helps build the brand through experience2

NOTES: 1. Source: The Long and the Short of It – Balancing short and long-term marketing strategies 2. Source: Marketing Society: Library – The inexorable rise of brand response

TVNZ | AdEffect Series Summary Page 30 Focussing on brand building (long-term effects) generally also leads to a short-term impact. The reverse, however, isn’t true Short and long-term potential of campaigns(1)(2)(3) Long-term potential (AMP)

Yes No

Yes 32% 7% Short-term potential (Motivation) <1% If the campaign achieves the long-term criteria, 94% of the time it will also achieve the short- term criteria.

No However, if the campaign achieves the short-term criteria it will only achieve NOTES: the long-term criteria 1. Source: TVNZ AdEffect case study series 2. 61% of the time neither criteria achieved 18% of the time. 3. Details of metrics provided on page 40

TVNZ | AdEffect Series Summary Page 31 Dressing appropriately for the channel: does tailored content deliver greater impact? The average campaign in the TVNZ AdEffect case study series made use of 4.4 channels

4.4 Channels used on average

Broadcast TV Online Video Online Display OOH Radio Print Cinema 7/7 case studies 7/7 case studies 6/7 case studies 4/7 case studies 3/7 case studies 2/7 case studies 2/7 case studies

NOTES: Source: TVNZ AdEffect Series

TVNZ | AdEffect Series Summary Page 33 Exposure to multiple channels significantly extends the campaign’s total impact Impact of single verse multiple channel exposures (index, 18-54)(1)(2)

Global research estimates that a quarter of the impact campaigns have is driven by no specific channel, but rather the magic that occurs when a message is synergistically delivered by 114 140 multiple channels3

Average impact on KPI of high Average impact on KPI of high exposure to one channel exposure to multiple channels

NOTES: 1. Source: TVNZ AdEffect case study series. 2. Control group (low exposure) = 100. 3. Source: Media in Focus: Marketing effectiveness in the digital era

TVNZ | AdEffect Series Summary Page 34 The extent to which multiple channels work together depends on the extent to which the creative is designed to match the channel’s strengths and form a cohesive campaign Impact of single verse multiple channel exposures (index, 18-54)(1)(2)

Increase over Combined brand Consistent Messaging Multiple channels One channel and above single and response campaign matched to channel exposure messaging assets channel strengths

172 FMCG 116 +56 144 Automotive 108 +35

FMCG new 156 product launch 126 +31

Consumer 128 electronics 110 +19

125 Retail 112 +13

116 QSR 105 +12

128 Finance 117 +11

NOTES: 1. Source: TVNZ AdEffect case study series. 2. Control group (low exposure) = 100.

TVNZ | AdEffect Series Summary Page 35 While there are clear-cut differences between many of the traditional channels, the way consumers consume content across different online video channels means online video cannot be treated as one channel

TVNZ OnDemand YouTube Facebook Audience1 (Watch alone) 52% 67% 84% Reaction1 9% 16% 22% (Stop watching when to many ads appear) (56% skip ads as quickly as possible all the time)

2 Greater use of Greater use of Size of device larger devices mobile devices

2 Leisure viewing. Task orientated, watching content Information scanning, scrolling Mind state Relaxed state watching for a specific reason i.e. education Or, through content looking for (Stop watching when too many ads appear) user selected content relaxing watching user selected content things that are relevant

Additionally, 48% of YouTube users will listen to talks/ music and not actively watch the screen, these users report this make up 39% of their time on YouTube.

NOTES: 1. Source: TVNZ VOD Case study. 2. Qualitative observation from Kantar TNS

TVNZ | AdEffect Series Summary Page 36 Therefore different creative executions are suited to, or need to be created for, different channels

6 15 30 60 Long Sec Sec Sec Sec form

Basic ideas Complex Resonance As an event As a reward & reminding messages Suited to the Youtube or social environment, where Best suited to social Suited to the TV OnDemand platform, where the viewing behaviour it is very important to channels where the and mind state most closely resembles Broadcast TV engage the viewer quickly viewer can self-select and minimise disruption

> Potential high reach > To stretch budget > New product > Developing stories > Assume low reach/social and resonance > Cut downs > New campaign or after > For high involvement potential or niche/transient > Surprise and delight > Simple communication needs long hiatus > Buzz relevance > Salience > Complex or multiple > Inspiring stories > Reinforcement messages > For fans who can’t > Limited messaging > New variant get enough > Immersion

TVNZ | AdEffect Series Summary Page 37 Appendix

TVNZ | AdEffect Series Summary Page 38 WARC: Sources and CPM calculation methodology

Medium Ad size Source CPM calculation What vendors have Are there any changes in What assumptions methodology you used? methodology vs. last year? have you made?

Television 30 seconds Nielsen/PARC Total Revenue/Total Audience TVNZ/MediaWorks/ No Assume network split for Network,cost-efficient mix converting ASA figures to based on demo net-net

Newspapers A3, full colour ASA, Adquest, SMI, Total Revenue/ NZ Herald/Waikato Times/ No Assume network split for Clear Descisions (Total Readership x14) Dominion Post converting ASA figures to net- net, assuming an average OTS of 14 full pages of advertising per issue

Magazines Full page colour ASA, Adquest, SMI, Total Revenue/ Womans Weekly No Assume network split for Clear Descisions (Total Readership x8) (NZ and AUS)/Woman’s Day/ converting ASA figures to NEXT/ North & South/YH&G net-net, assuming an average of 8 full pages of advertising per issue

Radio 30 seconds ASA, Adquest, Total Revenue/ TRN/RadioWorks No Assume an average of 8 Clear Descisions (Total Listenership*8) impacts per listen

Out of home Various ASA, Adquest, Total Revenue/ iSite/APNO No Assumes 70% of population 2013 Census Data (Population Main Mets*70%*20) of 5 main mets see an average of 20 OOH ads per day

Cinema 30 seconds ASA, Adquest, Val Morgan Total Revenue/ Val Morgan No Estimated impacts come (Total Admissions x3) from Val Morgan are not independently audited and are based on admission levels

Online Banner 468x60 Yahoo/Stuff/NZHerald/ Based on market Yahoo/Stuff/NZHerald/ No Estimation based on Premium Ad2One/Adhub/Mediaworks/ knowledge, client database, Ad2One/Adhub/Mediaworks/ Display advertising only MSN/TVNZ/Metservice and historical dealings MSN/TVNZ/Metservice

TVNZ | AdEffect Series Summary Page 39 Kantar TNS communications evaluation measurements

Motivation AMP Long-term Impact, i.e. creating memories Short-term impact, i.e. driving immediate that build brand equity by creating connections appeal towards the brand and strengthening neural pathways

AdEval questions determine an ad’s ability to change Affective memory potential (AMP) is based on neuroscience someone’s attitude or behaviour towards a brand findings about how the brain forms memories. It uses a gateway theory; in order to be motivated you We measure the three critical elements an ad requires in must first be involved in the ad. Motivation is proven to order to build or strengthen memories about the brand. be predictive of sales effects of advertising. Where all three elements are present, the ad can build long Driving short-term activation is key to immediate success term brand equity and influence future decision making. and we measure this via our AdEval questions which are validated against sales effects. It is validated through neuroscience testing and through modelling against brand equity effects.

TVNZ | AdEffect Series Summary Page 40 Thank You