INSIGHT BYTES BLOG

Insight Bytes is the official MRIA blog, capturing data, insights, and stories by research professionals from across Canada.

Our monthly newsletter is shared across all MRIA-ARIM's communications channels. Do you want to add your voice to the mix of innovations with a thoughtful and well-researched blog post? We want to hear from you!

To feature your latest research project, ideas or perspective in this space, please contact MRIA-ARIM's Chief Editor Arundati Dandapani blog@mria-arim.ca. 

<< First  < Prev   1   2   Next >  Last >> 
  • 03/05/2020 5:37 PM | Dan Jackson (Administrator)

    Michele Ronsen, Founder and CEO of Curiosity Tank, led a recent Remesh webinar on the role of bias in introducing incentives to research respondents. You can watch the full recording here. According to Ronsen, there are at least 180 documented types of biases. Some are captured below, followed by additional insights on bias in research design in our conversation with her described below.



    According to Ronsen, in research studies, the more perspective and experience you have, the more positioned you are to shape and identify bias. Self-awareness is the first step in the identification of bias. Then a conversation with one’s colleagues and team around biases assumptions, whether it’s about the topic you are learning about, the stimuli you are testing, the moderator, participants, or even the researchers. Bias is best analyzed in “360 degrees”. Among the top ways to reduce bias, suggest Ronsen, is to involve more colleagues. It is always good to have multiple perspectives involved in the study including a designer, product manager, a more senior member of either the research or product team, sometimes a content strategist and someone from engineering --typically around 4 to 5 people. A pilot for every study is critical and providing feedback on the pilot and the context are important to reducing bias in any research study.

    While income levels might have an impact on respondents’ motivation to participate in an incentivized research study, it really is not the central reason why people participate. Moreover, it is respectful and a good gesture to understand the time commitment a study involves, and incentives demonstrate that researchers care about the respondent’s time— especially critical to time-scarce professionals like doctors and attorneys. An incentive could be monetary or non-monetary. Her own clients tend to orbit in the big tech space, where access to food and shelter is not an issue.

    With team biases, says Ronsen, teams are much closer to the problem space and frequently think they know the best solutions prior to conducting research however they are often very surprised by participant’s feedback! Being cognizant of such (self-confirmation) biases mixing up the sequence you show concepts to participants, and making sure the moderator is prepared is one of the best ways to mitigate “bias”.With team biases, says Ronsen, teams are much closer to the problem space and frequently think they know the best solutions prior to conducting research however they are often very surprised by participant’s feedback! Being cognizant of such (self-confirmation) biases mixing up the sequence you show concepts to participants, and making sure the moderator is prepared is one of the best ways to mitigate “bias”.

    While Ronsen talks about improv being an important skill or tactic to exercise in conducting research, she emphasizes the importance of preparation and structure in every study. A great moderator and user researcher leverage the fundamentals of improv when getting into character. Thinking on your feet and pivoting happens on the job, and that’s where the improv comes into play especially when respondents don’t respond the way you expect. There are opportunities to pivot throughout the study, and getting into character and into the mindset with the confidence to pivot is learned with time and experience, she suggests.

    Is there a place for AI in reducing bias in research studies? When talking about AI powered bias detection tools in the marketplace (example BiasCorrect, a Slack plugin app that flags problematic phrases or language before use in the workplace used by the bank of Nova Scotia), she says she hasn’t encountered many. The prevalence of co-parenting apps that mediate language between divorced co-parents, to make sure their communication is respectful, sounds like an interesting application that could or might extend to research design. Co-parenting apps for people who share custody of a child can help you monitor the language and tone you use when communicating. The detector feeds off some key words, tone or even emojis if the communication is via text. Its goal is to help partners communicate in a respectful manner before they regret their interactions. So much has to do with tone, even in any research study, points out Ronsen.

    When developing a set of best practices or guidelines that control for bias, Ronsen reminds that it is critical to first identify biases with yourself and your teams. Then lay out those assumptions in order to know how to check them at the door while in session. While there will never be such a thing as bias-free research, we can reduce it. Self-care is another important element for researchers that calls for taking the time before a study to make sure we are recruiting the right people and screening them properly and reviewing the discussion guide to make sure it is fair and balanced so that we are fully present.

