Researching Authenticity. An Interview with Professor Watson and Danielle Wiley

What is authenticity? Dr. Jared Watson, an assistant professor of marketing at NYU Stern and Danielle Wiley, Founder & CEO,  SWAY GROUP, partnered on a study to find out exactly that. The study was launched pre-pandemic and offers insights into what we can expect going forward..

“Authenticity has been super important to us from day one,” Wiley explained, “We always had a gut feeling that our push for authentic content was a major driver of the high engagement rates on our campaigns. When we realized that Jared and his colleagues had the capability to actually MEASURE this for us, we couldn't resist.”

The study helped to reveal, how authenticity is actually manifested within a piece of content. “Certain words, use of emojis, type of photos - it's been just so interesting to get all of this juicy data and we can't wait to translate those learnings into improved post instructions for the creators in our network,” she added.

Charlene Weisler: Is this the first time it was done? If yes, are there plans for a follow-up? If not, what are the trends?

Dr. Jared Watson: To the best of our knowledge, this is the first time sentiment analysis has been used to explore the effectiveness of influencers. There has been other research that concludes that authenticity helps influencers, but this work is based on subjective evaluations of authenticity whereas we are using a methodology that tries to objectively quantify authenticity. We certainly plan to further our investigation beyond authenticity to explore related themes (e.g., negative and positive emotion) to understand the contexts by which it is better to express things like vulnerability or excitement. Beyond this, we also plan to explore when other signals (like credibility/expertise) may be more beneficial than authenticity and whether these effects are unique to individual influencers or if these effects hold when brands use similar language.

Weisler: What was the methodology?

Watson: Sway Group provided us with data from 20+ campaigns that include captions of the social media posts along with various engagement metrics like likes/comments/shares/etc. We then used the Linguistic Inquiry and Word Count (LIWC) software program to convert the text of the captions to quantitative scores for a variety of categories. Because our focus is currently on authenticity, we then used simple linear regressions to understand the impact of the authenticity scores (as quantified by LIWC) on the engagement metrics (as provided by Sway Group)

Weisler: What is the definition of influencer? of authenticity?

Danielle Wiley: These days, an influencer can be anyone, from a fellow mom at the playground to a celebrity that you follow on Instagram. That said, for the purposes of what we do at Sway Group, influencers are anyone with at least 1,000 followers on any social platform, usually Instagram, Facebook, TikTok, Twitch or blogs. We define authenticity as content that is relevant and meaningful to the influencer. Are they writing about a product or service that they actually use and are excited about? Are they using their own words and unique voice? Will the content feel real and relevant to the influencer's followers?

Weisler: What were the big takeaways?

Watson: I think there were two big takeaways. Authenticity is an industry buzzword and something influencers strive for, and the data suggests that their intuition is valid: posts that are deemed as authentic receive more engagement than those that do not (in this data, posts that receive a 50 or greater LIWC engagement score receive over 3x more engagement on average than those that score under 50). The second is that while we often think of authenticity as a subjective evaluation of the influencer, LIWC's classification seems to capture authenticity quite well, suggesting it can be easily measured and tracked.

Weisler: Were there any surprises?

Watson: I think the biggest surprise for us was how few posts scored high on authenticity. From a pure mean split perspective, less than 10% of posts in this dataset (and less than 5% in another dataset) recorded authenticity scores greater than 50 (out of 100). So while the industry continues to buzz about authenticity, sometimes the pursuit of it actually might lower perceptions of authenticity.

Weisler: Are there any differences in gender, age and region of the country?

Watson: The dataset provided was spread out throughout the country, and consisted mostly of millennial mothers. In some of our experimental work, we find that the effectiveness of influencers is highly predicated on gender (women are more influenced than men) and age (younger vs. older people accept influencers more easily). I'd also add that while we don't see much variance across the various campaigns in our dataset, most of these campaigns were affect-laden things that really impact people's feelings. In another dataset, we have more technology-focused products and we find that authenticity plays a much weaker role (probably because people are more concerned about quantifiable attributes like battery size, screen size, etc.)

Weisler: How can we best use these results?

