Marcello Sasso, MBA

💡𝐇𝐨𝐰 𝐯𝐞𝐫𝐭𝐢𝐜𝐚𝐥𝐥𝐲 𝐢𝐧𝐭𝐞𝐠𝐫𝐚𝐭𝐞𝐝 𝐢𝐧𝐬𝐮𝐫𝐞𝐫𝐬 𝐜𝐚𝐧 𝐠𝐚𝐦𝐞 𝐭𝐡𝐞 𝐬𝐲𝐬𝐭𝐞𝐦 Uh oh. Alternative payment models (APMs) and value-based arrangements can be structured to exploit the medical loss ratio (MLR) loophole: • Insurer A pays its affiliated clinic $500. • The clinic’s actual cost is only $300. • The $200 difference becomes internal profit, but is still counted as “medical spending” in the MLR calculation. • Lump-sum APM payments (per capitation or per episode) hide price increases Meanwhile, a non-integrated competitor paying the same clinic $300 records no internal profit. 👉 Massive consolidation and vertical integration is yet another unintended consequence of the Affordable Care Act 🙄 Full analysis in Health Affairs: https://drugch.nl/3VNNKFm

175

30 Comments

Market research fails not from lack of data, but lack of direction. We’re living in an age of information overload — on average, a company uses over 28 different data sources for decision-making (Forrester, 2023). Yet, McKinsey’s research shows that only 30% of analytics insights lead to measurable business impact. Why? Because data collection is often treated as the goal, not the tool. Before launching a survey, commissioning a focus group, or mining dashboards, we should ask: What decision will this research influence? What does success look like for the business, not just the dataset? Who will act on these insights, and how? Harvard Business Review calls this “decision-backward thinking” — starting with the end goal and working in reverse to design the research process. Without direction, market research becomes a repository of unused numbers. With direction, even small, targeted studies can lead to transformative outcomes. **Data without a plan is noise. Data with purpose is power.**

13

1 Comment

Surveys & Custom Research = Thought Leadership. One of the strongest ways to build trust and authority? Original research. When companies commission or publish their own surveys, whitepapers, or technical studies, they’re not just selling a product but shaping a narrative. Data-driven insights signal real expertise, making journalists, analysts, and even AI systems more likely to CITE and amplify your findings. And yes that word “CITE”? That’s exactly where our name came from. 😉 #MarketResearch #Cybersecurity #CITEResearch #PR

10

5 Comments

Inspired by talks & conversations at Quirk's Media LA last week, here are some thoughts on how market research AI offerings affect the world of survey respondents & data quality: In market research, AI offerings currently solve for 3 primary use cases: 📊 Structuring unstructured data  🦾 Chat bots / AI agents 🫥 Synthetic data to predict/expand on existing data AI allows researchers to move faster and/or do more with the same amount of time. There are tons of companies vying to be the leaders in each of the above categories, making the number of AI conversations greater than the actual number of uniquely positioned solutions. AI also creates market research problems: ��� AI’s ability to take/fake surveys as a respondent, ie fraud 😵‍💫 Hallucinations or incorrect return, requiring time spent double checking results 📈 Shifted baseline of demand for everything to be faster, even when not a fit for a given methodology It also creates anxiety around jobs, layoffs & how to reinvent yourself to stay employable in this quickly evolving landscape. But here’s something AI doesn’t currently solve for:  How do we reach & motivate real audiences of interest to take surveys with honesty? (Isn’t this kind of the point of what we’re all doing here anyway?) This question is deeply impacted by the AI induced pressure to move faster, and the answer is deeply rooted in the natural human world: There are a limited number of real people taking surveys at any given time. The gap between the number of real people taking a survey any given day and the demand for a certain number of people to fill a survey = the void filled with fraud. This is the world that encourages the “professional” survey taker. There are tons of tools that can identify the likelihood of fraud, but not enough solutions for filling that void with real respondents. So what can you do if you care about quality? Know your panel sources. Learn about their recruitment, opt-in, incentive, and quality control processes. Understand what a realistic sample size is for a given audience in a given timeframe. Build in field time accordingly. Pay attention when it sounds too good to be true. Design surveys with fraud checks & awareness of the impact survey experience has on completion rates. Respondents are still the fundamental building block of primary market research. We can’t look at the need for results yesterday as separate from the perspective of people taking surveys today. If you want better data quality, there are ways you can benefit from the speed of AI tools and still get results you can trust. To get the best of both worlds, you just have to understand the new rules of the game. Reach out if you want to talk sample strategy in the age of AI.  p.s. I spent time at Quirk’s listening to different AI presentations to summarize the use cases, but if you think I’m missing any, please comment. #industryinsights #ai #marketresearch #dataquality

22

8 Comments

Op-Ed: Market Research Is at Risk — But Behavioral Data Is the Savior Traditional market research is on the verge of obsolescence. Survey fraud, synthetic responses, and outdated methods have eroded trust—and generative AI will finish the job. Why pay for human-written reports when LLMs can synthesize thousands of responses in seconds? But not all is lost. The real future lies in behavioral data—what people actually do, not just what they say. App usage, website visits, and location patterns offer a far more reliable picture of the consumer. Companies that own first-party behavioral data will become the power brokers of the AI age. As AI models hunger for real-world inputs to stay relevant, proprietary behavioral datasets will become the new oil—fueling training, personalization, and prediction across every industry. The value won’t be in asking better survey questions, but in knowing the customer before the question is asked. The future of market research isn’t in sample sizes or Likert scales—it’s in human behavior at scale. Firms that capture, unify, and ethically apply this data will lead. The rest will be replaced. Behavior doesn’t lie. And in the world of AI, it’s the only truth that counts.

25

🚀 EXPERIENCE MATTERS!!!! #TUESDAYTRENDS 🚀 Great job on last week’s poll question, sales experience beats other types of experience on overall performance! In order of importance: 💼   Sales Experience (Sales EXP) → 0.31 (the strongest positive impact) 📞   Business Experience (Bus EXP) → 0.26 (moderate positive impact) 🏢  Company Experience (Comp EXP) → 0.09 (a weaker correlation) 🔎 Interpretation Sales Experience is the clearest differentiator overall — more time spent actually selling drives better performance. Business Experience also helps, likely because it gives reps broader context and credibility. Company Experience shows a more nuanced story: -At low tenure, company experience doesn’t matter yet. -At high tenure, company experience does make collaboration, customer service, and relationship-building more impactful. ✅ Takeaway: If we were to rank the true drivers based on our assessment dataset: 1) Sales Experience → strongest positive link 2) Business Experience → meaningful but secondary 3) Company Experience → non-linear (not helpful at low tenure, but critical later on for collaboration and service skills) But wait, there’s more!!!!! Let's check out some moderating relationships! 🏢 Business Experience -High business experience individuals show a strong performance lift from formal presentation skills, integrity, and emotional intelligence. -Those with low business experience don’t get as much performance benefit from these skills — suggesting these are “later stage” differentiators once someone has a broader business context. -In short: the more experienced in business, the more soft skills and credibility traits (like integrity, EI) drive performance. 💼 Sales Experience -High sales experience makes technical selling skills far more predictive of performance: Adaptability (+0.31 correlation gap high vs low) Time management (+0.31) Financial acumen (+0.31) Objection handling & recommendations (+0.27–0.28) -These skills really start to “pay off” once reps have logged more sales time. 🏢 Company Experience This is the most striking set of differences: -Relationship building’s correlation with performance grows by 0.35 from low tenure to high tenure. -Customer service also increases strongly by 0.40. -Engaging others and collaboration are far more predictive at high company experience. Early-tenure folks may be focused on execution, but once they “know the org,” collaboration and customer connection skills become critical to performance. ✅ Takeaway: Business experience → soft skills shine Sales experience → selling mechanics drive performance Company experience → collaboration and customer service emerge as the differentiators #SalesLeadership #SalesPerformance #SalesEnablement #TalentAcquisition Tyson Group

51

1 Comment

Segmentation analysis is an extremely powerful tool for providing stakeholders with strategic guidance - but it’s also a potential minefield if you’re not careful with how you do it… Here’s a quick reality check on the kinds of pitfalls you should avoid: 1. Too Many Variables, Not Enough Focus Throwing every variable into the mix sounds thorough, but it’s a recipe for confusion. The “curse of dimensionality” means that with too many variables, meaningful segments disappear, and everyone starts looking the same. Imagine trying to find patterns in a haystack made of haystacks. 2. Sample Size Struggles Small samples and lots of variables? That’s a fast track to unreliable segments. If you don’t have enough data, your results won’t hold up - and your segments may just be random groupings. 3. Masking & Redundant Variables Including variables that don’t actually differentiate customers (masking variables), or those that are just duplicates (highly correlated), can hide real structure and create “rubbish bin” segments. It’s like trying to tune a radio with too much static. 4. Algorithm Instability Many clustering methods are sensitive to starting points and can produce wildly different results each time you run them - especially if your data isn’t well-prepped. 5. Response Biases Rating scales can often serve the creation of artificial segments (like “high raters” vs. “low raters”) as different geographies (and even individuals) can use scales entirely differently, responding with their own bias. The output of your analysis then becomes just a descriptor of these biases. Ultimately, segmentation analysis comes with a myriad of issues and simply running a clustering algorithm on your collected data is not likely to be a good enough approach if you care about a quality result. At our Nuremberg event, my colleague Keith Chrzan will be taking the Intro track through his fantastic training on how you too can undertake Segmentation analysis badly – and well (we recommend you do it the latter way).

58

15 Comments

Market research is evolving rapidly, and will bring even bigger shifts in the current and succeeding years. Synthetic data is emerging as a game-changer, helping address data quality challenges and respondent fatigue. At the same time, researchers are under growing pressure to demonstrate ROI and tie insights directly to business outcomes. With increasingly complex products and pricing models, traditional research methods need a more customized, strategic approach. Success in this new era will require embracing AI-driven solutions while ensuring research remains actionable and relevant to decision-makers. The future belongs to those who can combine technology, strategy, and insights to drive real business impact. #MarketResearch #AI #SyntheticData #BusinessImpact #FutureOfInsights

17

2 Comments

One thing I NEVER understood when I was a client-side researcher, was the hoarding and protection of information...FROM OTHERS INSIDE THE BUSINESS! Sounds wild, right?!...but it happens so much. Insights were often deemed to be approved for dissemination on a "need to know" basis...but shouldn't the entire business have access to information that helps them become smarter about their consumer? Efforts to more broadly send out insights were met with barriers and carefully curated distribution lists, and focus was placed ONLY on those in the direct circles of those who requested it. But I firmly believe that successful organizations are informed organizations. When more people understand the consumer, better decisions get made at every level. Businesses talk about being "consumer-first", but this doesn't happen in silos, it happens as a collective. Information and insights should be treated like a catalyst for the collective instead of like currency for the c-suite!

24

7 Comments

Synthetic data is on the rise in #marketresearch, used to augment datasets and create synthetic personas. In this final video in our series discussing the issues around survey data quality, we explore why synthetic data cannot fully replace real, human participants. While it can be highly plausible, it may not always be accurate or representative, especially for edge cases, emerging trends, or new categories. TL;DR : Robust research with verified human participants remains essential for generating new insights and making high-stakes decisions. #TheFutureofQuantisQual inca | Nexxt Intelligence Josh Seltzer Kathy Cheng Ben Leet Tim Benner Mary An Louis Devane-Waters Ellie Ma  

30

1 Comment