    The goal with research is about soliciting multiple perspectives to reduce inclinations or biases when you’re in character and when conducting studies. How does one learn to reduce moderation bias? Record all your sessions, go back through your sessions with a checklist,  and answer questions very candidly, are this leader's last words of advice: “Record everything. Rehearse everything. Bias comes from so many different sources.” 


    Further reading: Portigal, NN Group, Curiosity TankSign up for Curiosity Tank's newsletter here.




  • 03/05/2020 5:36 PM | Dan Jackson (Administrator)

    By Christie Christelis

    Companies succeed or fail based on their ability to innovate. Over the last century, we have seen how rapid changes in technology can alter the economic landscape dramatically – and increasingly quickly. One of the challenges that businesses face is understanding which factors influence the uptake of innovations and how those factors might influence their success. Diffusion of innovations amongst consumers is a complex phenomenon based on the individual decisions of consumers, brands, suppliers, etc. In many cases consumers' choices are based on a simple set of factors (rules and properties), such as awareness, availability, social influence, and the availability of substitutes, among many other factors. Consumers are also heterogeneous, which manifests itself as differences in their behaviours.

    Marketers and strategists are often charged with developing forecasts and scenarios for assessing the impact of innovations. While there are many approaches for addressing this, agent-based models (ABMs) have the unique ability to reveal insights into emergent behaviour in complex markets. One example where agent-based modeling has been particularly useful is in gaining insight into the dynamics of diffusion in the mobile payments market.

    For a mobile payment to take place, the consumer must have a device that is able to conduct a mobile payment transaction with one or more of their accepted payment credentials (e.g. Apple Pay set up for credit card payments on an iPhone). However, that is not enough. The merchant where the consumer shops must also be able to accept the mobile payment. Only when both have the capability to execute a mobile payment can a mobile payment be conducted, otherwise the consumer and merchant have to revert to a traditional form of payment.

    In an agent-based mobile payment model, consumers could be represented as agents, as could merchants. Each individual consumer agent decides to adopt the mobile payment innovation based on several factors, for example, awareness, social influence, affordability and availability of the device, preference, trust, among others. Social influence is a powerful factor, and it relies on linkages or connections between consumers. In an agent-based model, these linkages between consumer agents are themselves agents.

    On the merchant side, acceptance may depend on other factors, such as merchant awareness and understanding, the business case for adopting mobile payments (since there are usually costs involved), the interest that consumers display in using the payment mode, the difficulty of adopting mobile payments, and so on. The power of ABM is in being able to vary each of these parameters to see what the overall impact on adoption and usage might be.

    In the agent-based model shown in the accompanying video we demonstrate the implementation of a relatively a simple mobile payments market. Each agent’s decision (merchants and consumers) has been modeled stochastically to represent diversity in the agent populations. For every time that the model runs each agent makes a decision to adopt mobile payments or not based on the factors that influence that agent. It is important to note that this particular model does not include post-adoption attrition or rejection i.e., a consumer deciding to stop using mobile payments after initially accepting it. The following video I made offers a quick illustration:

    One of the ways that ABM can be of value to stakeholders in the payments ecosystem would be to provide key insights into what the optimal balance might be between building out the acceptance infrastructure and promoting consumer adoption of mobile payments. Running the model a few thousand times under different innovation uptake scenarios (by varying each of the parameters in the model) could provide important metrics on how consumer uptake and merchant acceptance interact with each other. In the analysis of the model output, it would be fairly simple to add cost functions to each output scenario, allowing payments system stakeholders to assess the business case on either side.

    While it is possible to model each agent’s decisions in an extremely complex manner, a cornerstone of ABM is to adopt Albert Einstein’s maxim, namely that “everything should be made as simple as possible, but no simpler.” The goal with ABM is not to recreate reality, but to create a useful model, one that can demonstrate interesting emergent behaviours that may arise from interactions between agents.

    Agent-based modeling or ABM is an excellent exploratory tool for gaining insights into complex phenomena. Emergent behaviour that arises out of complex multi-agent systems can reveal critical interactional dynamics that are not readily apparent from a purely analytical simulation. And as shown here, it can be a very useful tool for understanding the diffusion of innovations in payments.

    If you believe that agent-based modeling could be useful for your business, please do give us a call.


    Author: Christie Christelis is the President of Technology Strategies International and an established Fintech expert, entrepreneur, data scientist and organizational culture revolutionary.