Watson: These results suggest that there are some guidelines by which we can help influencers more authentically communicate their message. For example, use I-first language rather than you-first. That is, the influencer can speak about how a product has impacted their own life vs. telling their audience how the product might impact their lives. Similarly, there are benefits to speaking about feeling-states (e.g., this product makes me happy) vs. just product attributes (e.g., this product has a 12x zoom). Naturally, these guidelines may vary for individual products, but we can better provide advice for the influencer community on how to communicate their message more authentically.

This article first appeared on www.Mediapost.com

 

 

 

Measurement During a Pandemic with Analytic Partners. An Interview with Analytic Partners’ Konstantinos Spetsaris

Measurement at any time is filled with challenges. There are so many different datasets and parameters to use to quantify an increasingly complex consumer journey.

Now, during a pandemic where the entire landscape is shifting, the task becomes even more complicated. Konstantinos Spetsaris, SVP, Analytic Partners believes that we can apply specific methodology at this time to make greater sense for future forecasting. His company, he explained, offers, “adaptive solutions (that) integrate proprietary technology powered by the latest data science delivered through our platform and high-touch consulting.”

Charlene Weisler: What goes into accurate measurement of and during a pandemic?

Konstantinos Spetsaris: With so many forces at play, a holistic econometric model is best suited to accurately measure and decompose the impact of Covid-19 and its compounding impact on other business drivers such as media, operations and direct to consumer marketing. A simplified formulation of an econometric response model where all controllable and non-controllable drivers are in included as predictors (independent variables) would look like Response=f(Marketing, Non-Marketing and Macro Factors). The model lends itself to quantification and decomposition of impacts, reporting of core performance metrics (ROI, cost per acquisition, response/unit of support etc.) and scenario planning (simulation and optimization).

Weisler: How can you maintain quality data and feedback?

Spetsaris: We are expert at auditing, cleansing and validating data elements to be analyzed.  Our ROI Genome, an integrated database of benchmarking metrics, enables us to sense-check data inputs, which is complemented by a series of algorithmic validations, business logic, and human checks. Our data audits score all model inputs for both aggregate (accuracy, consistency, granularity, completeness) and user level data (coverage, quality, detail, integration).  We work with our clients to identify and resolve data gaps by supplementing (i.e. augmenting the partner ecosystem) or by creating proxies for missing data.

Weisler: Does measurement vary by industry, consumer category etc? If so how?

Spetsaris: Measurement varies in the sense of which KPIs are most critical to any given brand within any given vertical, as well as what data is available per industry. For example, there are industries with an immense amount of 1st party data, like financial services, which allows for extreme deep dives. Conversely, in industries like CPG there is a lack of 1st party data, which calls for a different process to draw out insights. 

Based on the industry, type of data available and KPIs, we can align candidate data inputs to be tested empirically in the model. Our fully specified model includes marketing, non-marketing and known external factors, but given the atypical and disruptive nature of Covid-19, we further explore other indicators which may have a significant impact on sales. These factors may provide additional insights into how changes in consumer behavior: e.g. reduced mobility, increased online shopping impact business performance. Inputs are rigorously tested on significance, independence (multicollinearity), in and out of sample fit (for predictive strength) and data source sustainability.

Weisler: What data is most important?

Spetsaris: What data is most important really depends on what business question is being asked. For that reason, it is critical to have a holistic measurement system in place that allows available data to be viewed through different lenses and dimensions, in order to extract the most relevant answer.

A few examples of the most important factors to consider during the Covid-19 crisis may include:       

> Bayesian Causal Impact – measures a signal in sales itself and the difference beyond expected response variable (synthetic baseline)

> Human Mobility data to account for restricted movement based on government stay-at-home orders

> Macroeconomic Indicators such as consumer sentiment, consumer confidence etc.

> Financial Indicators such as the VIX (Volatility Index) for financial services firms

> Store Closings and Operational Changes in Services e.g. no longer offering dining in for restaurants or adding a service e.g. curbside pick-up, changes in business model B2B to B2C etc.

> Category Base Sales to capture shifts in consumer demand towards certain product classes e.g. disinfectants, cleaners, shelf stable food

> Scaled Indicator Variables to capture Out of Stocks and the impact of stockpiling as a result of the initial panic mode buying at the onset of the pandemic

> Covid-19 Incidence as a leading indicator to government orders to shelter in place etc.

