Bob Hayes, PhD

Organizations & AI impact: Interesting AI paper from April '24 - "The Adoptionof ChatGPT" by Humlum and Vestergaard examines the rapid uptake of ChatGPT in Denmark, focusing on the factors driving its adoption across 11 occupations. They surveyed 100,000 workers and found that younger, higher-achieving, and male workers are more likely to use ChatGPT. Employer restrictions and the need for training are the main barriers to wider usage. So already there is a substantial gender gap in AI use seen in this data. Their research suggests that companies need to be proactive in developing policies that will promote safe, widespread usage so that everyone in an organization can benefit from the productivity gains. Seems obvious - however does your organization have such deliberate policies? https://lnkd.in/ewfMG2WA

6

2 Comments

Many people are talking about how LLMs can't count the number of r's in "strawberry." Most people say it's because LLMs aren't good at math, but that's not the issue. To understand the issue, you need to understand how LLMs work and how the internals interact with the task at hand, which comes down to: 1. Tokenization (how LLMs interpret words) 2. Training data -- Tokenization -- LLMs break words down into tokens. Tokens are small words or parts of larger words, but importantly, they're usually not broken up by letters. Why does this matter? It means they just see a word like strawberry as a single unit as opposed to the letters that make up the word, like we see it. This is efficient for processing (and efficient for training the model). See the screenshot below using the GPT-4 tokenizer. Interestingly, strawberry is a single token when there is a space before it (most of the time) and it is broken into multiple sub-tokens when it either starts a sentence or when a quote comes before it. This tokenization quirk would likely add to its confusion slightly and may be partly why strawberry is picked on. "Blueberry," on the other hand, is broken into "blue" and "berry" tokens in both cases. -- Training data -- It still knows "strawberry" is a compound word made up of the words "straw" and "berry" because that would be in its training data many times (one of those things that just comes up in conversation). But there are too many combinations of words and letters for all combinations to be in the training data enough for LLMs to learn it. Note that LLMs still count the letters in many words accurately; that's because they do have some examples in their training data, and they're pretty good at pattern matching to generalize this to other words.

34

2 Comments

Two days until Mercury Analytics' webinar, detailing our unique AI application, which maximizes the value of responses from survey respondents. This webinar is a real eye-opener (not to mention mind-opener) and, to say the least, well worth your attention...which is why so many researchers have already signed up. That said, however, there's still room for more. So I am reposting the invitation below. If you can make it, be sure to do so. You won't regret it! ------------------------------ Tired of fraudulent survey responses? Wishing for richer responses from your survey takers? Well, it's Qual plus AI to the rescue! Meet MercuryIQ - a real-time survey tool that enhances data integrity and helps you get more out of your surveys. Join Mercury Analytics on Oct 24 at 1:00 pm EST where we’ll showcase how MercuryIQ uses AI to flesh out richer open-end responses from survey takers through smart follow ups, and simultaneously identifies and screens for disingenuous respondents. Sign up for the 30 min webinar here: https://lnkd.in/em_SGzhU And if you'd like to learn more about MercuryIQ outside of the webinar, reply to this email or reach out to Leigh Kessler, our VP of Client Solutions, at [email protected].

4

2 Comments

🔍 Just read a fascinating paper on the future of data ecosystem business models in the emergent "Artificial Intelligence of Thinks" (AIoT) space! Very useful systematic research that yields some key insights on how companies are transforming data into value in this developing ecosystem. https://lnkd.in/ert4nU7X 🔄 FOUR DISTINCT BUSINESS MODEL ARCHETYPES IN THE AGE OF AI ARISING 🤝 Adaptive Partnership - High control + High customization - Focuses on developing close, collaborative relationships with customers - Offers deep insights and personalized solutions - Best suited for strategic, long-term business relationships 👥 User-Centric - Low control + High customization - Prioritizes customer experience and flexibility - Enables broad market reach and scalability - Emphasizes user autonomy in configuring solutions ⚡ Standardized Efficiency - High control + Low customization - Focus on operational excellence and cost reduction - Streamlined processes and standardized offerings - Ideal for high-volume, routine operations [3] 🔄 Flexible Transaction - Low control + Low customization - Transaction-focused with minimal ongoing engagement - Emphasizes automation and self-service - Suitable for commoditized offerings 💎 AIOT BUSINESS MODEL VALUE CAPTURE STRATEGIES: 📊 Responsive Monitoring: Goes beyond traditional monitoring by enabling instant actions based on real-time data, including automatic corrective functions and predictive maintenance to prevent failures 🤖 Object Self-Service: Drives objects toward independent functionality with less human intervention, incorporating self-adjustment, replenishment, and self-operation capabilities 💡 Insights as a Service: A data-driven value proposition that leverages AIoT to collect, analyze and deliver real-time and collective insights that give meaning to massive amounts of data, enabling customers to make informed decisions   ✨ Enhanced User Experience: Focuses on improving customer interaction through digital add-ons and customizable interfaces and content that adapt to individual users' preferences and behaviors over time 💡💡 The research shows successful AIoT implementations balance control and customization to create sustainable competitive advantages. 💡💡 🤔 Where do you see your organization fitting in this framework? #AIoT #DigitalTransformation #Innovation #BusinessStrategy

11

Great webinar to check out!

6

Had a fantastic time attending the recent "Reframing Data Storytelling" event by Peekator, where Adam Baker shared valuable insights! Here are some key takeaways: 🔄 Evolving Research Roles: With advancements in tech, researchers are shifting focus from technical tasks to spending more time with people. 📊 Survey Stability: Traditional surveys still have a place in the landscape, proving their enduring value. 📖 Time is the Essence: A major challenge in data storytelling is having enough time to dive into the story and engage audiences. ✍️ Start Early: Design your report alongside data collection, not after, it saves time and sharpens your story. 🙌 Efficiency Wins: Streamlined workflows uplift teams, creating space for richer client and business interactions. 🔍 AI’s Impact: We’re seeing more blending of qual and quant data. Reporting practices will need to keep pace with these changes. https://lnkd.in/eyCYVJ5s

11

Check out The Path https://lnkd.in/gkmMQi_p for career reflections and advice from top Insights & Analytics leaders, across a variety of industries. What does it take to become Medtronic’s VP of Data, Insights & Performance Marketing? Take a look at Aniz Ruda’s career path for inspiration. Register to receive regular updates on featured leaders throughout 2024. #thepathtoinsights

1

There is so much noise in bad data I can’t even😵 In this excellent article “Ensuring quality survey responses in an age of uncertainty” featured in Quirk's Media, Michael Booth, PhD brings up an important point that is often overlooked in data quality discussions (link in the comment 👇 ). Bad data is filled with noise, and although it may appear 'fine' when looking at it in aggregate, it lacks cohesion. This leads researchers to waste precious time hunting for a story that just isn't there. Your ultimate test is a cluster analysis - there is no way to get sensible segments with bad data. As Michael says “Why should you, as a researcher, care about data quality and go to the lengths required to verify that you are receiving high-quality respondents? Like, really? When data is bad it’s hard for you to do your job.” He also has the best illustration to explain noise in the data. 🙌 Researchers - what do you think? Have you been there before? #mrx #dataquality #surveys

134

22 Comments

Brands are still in the midst of figuring out how to best integrate #AI into their day to day business. In this @Business Reporter article, my colleague @Eda Cetinok dives into Ipsos’ GET.AI framework that helps organizations prepare for AI adoption. Check it out: https://lnkd.in/gdDAVKW5

14

While we very much appreciated the attendance at last month's webinar on MercuryIQ, a good many people were unable to be there... ...so we have created a short recap video: https://lnkd.in/ekrZje9j I highly recommend watching at least some of this 9 minute clip. You’ll see how our tool for Qual-at-Scale helps you -Add automated, AI-assisted follow ups to Open End questions in surveys.  -Expand the richness of open-ends (over 200% more responsiveness!) -Remove gibberish -Screen out bad survey takers -And, based on responses we have gotten, make respondents feel more like they are being listened to! If you believe AI has “unlocked a golden age of Qual”, a sentiment we heard from L&E Research's 7th Annual Future Trends of Marketing Research & Technology webinar earlier this month, then I hope you’ll give it a watch to see how MercuryIQ, paired with our AI-assisted Open End Analytics & coding tools, is helping unlock a new age of research for you too. I believe it will invite ideas on how you can deliver qual to your clients with methodologies unimaginable just a few years ago. Our Mercury team is here to help!

4

This article has great suggestions on how to incentivize respondents and increase response rate. Virtual Incentives

19

Serious question about synthetic participants. If we assume that synthetic participants can fill out a survey good enough to stand in for real participants, why do we need the survey at all? If we have a question, and we're saying that we believe the AI + data (synthetic participants) knows the answer, then why have the AI pretend to be 500 people, take a survey, analyze the survey, and get at the answer through a very labor intensive market research pipeline? Why not just ask the AI what the answer is?

6

15 Comments

Prosper Insights & Analytics Uncovers a Shift in National Sentiment on the Vitality of the American Dream The survey reveals a lukewarm belief in the American Dream across all age groups that hard work and sacrifice can lead to home ownership and upward mobility. https://lnkd.in/gb3c-43d #Americandream #economicexpectations #upwardmobility #Prosperinsights&analytics #nationalsentiment #homeownership

1

We often say that we slice data to the point of transparency, and so it is with our analysis of all non-Early Decision application outcomes. We link you to a viz on our Tableau Public page. Spoiler: the regular decision landscape is treacherous. https://ow.ly/1LnV50TsJK4

8

Benchmark Reports are great ... but ... You never really know what a number represent. Ultimately making the benchmark report more interesting than useful How is the number calculated and more importantly: if I did the same calculation with my data - what would I get? Better or worse than the benchmark number in the report? With Dreamdata B2B benchmarks we seek to change this See your data directly next to the benchmark updated live over time !

19

3 Comments

Just a reminder that we released our US social listening report on the Top 30 auto brands. Get it here: https://lnkd.in/gfA77Gwv #automotiveindustry #cars #tesla #insights #sociallistening #data #analytics

5

We get it - time is your most limited resource. When you're overloaded, it’s hard to focus on data and research needs. That’s where Stratus steps in. Our research goes beyond collecting information to delivering insights you can immediately act on. We break down complex data into clear, actionable steps that integrate seamlessly into your existing strategies. No extra burden, just the essential insights to help you stay competitive and make informed decisions faster. #AgBusiness #AgIndustry #AgResearch

9

When I read a book that educates and inspires business success, I have to share it with other people. Today, that book is by Brett Townsend and Tim Hoskins and is called, “Insights on the Brink: Revitalizing the Market Research and Analytics Industry.” The struggle it addresses is summarized with this quote: “Consumer insights is, almost without exception, classified as nonworking dollars, meaning that we’re considered a cost center instead of a revenue-generating part of the business. So from the get-go, we’re fighting an uphill battle to prove our value.” Many of the concepts resonated deeply with my real-world experience, deeply held beliefs, and lifelong passion. Research insights, analytics, and collaboration are the foundation for driving business impact. We hold the key to unlocking opportunities for change, growth, and transformation—yet we often resist the changes needed to fully unleash that potential. Insights on the Brink challenges the traditional mindsets and practices of our industry, urging us to rethink how we approach our work. It offers fresh ideas and innovative methods that help us achieve the meaningful impact we strive for but don’t always attain. Here are a few excerpts that truly resonated with me. Every time I re-read them, I find myself internally shouting, "YES! This is spot on!" • “If you can’t tell me your story in three slides, you don’t have a story.” • “I don’t need to know everything you know; I just need to know what I need to know.” • “The most successful projects are the ones where the client and agency go on the journey together.” • “And further, the agency role should not end with a report, but should follow through on implications and strategy.” • “It’s not about methodologies: We need to give ourselves permission to break free from methodology and inconsequential data and flip our mindset to focus on outcomes so we can play a more powerful role.” I invite you to get the book and discover how it inspires new thinking and actions that can make a real difference for you! Kudos to Brett Townsend and Tim Hoskins for this remarkable book! https://lnkd.in/g5Pmya_6 #MRX #Insights #ConsumerInsight #CustomerInsight #MarketResearch #MarketingResearch #BusinessImpact

28

15 Comments

Love this! #AI can do a lot for us, but over-reliance on certain things cannot be replaced. #AI works best to help facilitate ideas, but there really is not replacement for the human brain. Take a read at the below - a pretty cool paper on this topic by my esteemed colleague Diana Livadic Enjoy! #AI #artificialintelligence #creative #advertising #Creativexcellence #Ipsos https://lnkd.in/gDE4uWuC

6