  • 03/05/2020 5:34 PM | Dan Jackson (Administrator)

    By Olga Dubanevych, CMRP

    Marketing in an Era of Search, Experiences and Referrals

    Technology impacts lifestyles, communications, and viewpoints across the generations: millennials, generation Z, boomers, seniors, etc. Videos, visually striking images, and social media platforms inspire people to seek new experiences, emotions, and excitement. According to a Boston Consulting Group study, 72% of millennials say they’d like to increase their spending on experiences than physical things in the next year, pointing to a growing consumer appetite for experiences over products.

    In fact, consumer behaviour is complex and often driven by emotions consumers don’t always understand. Augmented reality (AR) and Virtual Reality (VR) could help channel some powerful marketing and consumer insights tools that delve into the desires and fantasies of consumers. A 2016 Deloitte study reveals that by 2025 annual revenues for the AR/VR industry would reach $692 billion. The study also revealed that 88% of mid-market companies are already implementing some forms of virtual / augmented reality in their marketing campaigns.

    Think about the growing popularity of e-commerce and the share of products and services purchased online today. Virtually any customer journey begins with an internet search that compares various products and services available in market. Using AR/VR, online shoppers can now simulate real-life experiences with products or services, growing their product awareness, interest and purchase decisions. Simulated experiences help build trust, loyalty, and retention as well as encourage consumers to share their experiences with peers.

    A study by Vibrant Media evaluates the influence of its Vibrant 360° software which enables consumers to put themselves at the center of a brand, product, or place. With new formats consumers can virtually visit a tourist destination, explore a computer game or watch a movie through the eyes of the protagonist, attend a fashion show, performance or special event. The same study revealed that their VR software demonstrated 600 percent higher interaction rates, 700 percent better content recall, 2,700 percent improved brand recall, and 200 percent increase in purchase intent compared to standard 2D video ads.

    It is evident how VR/AR experiences enable not just powerfully personal and interactive experiences, but also help unite brands with people and networks in ways that can offset more brand chatter than conventional marketing methods. A Touchstone Research study revealed that 79% of customers will search for additional AR/VR experiences and 81% would tell their family and friends about it. Word of mouth marketing plays a strong role in building better brands, and AR/VR only amplifies this effect.


    Building Consumer Trust with AR/VR

    When shopping at a physical store, it might be challenging to imagine how some products or services work in real life without proper demonstration. Have you ever looked at a closed jar of paint on the store shelf and wondered how it would look like on the actual wall, in your apartment? Would it not be more gratifying to know in advance how a new haircut and colouring would work particularly on your hair, or how it would fit with your face or appearance? Wouldn’t it also be exciting to have a 360-degree view of the hotel room you are planning to book for your vacation? Wouldn’t it increase your customer experience and satisfaction?

    "Forty-five per cent of customers say they would like to try AR/VR while shopping, while 30% of consumers report that they would not visit other stores again if AR enabled them to buy the right size of clothing with confidence online."

    NIKKI BAIRD, VP OF RETAIL INNOVATION AT APTOS , NIKKI BAIRD, THE VP OF RETAIL INNOVATION AT APTOS. MAR 18, 2019.

    AR/VR could also help in educating customers about products and processes. It could be really challenging to gain consumer trust when your potential clients don’t know what actions to expect during complex or detailed vehicle maintenance, medical or dental procedures. However, AR/VR technology could help consumers visualize even complex processing and help people overcome their fears. Another interesting application of consumer education via VR/AR is the vacuuming robot that cleans their surface or the video journey of how a tea leaf becomes a flavoured beverage. The adventure would start with the montage of a real-life tea plantation and customers watching a pre-filmed documentary of nurturing, collecting and sorting tea leaves without having to travel to India or Sri Lanka. Then they would follow the tea leaves from the plantation to the factories as they made their way into stores across the globe. Finally, viewers would discover how tea leaves lend to the taste and water to become a flavoured beverage. Finally, you can watch the simulation of a real-life tea ceremony which traces its roots to ancient Chinese culture.

    Absence Makes the Heart Grow Fonder


    Sports or music fans who did not have a chance to attend a concert often wish they could experience the event remotely. AR/ VR fills this gap between fantasy and experience, offering the perfect opportunity to grow a sporting event’s viewership exponentially. According to Greenlight Insights’ 2017 survey, most popular consumer applications for VR were in the categories of concerts, sports and exercise. Their report reveals that 65% of consumers would be interested in live-streamed events using VR, with the same number being interested in using virtual reality capabilities to watch sports-games, and 58% preferring to use AR/VR to watch concerts.

    Gamification Improves Participation

    Wouldn’t it be fascinating to gamify your brand study to enable your research respondents to have real-life experiences with a product prototype or to test interactive advertising instead of with static routine videos? VR/AR can contribute to better participation rates among survey respondents, enabling the collection of context-rich data, a powerful tool in breaking through the endless sea of products and services for any brand.


    Author: Olga Dubanevych is an Agent at Bill Gosling Outsourcing and is passionate about new technologies and experiences. 

    Article source: Generation1.ca

    Image credits: (1) David Grandmougin and (2)Stephan Sorkin on Unsplash 


  • 03/05/2020 5:31 PM | Dan Jackson (Administrator)

    By David Phillips

    In this age of big data it's easy to believe that we inevitably understand more and more of the world around us.  In fact, the opposite can be true.  We can become blinded by the data we have to the realities not covered numerically.  So how do you uncover your blind-spots, how do you use data smartly, and why should you never, ever have a data strategy?  Here are my thoughts.

    Having worked with data for a long time, both with our clients and also within our company, I know data’s limitations and I’ve noticed how we have come to have a weird relationship with numbers: they cause us to have blind spots.  

    "MARKETING IS NOT A SCIENCE LIKE PHYSICS WHERE THERE IS A RIGHT AND A WRONG ANSWER.  IT'S MORE LIKE WEATHER FORECASTING - THERE ARE PATTERNS OF BEHAVIOUR FROM WHICH WE CAN LEARN"  (SAM GAUNT, HEAD OF MEDIA, LIDL)

    Three reasons why you shouldn’t have a data strategy:

    1. Data strategy is not a thing

    It doesn't make any sense as a concept.  It's like saying you have a 'sockets and wrenches strategy' when faced with a broken-down car.

    Data is a means to an end, not the other way around.

    2. Blind-spots are the downsides of a data-first mindset

    We’ve moved to an era of ‘data or it didn't happen’. We need to recognize that all data is limited, in that there are things it cannot tell you.  Or, more frequently, it creates the impression of being a holistic picture whereas in actual fact it isn't. Take the 2016 US elections where Trump surprised the entire world by being elected the 45th president of the United States. All the data pointed to Hillary Clinton but the data wasn’t showing us the complete picture. (You can read about my thoughts on this here.)

    The problem is that once you prioritize what you see, you de-prioritize what you don’t see, causing blind-spots.

    This has become even more evident with the growth of digital data where reams of data are quickly available.  We focus on things that give us data, rather than on what our customers want, as highlighted in this Marketing Sherpa article.

    3. A data-created brand crisis

    Not only are we potentially missing the complete picture and prioritizing the wrong things, we are creating our own brand crisis by focusing on (the wrong) data.  With the growth on online activation, the rise of white-label products, with the reduction of purchasing friction through online shopping and voice assistants, branding has become more important than ever.  When almost any product can be delivered to your door within a day or so just by talking into a speaker, getting someone to say your brand vs. someone else’s (or Amazon’s own brand) is vital.  And yet, blinded by data, too many brands are doubling-down on short-term thinking and harming their brands and their businesses as a consequence.

    So, what should you do?


    1. Choose a business strategy over a data strategy

    Identify which parts of that strategy can be informed by your data assets and then see where you have blind-spots. Determine whether the blind-spots are in strategically important areas that would be helped if you had more data. This will help you overcome 'number bias', where we think something must be true because it's a number.

    2. Realize what it is, emotionally or politically, that data is actually providing

    We need to be very conscious of the role that data can play when making decisions. The role of data is to help us make decisions, it does not make the decision for us.  We have a tendency to automatically consider data to be facts, making it difficult for us to contradict an argument that uses it.  And yet we know – at least at some level – that not all numbers are factual.  

    Not getting too caught up in the data will help you balance its relative importance around any decision and instead provide you with a better perspective.

    3. Fix the blind-spots

    As mentioned before, the key is not to pretend blind-spots aren’t there, but instead to identify them and find a solution. Once identified, you can get the data, tools or support you need to address them.

    We’ve been working with brands, broadcasters and agencies for ages, so if you are having a hard time convincing your team to focus on a business strategy vs. a data strategy, or not sure how to fix the blind-spots, we can help.



    Source: Nlogic


    David Phillips is the President and COO of NLogic, and is known for creating innovative, profitable company cultures that make clients happy. 


  • 03/05/2020 5:24 PM | Dan Jackson (Administrator)

    By Simran Sethi

    A Changing Fashion Industry

    The fashion industry has significantly evolved over the last decade with the global online fashion market expected to be valued at $765 billion by 2022 and e-commerce constituting about 36% of total fashion sales. Fashion has evolved from pure design to marketing with every brand rushing to social media and mobile marketing innovation today.  New tech devices are being launched constantly that open new doors of customer engagement.  As online and 3D services like Stitch Fix and Matchesfashion.com  become popular, we must also discuss how brands can utilize technology to enforce the idea of a “third space”. Fashion is witnessing a new universal technological revolution via personalization. 83% of Canadians already own some form of smart device in their homes led by smartphones and smart TV and about 2.6 million Canadians already own a smart speaker (Vividata Vivintel Spring Survey 2019). More AI and Experiential Marketing will catalyze the success of fashion retailers.

    Personalize for “Customer Delight”

    Consumers are overwhelmed with choice and competitive pricing. How will fashion brands maintain their niche and go beyond to serve “customer delight”? They monetize personalization, email marketing strategies and automation to target people based on their behaviour, interests, and demographics. Customer experience initiatives need to go even further with individualization –Personalization 2.0–creating a one-to-one relationship with consumers to drive revenues and loyalty. Many fashion brands are realizing their use of AI in ibeacons, algorithms, chatbots, data analytics and 3D printing will lead to more personalized products, curated recommendations, discounted offers, customized packaging, and storytelling that connects to individuals digitally.

    Sephora disrupted the fashion industry with its product variety and “assisted self-service” philosophy, earning double-digit growth in profits and achieving more than 100 stores. Its virtual assistant is a game changer, specifically tailoring cosmetics recommendations to a customer’s exact skin type, tone, eye or hair colour. Sephora’s Virtual Beauty app uses AR to allow consumers to explore different beauty styles by digitally overlaying Sephora products on web or phone-cam selfies.

    Retailers like Fendi, Tommy Hilfiger, Coach and Burberry are already offering monogrammed products to meet consumers’ individual tastes. Consumers’ desires to express themselves through fashion choices and emotional triggers allow companies to innovate with design by hosting a puppy’s face on a scarf or etching their initials or name on clothing. A technology company Carbon is reviving the 3D printing industry and has been working with sports brand Adidas to develop a revolutionary midsole that paves the way for custom high-performance shoes or Future Craft 4D sneakers to meet the needs of every consumer. Released in November 2018, the shoes have gained high popularity.

    Machine Learning, In-Store Experiences, and Community

    The savviest of retailers are leveraging advancements in machine learning and according to Deloitte, one in three consumers surveyed were interested in personalized products, with 71% of them ready to pay a premium for such embellishments. Moving forward, the biggest challenge for fashion stores will be to leave an everlasting impression on consumers, like the excitement felt by Rebecca in The Confessions Of a Shopaholic or  Holly Golightly’s “Nothing bad can happen” feeling in Breakfast at Tiffany’s.

    As consumers now shop from the comfort of their homes, they will choose to invest in experiences rather than products. The store could be a place to meet your friend for a drink, enjoy salsa nights or experience a live fashion event using VR. Given how emotions define who we are, how we live, and form our perceptions of brands, experiential marketing and communities will transform the relationship between brands and consumers forever.

    IKEA picked 100 consumers in the UK (via a Facebook contest) to sleepover in their warehouse in Essex. The experience was unique as it was tailored to pamper the chosen guests with massages, salon services, and options to select their mattresses, sheets, and pillows for the sleepover for free. The company tried to integrate social with experiential marketing, and Lois Blenkinsop, Ikea’s UK PR manager, stated: “Social media has opened up a unique platform for us to interact directly with our customers. Listening to what they want is what we do best, and the Big Sleepover is just one example of how we’re using such instant and open feedback to better inform our marketing activity.”