> Google Query Volume for specific search terms such as “lockdown”, “virus” etc.

Weisler: How can we effectively capture the impact of Covid-19 with so many other forces at play?

Spetsaris: We recommend starting with a holistic measurement framework such as Commercial Mix Modeling that incorporates controllable, non-controllable, and macro-factors in order to isolate the impact of Covid-19 on the business. From a measurement perspective, there are several factors to consider including: time horizon (immediate vs. longer term impact), industry (benefiting or negatively influenced and to what extent) and unique brand / business dynamics (% of sales impacted, geographic footprint, etc.). As an initial analytical objective, we recommend beginning with descriptive data analysis to help define the impact window in terms of business units, sales channels, consumer segments, and time. The goal is to identify where the impact of Covid-19 may be manifested in the dependent variable and gauge the order of magnitude vs. expectation. This helps refine our search for the right data inputs for Covid-19 as businesses are impacted differently.

Weisler: Can we still leverage historical results to predict outcomes given this unprecedented, non-regular event?

Spetsaris: In a word, yes. Covid-19 has disrupted every business in some capacity, which has influenced business and marketing plans and forecasted performance. In this chaotic state, data and analytics become even more important and measurement approaches must adapt. It is critical to update existing models to reflect new consumer behavior and continuously refresh to assess how these changes impact business performance. But, without an accurate understanding of historical insights and principle based learning as a foundation for these updates, there is no way to measure progress or success – nor is there a way to understand when and if consumer behavior and other key factors have returned to “normal.” 

Weisler: How will we know the lagging impact of Covid-19 as we shift to stabilization / recovery and revitalization phases?

Spetsaris: In the current environment, it is not enough to just know how much a business has been impacted by Covid-19. Brands need to know how the underlying consumer behavior has changed, and how it may continue to change, and better understand the pandemic’s impact on marketing and media channels, shopping habits, competitive actions, and overall business performance. Given the disruptive nature of Covid-19 its epidemic and economic consequences, decisions taken in the short term will have significant ripple effects down the road. With a robust and holistic model framework in place that incorporates all business drivers including the economic and Covid-19 impacts detailed above, brands will have a foundation to monitor business performance on an ongoing basis as we shift into stabilization, recovery and revitalization phases.

This article first appeared in www.Mediapost.com

What about TV Viewability? An Interview with TVision’s Luke McGuinness

Viewability is a challenge for many forms of visual media. How do we really know if a piece of content is seen? “IPG recognized that the industry had never effectively measured TV viewability,” explained Luke McGuinness, President of TVision. “Ads are bought and sold on the assumption of 100% viewability - but that assumption is false.” He should know since his company tracks a range of television consumption behavior.

IPG MAGNA recently released a TV Viewability study using TVision data. The study noted that regardless of device, viewability indicates whether an ad has the opportunity to be seen but it does not guarantee that a viewer has actually seen the ad or whether that ad is effective. The difference between digital and TV viewability is that with digital, the consumer is present, but not all served ads appear on the screen. With TV all ads appear on the screen, but the consumer may or may not be present.

In this interview, McGuinness discusses the study, its genus and its future.

Charlene Weisler: What prompted the study?

Luke McGuinness: More than $59 billion will be spent on TV ads in 2019, without knowing if the ads are viewable. IPG sought to quantify TV viewability so that TV advertising could be measured in a manner similar to digital. 

Weisler: What are its implications?

McGuinness: The study from IPG Media Labs, using TVision data, found that 29% of all TV ads are not viewable. That means they air to an empty room. When we think about the $59 billion (or more depending on the source), ads that air to an empty room are costing advertisers quite a large sum. By evaluating TV for viewability, advertisers can determine what is working, what is not, and optimize for better performance, and therefore substantially reduce ad waste.

Weisler: Please give me an overview of the methodology.

McGuinness: IPG reviewed 6 months of TVision data, tracking 5,388 individuals in a nationally representative panel, tracking 39,464 hours of ads for households, 2,992,414 unique ads, 5,961,757 impressions, persons 2+ and C3. Programming and ads were captured via ACR (automatic content recognition). Participants opted to install TV viewability detection technology in their household. Viewability and attention were measured by using computer vision algorithms.

