Embed Size (px)
Citation preview
34 Market Leader Quarter 1, 2016
Measuring the contribution ofsocial media
CASSIDY, POYNTER ET AL.S O C I A L M E D I A M E T R I C S
Social media has emerged as significant and pervasive. As a new media format, it is also elusive in the sense of the complex challenges it presents to understanding its contribution to marketing effectiveness. In this edited extract from the #IPASocialWorks project, Fran Cassidy, Ray Poynter, Simeon Duckworth and Celina Burnett summarise the key lessons in measuring social media
THE CURRENT view of social media measurement is that there is no ‘best practice’. Indeed, even when best
practice is achieved, it is transitory. Yet the gleaming potential of social media measurement continues to beckon like a will-o-the-wisp over a boggy landscape of increasingly complex communication and insight options.
Since 2013 the #IPASocialWorks project has been developing a comprehensive knowledge bank of social media effectiveness – the equivalent of the IPA’s 30 years of peer-reviewed advertising effectiveness data set, or the
Marketing Society’s Awards for Excellence. #IPASocialWorks is an industry-wide project to help advertisers, agencies, researchers and social media owners develop a more robust approach to measuring the effectiveness of marketing through these new platforms.
It is supported by the IPA in partnership with the MRS and the Marketing Society. This is believed to be the first time that such a wide collaboration across the marketing and advertising industry has been formed.
The three specific areas of focus are Evaluation, Insight and Personalisation. This article documents the work to
date of the first of these: Evaluation. The project was created because many
businesses continue to struggle to see the connection between social media and business outcomes. The industry is now at a critical time, when resource and investment have reached levels way beyond experimentation and at the level where effects can, and should, be felt by the bottom line. This is augmented by the decline of organic reach and the rise of social as a paid channel.
HYPOTHESESWe have developed a number of basic hypotheses on how social media helps
CASE STUDY: RECKITT BENCKISER – NESTED MARKET MIX MODELLING
LYSOL is one of Reckitt Benckiser’s power brands in the US, with a range of cleaning and disinfecting products. Rather than measure at a campaign-specific level, Reckitt Benckiser was looking to measure the effectiveness of each part of the marketing mix overall, across all activity carried out for the Lysol brand over a period of 18 months using market mix modelling (MMM).
Conventional MMM typically measures short- to medium-term response in sales, but advertising often works over the longer term by reinforcing existing habits. If long-term effects
are being sought then the MMM needs to be appropriately configured.
The approach Reckitt Benckiser used to measure its Facebook marketing activity for Lysol is one example of this. As an established FMCG brand, the direct impact of Facebook and other social media platforms on purchases was expected to be low. As a result, the team worked with marketing analytics agency Gain Theory to focus on understanding the role of social media across all stages of the purchase funnel and how these work together to drive sales.
The analysis determined that Facebook
activity was primarily driving the awareness and active consideration stages, and furthermore that viral activity from Facebook users liking, sharing and commenting was generating the greatest incremental awareness, while targeted paid advertising was driving active consideration.
So, while the analysis confirmed that Facebook had a minimal direct impact on sales for the brand, the impact of these customer journey stages on sales shows that Facebook accounted for around 0.2–0.5% of total sales for the relevant period.
Market Leader Quarter 1, 2016 35
CASSIDY, POYNTER ET AL.S O C I A L M E D I A M E T R I C S
The stumbling block that many businesses face is how to measure the business value of these applications robustly to justify and optimise investment in the channel
marketing communications deliver results. Social can be used to:∆ Tell deeper and richer stories than many
other media.∆ Create meaningful contact with
customers and amplify other activities by leveraging and sharing other social behaviours.
∆ Facilitate effective consumer targeting, particularly younger audiences.
∆ Enable real-time management and development of campaigns.
∆ Generate valuable data for enhanced customer insight and delivery of personalised customer communications and experience.
The stumbling block that many businesses
face is how to measure the business value of these applications robustly to justify and optimise investment in the channel.
THE CHALLENGE OF MEASURING SOCIALJust 7% of case studies reviewed by the #IPASocialWorks project have passed peer review to date by demonstrating robust measurement of the business value of social. There are many reasons why measuring social is hard but some of the key reasons stem from the interaction with other channels, the complex interaction between paid and earned media, the lack of control, the multitude of possible data choices and others (described in detail in the report).
36 Market Leader Quarter 1, 2016
HOW TO MEASURE SOCIAL: THE FIVE-POINT PLANTo properly evaluate social media marketing and activities, the measurement process should be baked in from the start. This may sound elementary for any channel or activity planning, but it is still not standard practice. Waiting until the end of a campaign to determine how to measure the campaign will fail, either because the right data has not been collected or because there are simply too many possible interpretations of what happened, or both.
We recommend a simple five-point plan:
1What is the campaign/activity designed to do? How does this link to
the wider business or organisational needs?
2Why use social? Which channels are going to be used? Is social being
used on its own or in addition to other components? Consider its contribution to the roles of paid, owned and earned media.
CASSIDY, POYNTER ET AL.S O C I A L M E D I A M E T R I C S
3What decisions will be made on the strength of the evaluation? And at
what point? During or at the end of the campaign? How would these decisions be made otherwise?
4What are the most appropriate data sets and metrics? How will they be collected?
No single book or document can list and review all the metrics available, although Stephen Rappaport’s The Digital Metrics Field Guide does a good job of reviewing 197 key metrics in 12 key fields. However, metrics generally fall under the following headings:∆ Audit metrics, e.g. what was sent out and
who was reached, including virality.∆ Resonance metrics, e.g. engagement, attitud-
inal, customer satisfaction, reputation effects.∆ Response metrics, e.g. sales, customer
acquisition and other calls to action∆ Metrics that differentiate between the
channels and between the combination of social and non-social components, e.g. TV and Twitter, search and social.
5Design the evaluation process. Determine how the data, metrics and
objectives are going to be analysed to understand the effectiveness of the campaign or activity, and to assess incremental value, including assessing where the value of social comes from.
Causality is, of course, a critical issue in designing an effective approach. Understanding ‘why’ something works is the key to repeatability, learning and scaling. Most importantly, causality is a cornerstone of prediction.
The #IPASocialWorks knowledge bank demonstrates five core approaches that organisations can use:∆ market mix modelling∆ A/B testing∆ brand tracking∆ direct conversion tracking∆ cost reduction.
The case studies herein provide
CASE STUDY: IKEA – GEO-TARGETED A/B TEST
IKEA set out to determine whether social media investment, specifically Facebook, was a driver of new and existing customers to IKEA stores. Two identical groups were created: one group was a control and the other would be exposed to advertising on Facebook. The idea was to physically track each group using their mobile signal to see who actually visited an IKEA store during the campaign period.
The team worked collaboratively with EE and Facebook. Facebook has the mobile numbers of all the users of its smartphone app. This group was then matched against EE’s customer database and split into two random sets. The first ‘test’ group consisted of people whose mobile phone number ended with an even digit. News Feed ads detailing IKEA’s Christmas sale would be served only to this group. Those with an odd number formed the second ‘control’ group, who received no advertising.
EE then helped to create a geo-fenced area around the Cardiff IKEA store. Mobile behaviour of both groups was recorded and analysed. IKEA staff and residents of the ring-fenced area were excluded. The results demonstrated that advertising successfully attracted extra visitors to IKEA Cardiff: geo-targeted Facebook ads drove 31% uplift in store visits among 22–25-year-olds and an 11% increase in visits across all age groups. These additional visitors delivered an ROI of 6:1 against the paid media spend.
CASSIDY, POYNTER ET AL.S O C I A L M E D I A M E T R I C S
Market Leader Quarter 1, 2016 37
examples of these methodologies and demonstrate that social media measurement is not a future fantasy. There are challenges, but the engagement and focus of the industry, as evidenced by the work above, is more than a match for them.
SEVEN KEY TAKEAWAYS FROM THE #IPASOCIALWORKS PROJECTThe #IPASocialWorks body of work highlights a number of broader trends, opportunities and challenges for the effective use and evaluation of social media:
1Social is more than marketing communications. Social is broadening
and changing the definition of media, challenging the concept of a campaign with a clear start and end, and operating as both a service delivery platform and a source of insight, in addition to its marketing communications role.
2Social is changing the way we measure. The rich and real-time
data set provided by social is challenging the evaluation process. Its real-time nature means measurement is now often integrated with the management of campaigns, enabling a more agile approach. This brings a new set of opportunities and challenges for evaluation. Data is more dispersed and collaboration is important.
3Avoid measuring social in silo. Social needs its planning and evaluation
to be integrated with other channels to maximise its benefits, establish its value and be trusted as a mainstream option. We need to prioritise how social works and a siloed approach to social is likely to over- or understate its impact.
4It is easy to overestimate the value of earned media and influencers.
Social can appear in any part of the POEM model, but the value of social is clearly more than ‘free media’: in most cases, social media success is driven by paid media. A study by Convertro in 2014 stated that paid social leads to 25% more conversions than organic social – almost certainly because it is more targeted.
5The commercial value of social will increasingly lie in the richness of its
data. One of the key knowledge banks that the #IPASocialWorks project is beginning to assemble is the use of social for insight and for personalisation of campaigns, products and services.
6Social can learn from traditional planning. Social may be new, but
the planning process is not. The rigour of linking to objectives, based on clear assumptions, using comparable metrics, and balancing short- and long-term effects are all basic planning processes that can be applied to the use of social media.
7Even short-term results need a long-term context. One key issue for
the evaluation of marketing is that the long-term effects are not simply the aggregation of short-term effects, and focusing on short-term benefits, such as numbers of clicks, can make it harder to achieve the long-term benefits. Short-term benefits are often specific actions, whereas the long-term effects are more likely to be related to brand equity and price elasticity.
Fran Cassidy is founder of Cassidy Media [email protected] Burnett is partner at Gain [email protected] Poynter is managing director of The Future [email protected] Duckworth is global head of analytics and research at [email protected]
The knowledge bank comprises a series of case studies that have passed a peer review process using leading members of the industry bodies, and the publication Measuring, Not Counting - How to evaluate social media campaigns, of which there is a ‘Short’ and an ‘Expert’ version. We are continually looking for cases for the knowledge bank. If you feel you may have a case or need further information, please contact Janet Hull at [email protected] or Fran Cassidy at [email protected].
CASE STUDY: BT – COST REDUCTION
BT uses social media (Twitter, YouTube, Facebook and community forums) to deliver customer service excellence – a key part of the organisation’s overall strategy to transform its business and drive growth.
The group identified social media as a priority channel for service delivery due to its ability to reduce demand on traditional, more expensive channels as well as to customer preference and ease of use. Research showed that nearly 50% of customers preferred live chat or social media to more traditional channels, and that 27% of BT customers under 30 had previously used social media to obtain service. In terms of ease of use, social media outscored other channels, including email, online self-service and voice. According to BT’s ‘Net Easy’ scores, live chat and social media outperformed BT.com by 2:1 and voice by 4:1. This score has been proven a critical KPI in terms of churn reduction, increased level of spend and repurchase.
BT has demonstrated a measurable business impact of £2m cost reduction through the use of social media for customer service as a result of deflecting 600,000 contacts a year from the more expensive phone channel. It has also resulted in a measurable impact on the critical Net Easy score, indicating a reduction in churn and propensity to repurchase or increase spend.
The rich and real-time data set provided by social is challenging the evaluation process. Its real-time nature means measurement is now often integrated with the management of campaigns, enabling a more agile approach