    In my own previous organization, I was able to execute the same strategy at designer fashion concept store BasantiKeKapdeAurKoffe (clothes and coffee). We invited our existing clients by email, text and in person, to a kitty party with their near and dear ones, sponsoring their food, beverage, placards, games and a live performance in-store for a night. The event helped us engage our loyal customers and brought us new ones. We did this over a period of three months of hosting an event each month and were able to measure the increase in revenue as our sales improved by 30 percent! I agree with Doug Stephens, author of Retail Prophet, when he says that community is the buzzword to master as retailers awaken to the idea that their consumer’s primary problem is no longer the scarcity of products but the scarcity of genuine human connections. In-store experience is king.

    The Innovation Challenge and Opportunity

    Building on customer experience and communities are ambitious marketing strategies that also come with their challenges. AR, VR, and AI call for elaborate and expensive pitches (between $5250 to $28000 to just develop an application) and difficulty in measuring return on investment (ROI). While personalization helps improve customer experience, there will be a thin line between “cool” and “creepy” – that somebody is constantly watching over you or bombarding you with offers. Other caveats are the accuracy of big data mining, ethical AI and data privacy. A brand’s reputation can be compromised if its consumer data is inaccurate because of Dark Social (unattributed online browsing/sharing analytics) as consumers increasingly use messaging apps, email and browse their webs in Incognito Mode.

    Fashion is a rapidly growing, ever-changing industry. According to the BOF and McKinsey in 2019, developing markets were becoming a  “focal point for the fashion industry in the coming years as its middle-class consumer base grows and manufacturing sector strengthens.” With the increasing purchasing power of these markets, the fashion world’s dynamics are changing dramatically. With great challenges come great opportunities, and by targeting customers’ psychology, emotions and developing deep human connections, marketers will thrive. Leading fashion companies will begin delivering on personalization in earnest and in their ability to create communities around their brands while using their retail stores as a media tool to join in the fashion tech revolution.

    Author: Simran Sethi is a brand and marketing strategist with four years of experience in the fashion industry and a Master’s degree in Fashion Management.

    Top blog: This article was also republished on GreenBook Blog and featured among their most widely read blogs of 2019.

    Article source: Generation1.ca

    Image credits: Photo by Jason Leung on Unsplash 


  • 03/05/2020 5:09 PM | Dan Jackson (Administrator)

    By John Stackhouse

    AI may be the new flu: everyone knows it’s coming but no one thinks it will hit them next.

    In other words, no need for that AI flu vaccine.

    In a new Gallup survey for Northeastern University, 61% of Canadians and 71% of Americans think artificial intelligence will eliminate more jobs than it creates. Roughly half of those respondents think AI will also decrease the number of good jobs out there.

    But the vast majority also think they’re immune to automation.


    Despite a general concern about the impact of AI, only 37% of Canadians and 17% of Americans think their job will be eliminated because of smart technologies.

    The Gallup results were presented at a recent Future of Work discussion at Northeastern’s Toronto campus, put on by the Ontario Chamber of Commerce and RBC.

    Helena Gottschling, RBC’s global head of human resources, told the forum that in financial services, automation is leading to new and different jobs – and a changing mix of skills – which means an ever-growing need for lifelong learning.

    More than half of RBC’s 55,000 Canadian employees are enrolled in a digital learning program, often offered in the form of mobile and modular programs for skills ranging from digital troubleshooting to communications.

    In the Gallup survey, 92% of Canadians (and 95% of Americans) said they want some form of ongoing learning through their careers – not surprising given that two thirds of them worry their skills will be outdated within a decade. 

    Despite those worries, few know what skills they’ll need to thrive in an AI-powered economy, but there was strong agreement across Canada, the US and Britain that people will need so-called soft skills like teamwork, communication, creativity and critical thinking.

    Gottschling agreed with that assessment, saying “the shelf life of tech skills will be shorter, while the shelf life of human skills will be just as long.” 

    A greater concern in the study was a lack of confidence in universities -- for Canadians as well as American and Britons – to prepare graduates for the jobs of tomorrow. A rating of high confidence was lowest (3%) in the US.

    A plurality of respondents viewed employers as best-equipped to provide career-long education and training.

    Gottschling said most employees need solutions that are affordable, time-efficient and relevant to their ambitions. And then there’s mindset. Do people have the will and skill to learn?


    Author: John Stackhouse is the Senior Vice President in the Office of the CEO at RBC. 

    Article source: LinkedIn Pulse

    Image credits: Arundati Dandapani at Distillery District's Christmas Market in 2018


  • 03/05/2020 4:58 PM | Dan Jackson (Administrator)


    By Robert Vagi

    If you are like me, you’re probably reading this article because of the buzzwords in the title. Even though everyone seems to be talking about predictive analytics, machine learning (ML) and AI, definitions for these emerging fields can be hard to come by and, more importantly, few people have a clear vision for how they can benefit their businesses. In this post, we will bring clarity to this topic and give you some ideas for how these emerging fields can be used to benefit your business. 

    Starting at the Beginning: Data Analytics

    We can’t have a conversation about machine learning and AI without first talking about data analytics. This is because, to some degree, both machine learning and AI either are based on data analytics or actually are data analytics. So, what do we mean by “data analytics?”  For the purpose of this post, we define data analytics as deriving insights from data – specifically, from data that is collected via means other than “traditional” market research. This could include purchase history, customer feedback on third party websites, and so on. As you can see, this casts a wide net – nearly any analysis can be considered data analytics.  So, where do predictive analytics, machine learning and AI fit in?

    Meet the Twins: Predictive Analytics and Machine Learning

    In An Introduction to Statistical Learning, the authors argue that analyses are done for one of two reasons: inference and prediction.

    What is the purpose of your analysis?

    Analysis for Inference

    • Focuses on describing relationships
    • Helps us understand how the world works
    • Example: Using ad spend and sales data to identify the most effective ad channels

    Analysis for Prediction

    • Focuses on knowing something that’s not currently in the data
    • Helps us understand what something is or what will happen in the future
    • Example: Predicting the likelihood that a customer will renew his or her subscription using historical transaction data

    The reason predictive analytics and machine learning are twins is that both are used for making predictions. In fact, when people refer to predictive analytics and machine learning, they’re often referring to the same thing (though sometimes not – which we’ll explain shortly). Predictive models can take many forms ranging from regression techniques, to clustering algorithms, to more sophisticated models like neural nets or boosted models.  In short, all these analytics tools are used for knowing something that isn’t observed in the data.

    Where predictive analytics and machine learning can differ are in the specific techniques that people refer to when using these terms. For instance, when speaking of predictive analytics, people are often referring to more traditional, though potentially less accurate, techniques like regression or time series analysis. Machine learning, on the other hand, typically refers to more recent developments in predictive modelling like boosted models or neural nets. These techniques are more complicated and require more data, but are also more accurate than traditional methods.

    The Machines are Alive! Artificial Intelligence and Machine Learning

    The definition of AI is the topic of much debate. However, for the purpose of this post, we’ll define AI as the process of teaching a computer to learn from data and act based on what it’s learned – in essence, to learn and act with minimal human intervention.  In most cases, this process is driven by predictive/machine learning models.  Thinking back to the example of whether someone will renew a subscription, if you want your online systems to identify customers that likely won’t renew and then automatically send them a special offer, you would want to look into an AI solution.


    Applying it to Market Insights: Choosing the Right Approach

    So which one do you need? The answer is found through understanding:

    1. Why are you conducting the analysis? and

    2. How will the results be used?

    To this end, here are some guidelines for each: 

    Approach

     Good if…

    Data Analytics (analytics for inference)

    • The goal is to make large-scale, strategic decisions
    • You are at a phase in a project where understanding the problem is key

    Predictive Analytics/Machine Learning

    • You need to understand what will happen in the future (i.e. project revenue, customer turnover, etc.)
    • You want to determine a characteristic of your customers that is not currently measured in your data (i.e. segment membership, phase in purchase cycle, etc.)
    • You have a good understanding of the problem and want to act strategically to promote a specific future outcome for individual customers (e.g. increase the probability of a sale, decrease the likelihood of leaving/churn, etc.)

    Artificial Intelligence

    • You have a good understanding of the problem, you have a large quantity of data, and you want to automate the process to save cost or reduce time spent on the task

    Do you have a data problem that you are struggling to solve? Feel free to reach out to us at robertv@phase-5.com.

    Robert Vagi, PhD, is Phase 5’s lead data scientist, based in Minneapolis, USA. Drawing from his background in both quantitative research and education, Rob is passionate about helping clients use data to tackle their most challenging business problems.


<< First  < Prev   1   2   Next >  Last >> 



© Marketing Research and Intelligence Association

Powered by Wild Apricot Membership Software