Weisler: Would you say that the results show that TV and digital are comparable in delivering ad messaging? What are the differences that you found?

McGuinness: The fact that the size of the Viewability issue for TV very closely mirrors digital video shows that the two face similar (29% for TV; 31% for digital video) challenges in delivering ad messaging, but there are some natural differences in viewability for TV and digital. And these differences are rooted in the fact that digital video on PC and mobile are inherently different experiences. With digital, it is presumed that the consumer is there because of the nature of the medium. Someone just clicked to watch a video. As we all know, TV is very different. People leave the room or even leave the home with the TV on. 

Advertisers measure digital viewability as they do not want to invest in ads that do not have the opportunity to be seen. The way we measure viewability for TV delivers the same value proposition - investing in an ad strategy that will deliver an opportunity for people to see the ads. The methodology is different, given that are different mediums, but the value proposition is consistent. 

Weisler: What is your recommended course of action to improve ad delivery and consumption?

McGuinness: TV’s viewability challenge lacks uniformity. It is not limited to specific networks, times of day, programs or content types. While viewability varies across industries, no advertiser is immune. Every brand can improve their return on ad spend with this data. The best step forward for brands is to measure what’s working or not for their historical TV advertising, benchmark versus competitors’ performance, plan a more effective strategy along with their existing planning tools, and measure and optimize on an ongoing basis. By using TVision viewability and attention data, combined with other data such as cost and ratings data, brands can identify higher performing opportunities. For example, the study suggests that brands can find value by buying ad spots outside of prime, and outside of the first spot in the pod.

Weisler: What about pod position?

McGuinness: In general, the first position in an ad pod may not be worth a premium. Ads appearing first in a pod had 72.2% viewability. Ads in the middle of the pod had 70.3% and those at the end of the pod had 69.9%. Longer ads have higher viewability but doubling the length of an ad does not double the viewability, so the cost of longer ads must be considered.

Weisler: What are next steps?

McGuinness: The immediate next steps are for advertisers and networks to incorporate viewability into their ad buying and selling - and many have already started to do so. They can do this by analyzing their historical performance for viewability, analyzing their competitors’ performance, and learning from that.  Additionally they’re leveraging broader planning data from TVision to best optimize their campaigns and they are now measuring their specific performance on an ongoing basis.

This article first appeared in www.Mediapost.com

Q&A Interview with Henry Schafer

Henry Schafer, EVP Marketing Evaluations, Inc. The Q Scores Company is a media research veteran from the consumer research / advertising agency sphere. His current company, Marketing Evaluations, is best known as the inventor of “Q-Scores” which indicate consumer favorability and marketing potential for celebrities as well as for programs, brands and licensed properties.

According to Henry, the media landscape has seen a sea change over the years with greater complexity and more challenges to research and methodologies. In this compelling interview, Henry discusses specific challenges to research, the future of media, measurement (and value) of celebrity favorability and negativity and the advancement of a branch of celebrity measurement that one might not have envisioned – the current media value of those stars who are deceased.

The five videos of the complete interview are as follows:

Subject                                             Length (in minutes)
Background and Predictions              (7:09)
Marketing Evaluations                       (4:20)
Deceased Celebrities                          (6:35)
Negative Q Scores                              (6:52)
Methodology                                      (7:29)


Charlene Weisler interviews Henry Schafer, EVP Marketing Evaluations. Henry discusses his background in this 7:09 minute video:


In this 4:20 minute video, Henry Schafer talks to Charlene Weisler about his company Marketing Evaluations:




Charlene Weisler interviews Henry Schafer, EVP Marketing Evaluations about the company's use of
Deceased Celebrities Q-Scores for licensing valuation. The video is 6:35 minutes:




In this 6:52 minute video, Henry Schafer, EVP Marketing Evaluations talks to Charlene Weisler about Negative Q-Scores adn their importance in ascertaining a celebrities connection to the audience:




Henry Schafer talks to Charlene Weisler about the methodology to calculate all the different forms of Q-Scores in this 7:29 minute video: